Bibliography, page 3

1. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
   UniProtKB Q00653 (NFKB2_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

   · UniProt Consortium 2026
   used by NFKB2 (p100/p52) — menin binds the non-canonical subunit too, and activating that arm SUPPRESSES MLL-AF9 leukaemia
2. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
   Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for NFKB2, fetched 2026-09-21

   · Human Protein Atlas 2026
   used by NFKB2 (p100/p52) — menin binds the non-canonical subunit too, and activating that arm SUPPRESSES MLL-AF9 leukaemia
3. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
   PubMed E-utilities esearch, db=pubmed, run 2026-09-20

   NCBI · PubMed 2026
   used by NFKBIA (IkB-alpha) — the switch itself. This graph has entries for everything the switch controls and none for the switch
4. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
   UniProtKB entry P25963 (IKBA_HUMAN), NF-kappa-B inhibitor alpha

   UniProt Consortium · UniProt 2026
   used by NFKBIA (IkB-alpha) — the switch itself. This graph has entries for everything the switch controls and none for the switch
5. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
   IntAct molecular interaction database (EMBL-EBI), PSICQUIC MITAB 2.7, pubid:35140242, retrieved 2026-09-23

   2026
   used by NFYC / NF-YC - a histone-fold transcription factor whose partner subunit carries an H2B-like ubiquitination, and the most selective of the eleven TF baits that reached KMT2A
6. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
   UniProtKB Q13952 (NFYC_HUMAN)

   2026
   used by NFYC / NF-YC - a histone-fold transcription factor whose partner subunit carries an H2B-like ubiquitination, and the most selective of the eleven TF baits that reached KMT2A
7. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
   NCBI PubMed esearch via scripts/pubmed_safe.py, run 2026-09-23; querytranslations recorded: ("NFYC"[Title/Abstract] OR "NF-YC"[Title/Abstract]) AND "KMT2A"[Title/Abstract] and the same with "MLL"[Title/Abstract]

   2026
   used by NFYC / NF-YC - a histone-fold transcription factor whose partner subunit carries an H2B-like ubiquitination, and the most selective of the eleven TF baits that reached KMT2A
8. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
   Human Protein Atlas v25.1, proteinatlas.json.gz record for NFYC (ENSG00000066136), retrieved 2026-09-23

   2026
   used by NFYC / NF-YC - a histone-fold transcription factor whose partner subunit carries an H2B-like ubiquitination, and the most selective of the eleven TF baits that reached KMT2A
9. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
   PubMed esearch, NG2[tiab] AND pericyte[tiab] -> 332; CSPG4[tiab] AND pericyte[tiab] -> 24; querytranslations "NG2"[Title/Abstract] AND "pericyte"[Title/Abstract] and "CSPG4"[Title/Abstract] AND "pericyte"[Title/Abstract]; run 2026-09-23

   · NCBI PubMed 2026
   used by NG2/CSPG4 is the one antigen in the whole surfaceome that KMT2A::AFF1 itself drives - and it is a pericyte and vascular protein, at 124.5 nTPM in normal blood vessel and 246 TPM in tibial artery
10. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    PubMed esearch, (CSPG4[tiab] OR NG2[tiab]) AND "chimeric antigen receptor"[tiab] -> 36 records, titles inspected 2026-09-23; includes CSPG4.CAR-CIK in soft-tissue sarcoma (Clin Cancer Res 2020), CSPG4-CAR-NK in glioblastoma (Cell Oncol 2025) and B7-H3/CSPG4 CAR-T in anaplastic thyroid cancer (J Exp Clin Cancer Res 2025)

    · NCBI PubMed 2026
    used by NG2/CSPG4 is the one antigen in the whole surfaceome that KMT2A::AFF1 itself drives - and it is a pericyte and vascular protein, at 124.5 nTPM in normal blood vessel and 246 TPM in tibial artery
11. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    IntAct molecular interaction database (EMBL-EBI), PSICQUIC MITAB 2.7, pubid:35140242, retrieved 2026-09-23

    2026
    used by NHLH1 / HEN1 / NSCL1 - a TAL1-family bHLH factor found by cross-hybridising with the T-ALL oncogene, whose knockout mouse is normal and whose only imaging call puts it in the GOLGI
12. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    UniProtKB Q02575 (HEN1_HUMAN / NHLH1)

    2026
    used by NHLH1 / HEN1 / NSCL1 - a TAL1-family bHLH factor found by cross-hybridising with the T-ALL oncogene, whose knockout mouse is normal and whose only imaging call puts it in the GOLGI
13. weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.database
    Human Protein Atlas v25.1, proteinatlas.json.gz record for NHLH1 (ENSG00000171786), retrieved 2026-09-23

    2026
    used by NHLH1 / HEN1 / NSCL1 - a TAL1-family bHLH factor found by cross-hybridising with the T-ALL oncogene, whose knockout mouse is normal and whose only imaging call puts it in the GOLGI
14. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    NCBI PubMed esearch via scripts/pubmed_safe.py, run 2026-09-23; querytranslation recorded: ("NHLH1"[Title/Abstract] OR "NSCL1"[Title/Abstract] OR "HEN1"[Title/Abstract]) AND "MLL"[Title/Abstract]

    2026
    used by NHLH1 / HEN1 / NSCL1 - a TAL1-family bHLH factor found by cross-hybridising with the T-ALL oncogene, whose knockout mouse is normal and whose only imaging call puts it in the GOLGI
15. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41371538
    Mixed-lineage leukemia cells undergo unique adaptations in the CNS niche

    · Exp Hematol 2026
    used by Niche biology HAS been measured in KMT2A-rearranged disease - but in the index fusion itself there is essentially one marrow-niche molecular experiment, from 2011, The CNS-adapted state — leukaemia-propagating cells reprogrammed by the central nervous system niche, and worse at restarting systemic disease afterwards, The lipid-dependent state — ACSL4 and polyunsaturated fatty acid utilisation in KMT2A-rearranged AML, against ELOVL1 and saturated elongation in KMT2A::AFF1 infant ALL, Where KMT2A-rearranged leukemia physically sits — marrow niche, CNS, blood, liver/spleen and skin
16. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    NCBI E-utilities esearch counts for the marrow-niche beat, both halves of every intersection, 2026-09-19

    NCBI E-utilities · NCBI Entrez 2026
    used by Niche components considered and not written: what the search actually returned for E-selectin, CAR cells, osteoclasts, marrow macrophages and the rest
17. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    NCBI E-utilities esearch: singular-versus-plural behaviour of multi-word Title/Abstract terms, demonstrated on PMID 33108353, 2026-09-19

    NCBI E-utilities · NCBI Entrez 2026
    used by Niche components considered and not written: what the search actually returned for E-selectin, CAR cells, osteoclasts, marrow macrophages and the rest
18. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paperPMID 42139054
    miR-93-mediated PTEN suppression and CNS-specific T cell exhaustion shape the leukemia niche in infant KMT2A::AFF1+ B-ALL

    Duguid A, Malouf C, Leah T, Nitsche L, Barrett NA, Smith OP, Halsey C, Ottersbach K · Cell Reports 2026
    used by Niche components considered and not written: what the search actually returned for E-selectin, CAR cells, osteoclasts, marrow macrophages and the rest, Which immune-evasion routes are actually documented in KMT2A-rearranged disease, and which are borrowed from AML and assumed - a ranked ledger with the bounded absences that define it
19. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    data/symbol-collisions.json - this project's symbol-collision registry, 177 rows at the time of reading

    · Oncofusion knowledge graph 2026
    used by Niche proteins enumerated and NOT written in this tranche, and four symbol traps measured while doing it - three of them new to the registry, SELE (E-selectin) - the vascular-niche chemoresistance receptor with a clinical antagonist, zero KMT2A papers, and a paediatric trial that a company restructuring closed
20. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB gene_exact:LEPR AND organism_id:9606 AND reviewed:true - two reviewed human records, O15243 (LEPROT, 131 aa) returned before P48357 (LEPR, 1165 aa)

    · UniProtKB 2026
    used by Niche proteins enumerated and NOT written in this tranche, and four symbol traps measured while doing it - three of them new to the registry
21. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB Q8N4C6 (NIN_HUMAN), human ninein, 2090 aa

    · UniProt Consortium 2026
    used by NIN (ninein) - a centrosomal scaffold recorded as a DOT1L partner under a UniProt isoform that UniProt itself calls dubious
22. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB Q6KC79 (NIPBL_HUMAN), Nipped-B-like protein, 2804 aa, primary gene name NIPBL

    · UniProt Consortium 2026
    used by NIPBL - the cohesin loader, the cleanest degron result in the whole architecture literature, and a second bounded absence in this disease
23. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    PubMed esearch, db=pubmed, terms '(MLL OR KMT2A) AND NIPBL' (count 13) and 'KMT2A AND "loop extrusion"' (count 0), run 2026-09-20; all 13 titles read from a full efetch record and judged off-topic; both querytranslations read

    · NCBI PubMed 2026
    used by NIPBL - the cohesin loader, the cleanest degron result in the whole architecture literature, and a second bounded absence in this disease
24. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB reviewed human entry Q86WI3 (NLRC5_HUMAN)

    UniProt Consortium · UniProtKB 2026
    used by NLRC5 - the MHC class I transactivator, and the route by which interferon activates class I when KMT2A and KMT2B are both gone
25. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    UniProtKB Q86WI3 (NLRC5_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

    · UniProt Consortium 2026
    used by NLRC5 - the MHC class I transactivator, and the route by which interferon activates class I when KMT2A and KMT2B are both gone
26. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for NLRC5, fetched 2026-09-21

    · Human Protein Atlas 2026
    used by NLRC5 - the MHC class I transactivator, and the route by which interferon activates class I when KMT2A and KMT2B are both gone
27. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    NCBI E-utilities esearch and esummary over PubMed, db=pubmed, retmode=json, 21 queries with querytranslation read back for each

    National Center for Biotechnology Information · NCBI Entrez Programming Utilities 2026
    used by No BRCA1, BARD1, EXO1, DNA2, HP1-gamma or KDM5A has ever been localised at the KMT2A locus, in any cell type: 21 queries, every querytranslation read back, 0 on-topic records - and the absence splits THREE ways by class, which is what decides what it costs to close
28. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UCSC Genome Browser REST API, hg38 refGene track, chr5:2,735,000-2,765,000

    UCSC Genome Browser, api.genome.ucsc.edu · UCSC Genome Browser 2026
    used by No gene-body-scale IRX2 measurement exists in this project - the one scale that could attribute the signal is the one nobody computed, The IRX2 / IRX2-DT geometry re-derived from three annotations: 215 bp and 89.2% under RefSeq, 437 bp and 78.2% under Ensembl, and the verdict does not move, Two-window deconvolution: at a typical shared promoter only 1.8% of the signal can be proved to belong to the gene it is filed under, and at IRX2's one near-positive it is 19.7%
29. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    ENCODE Portal experiment search, assay_title=TF ChIP-seq, status=released, one query per target label (SUV39H1, SUV39H2, SETDB1, CBX5, CBX1, CBX3, EHMT2, TRIM28)

    ENCODE Project Consortium · ENCODE Portal 2026
    used by No H3K9me3 writer and no H3K9me3 reader has ever been localised at the KMT2A locus IN THE PUBLISHED LITERATURE, in any cell type: 20 queries, 5 alias spellings, 0 on-topic records - and CORRECTED 2026-09-22: the SUV39H1 half is no longer unanswerable from archives. ENCODE still holds zero SUV39H1 experiments (verified by three routes with controls), but GEO holds two deposits, and Q114 read both at this locus and found the writer ABSENT, SUV39H1 (O43463, KMT1A) - the H3K9me3 writer: a menin effector whose binding region OVERLAPS the drug pocket at its N-terminus, a tumour suppressor in MLL-AF9 AML, recruited to double-strand breaks by CTCF, and never localised at the KMT2A locus in any cell type
30. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    PubMed E-utilities esearch: (MLL[tiab] OR KMT2A[tiab]) AND HDAC6[tiab] -> count 0, warninglist 'No items found', translation read; and (MLL[tiab] OR KMT2A[tiab]) AND SIRT6[tiab] -> count 1 (PMID 30448059), title-checked. Run 2026-09-20.

    NCBI E-utilities · PubMed 2026
    used by No HDAC-inhibitor result in this disease attributes to an HDAC enzyme - the only per-enzyme evidence is genetic, and it says the enzyme is not fusion-specific
31. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    NCBI esearch via scripts/pubmed_safe.py on 2026-09-24. MTA2[tiab] AND KMT2A[tiab] -> 0; MTA2[tiab] AND MLL[tiab] -> 0; MTA1L1[tiab] -> 4. MBD3[tiab] AND KMT2A[tiab] -> 1 (37686014); MBD3[tiab] AND MLL[tiab] -> 0. GATAD2A[tiab] AND MLL[tiab] -> 0; p66alpha[tiab] AND leukemia[tiab] -> 0. SIN3A[tiab] AND MLL[tiab] -> 2; SIN3A[tiab] AND KMT2A[tiab] -> 2. ARID4B[tiab] AND MLL[tiab] -> 0. Complex level: NuRD[tiab] AND KMT2A[tiab] -> 0; NuRD[tiab] AND MLL[tiab] -> 6 (39062462, 30073149, 29694837, 26714524, 22231400, 17543573). Positive control MEN1[tiab] AND MLL[tiab] -> 55.

    · NCBI PubMed E-utilities 2026
    used by No histone-mark readout has ever been measured after perturbing ANY NuRD or Sin3 subunit in a KMT2A-rearranged cell - the two complexes whose entire job is acetylation have zero acetylation data in this disease
32. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
    PubMed esearch counted-absence query set for IDH inhibitors in KMT2A-rearranged leukaemia, with rule-67 positive and operand controls

    Oncofusion knowledge-graph project, Q81-methylome-axis · NCBI E-utilities esearch, run 2026-09-21 with querytranslation inspected for every query 2026
    used by No IDH inhibitor has ever been reported in KMT2A-rearranged leukaemia: 4 literature records, all reviews, and 1 registry record, an adult master protocol - and the first version of this query returned a false zero
33. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    PubMed E-utilities esearch, db=pubmed, run 2026-09-20

    · NCBI 2026
    used by No KMT2A fusion junction peptide has ever been reported - the one measurement that would prove the protein exists has not been published
34. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
    A Phase 1, Open-Label, Dose-Escalation and Expanded Cohort, Continuous Intravenous Infusion Study of EPZ-5676 in Subjects With Advanced Hematologic Malignancies

    · ClinicalTrials.gov 2026
    used by No menin-inhibitor or DOT1L-inhibitor trial has ever named this fusion - and the DOT1L trials excluded it by construction
35. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
    Pinometostat With Standard Chemotherapy in Treating Patients With Newly Diagnosed Acute Myeloid Leukemia With MLL Gene Rearrangement

    · ClinicalTrials.gov 2026
    used by No menin-inhibitor or DOT1L-inhibitor trial has ever named this fusion - and the DOT1L trials excluded it by construction
36. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
    preprint-exposure: every citation in the graph classified by peer-review status, by venue string and DOI shape

    · Oncofusion knowledge graph (internal analysis) 2026
    used by No preprint in this graph is carried at strong or established - and the audit that found that first misclassified Genes & Development as a preprint
37. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    Bounded PubMed absence for PML nuclear bodies against MLL and KMT2A, run through scripts/pubmed_safe.py with operand controls and title checking of every hit

    · PubMed (NCBI E-utilities), via this project's pubmed_safe wrapper 2026
    used by Nobody has ever looked at whether a KMT2A fusion is at PML nuclear bodies - a counted zero, with the one biochemical hint pointing yes
38. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    Enumeration and title check of all 19 records returned by hDaxx[tiab] AND leukemia[tiab]

    · PubMed (NCBI E-utilities), titles fetched by efetch and read individually 2026
    used by Nobody has ever looked at whether a KMT2A fusion is at PML nuclear bodies - a counted zero, with the one biochemical hint pointing yes
39. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
    Bounded absence query with controls for proximity labelling of KMT2A fusions, PubMed esearch 2026-09-20

    · Oncofusion knowledge graph (internal analysis) 2026
    used by Nobody has ever run proximity labelling on a KMT2A fusion, and the interaction records that look like they did are RBP-atlas preys from HEK293
40. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
    Label-permutation and within-set nulls for the five-gene shortlist of unwritten interactors that are also selective dependencies

    · Oncofusion knowledge graph (internal analysis) 2026
    used by None of the five selective unwritten interactors survives a null, and FANCD2 - the one that carried a story - is the second weakest of them at p = 0.12
41. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
    check-retracted-sources.py: cited PMIDs queried against Europe PMC PUB_TYPE filters, with a spike control on the query shape actually run

    · Oncofusion knowledge graph (internal) 2026
    used by None of this graph's 2,196 cited papers is retracted - and I nearly reported that from a query that could not have found one
42. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    UniProtKB - P46087 (NOP2_HUMAN), 28S rRNA (cytosine(4447)-C(5))-methyltransferase

    · UniProt Knowledgebase 2026
    used by NOP2 / NSUN1 / p120 (P46087) - the 28S rRNA m5C4447 methyltransferase in the MLL-fusion nucleolar block, present in the index fusion's column at p = 0.0223, and the gene whose PubMed token is two-thirds contaminated by its NSUN paralogues
43. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    NCBI PubMed esearch, symbol-contamination measurement and bounded-absence query

    · NCBI E-utilities 2026
    used by NOP2 / NSUN1 / p120 (P46087) - the 28S rRNA m5C4447 methyltransferase in the MLL-fusion nucleolar block, present in the index fusion's column at p = 0.0223, and the gene whose PubMed token is two-thirds contaminated by its NSUN paralogues
44. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    PubMed esearch, three stated queries run 2026-09-19; counts 0, 0 and 1

    · NCBI E-utilities 2026
    used by Not one phosphorylation site has ever been mapped on a KMT2A fusion protein itself - every site in this beat is an inference across the junction
45. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    Europe PMC RESTful search, BODY: IDH-inhibitor name list x KMT2A clause, 2026-09-22

    · Europe PMC 2026
    used by Not re-bounded: an IDH-inhibitor name list crossed with KMT2A terms returns 489 full-text records and 131 even when tightened to the three approved drugs against the exact string KMT2A-rearranged, which is more than this beat could screen - the unrun step, stated with its size
46. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    PubMed esearch, five stated queries run 2026-09-19 with title checking of all returned records

    · NCBI E-utilities 2026
    used by Nothing connects ENL YEATS binding or ENL condensates to phosphorylation: zero records, and what that means for the phase-2 dossier
47. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    Proteins API proteomics-ptm Q03111 and P42568, fetched 2026-09-19; positions compared against the YEATS domain boundaries recorded in this graph

    · EMBL-EBI 2026
    used by Nothing connects ENL YEATS binding or ENL condensates to phosphorylation: zero records, and what that means for the phase-2 dossier, Phosphosite counts for the five 3' partners: the atlas beats UniProt 1.4-6.0x, and the retained/lost split is completely different for each partner
48. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB P06748 (NPM_HUMAN), Nucleophosmin

    · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
    used by NPM1 (nucleophosmin) — the nucleolar shuttle whose mutant form binds chromatin with KMT2A and buys the second menin-inhibitor indication
49. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
    UniProtKB P06748 (NPM1_HUMAN): sequence length, domain, zinc-finger, DNA-binding, repeat, motif and region features, and PDB cross-references

    UniProt Consortium · UniProtKB 2026
    used by NPM1 (nucleophosmin) domain architecture
50. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    RCSB PDB entry 2XP2: Structure of the Human Anaplastic Lymphoma Kinase in Complex with Crizotinib (PF-02341066)

    RCSB Protein Data Bank depositors · Protein Data Bank 2026
    used by NPM1::ALK (NPM-ALK), t(2;5)(p23;q35)
51. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB P55055 (NR1H2_HUMAN) and Q13133 (NR1H3_HUMAN), lengths and primary gene names fetched 2026-09-21

    · UniProtKB 2026
    used by NR1H2 / LXR-beta - the paralogue arm of the menin-LXR contact, present in both species and absent from the paper's conclusion
52. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB P55055 (NR1H2_HUMAN) SUBCELLULAR LOCATION and TISSUE SPECIFICITY comments, fetched 2026-09-21

    UniProt Consortium · UniProtKB 2026
    used by NR1H2 / LXR-beta - the paralogue arm of the menin-LXR contact, present in both species and absent from the paper's conclusion
53. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    UniProtKB Q13133 (NR1H3_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

    · UniProt Consortium 2026
    used by NR1H3 / LXR-alpha - the best-observed of the three MEN1 sterol partners, with both directions curated in two species, and menin as its corepressor
54. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for NR1H3, fetched 2026-09-21

    · Human Protein Atlas 2026
    used by NR1H3 / LXR-alpha - the best-observed of the three MEN1 sterol partners, with both directions curated in two species, and menin as its corepressor
55. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB reviewed human record P04150 (NR3C1, glucocorticoid receptor)

    · UniProt 2026
    used by NR3C1 - the receptor the single most prognostic drug in this disease acts through, and the fusion suppresses it
56. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB P01111 (RASN_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

    · UniProt Consortium 2026
    used by NRAS - the other RAS, the one whose 0.3% clone came back, and a UniProt record that lists HRAS1 as its own synonym
57. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for NRAS, fetched 2026-09-21

    · Human Protein Atlas 2026
    used by NRAS - the other RAS, the one whose 0.3% clone came back, and a UniProt record that lists HRAS1 as its own synonym
58. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB entry Q96L73 (NSD1_HUMAN): reviewed record, length, gene names and subcellular-location comments

    UniProt Consortium · UniProtKB/Swiss-Prot 2026
    used by NSD1 - the H3K36 dimethyltransferase whose catalytic activity a NUP98 fusion cannot do without, and which cells can otherwise lose almost for free
59. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    HGNC record for NSD1 (HGNC:14234)

    HUGO Gene Nomenclature Committee · HGNC REST 2026
    used by NSD1 - the H3K36 dimethyltransferase whose catalytic activity a NUP98 fusion cannot do without, and which cells can otherwise lose almost for free
60. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
    UniProtKB Q96L73 (NSD1_HUMAN): sequence length, domain, zinc-finger, DNA-binding, repeat, motif and region features, and PDB cross-references

    UniProt Consortium · UniProtKB 2026
    used by NSD1 (nuclear receptor-binding SET domain protein 1) domain architecture
61. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
    UniProtKB Q16288 (NTRK3_HUMAN): sequence length, domain, zinc-finger, DNA-binding, repeat, motif and region features, and PDB cross-references

    UniProt Consortium · UniProtKB 2026
    used by NTRK3 (TrkC neurotrophin receptor) domain architecture
62. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41926643
    NUDT21 Drives T-Cell Acute Lymphoblastic Leukemia Through Dual Regulation of Alternative Polyadenylation and Transcriptional Activation.

    Qiu C, Wang J, Li Z, Ji Q, Zhang H, Zhang Q, et al. · Adv Sci (Weinh) 2026
    used by NUDT21 / CFIm25 (O43809) - the CFIm subunit that picks poly(A) sites, in the pan-MLL-fusion core, and a paraspeckle protein
63. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB P35658 (NU214_HUMAN), Nuclear pore complex protein Nup214 / Protein CAN, 2090 aa

    · UniProt 2026
    used by NUP214 / CAN - the cytoplasmic-face nucleoporin that CRM1 docks against, and the 3' partner of four leukaemia fusions
64. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    Ensembl REST lookup for ENSG00000119335 (SET) and ENSG00000126883 (NUP214), GRCh38

    · Ensembl 2026
    used by NUP214 / CAN - the cytoplasmic-face nucleoporin that CRM1 docks against, and the 3' partner of four leukaemia fusions
65. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB P35658 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (one location value and one note, both from the founding paper)

    UniProt Consortium · UniProtKB 2026
    used by NUP214 / CAN - the cytoplasmic-face nucleoporin that CRM1 docks against, and the 3' partner of four leukaemia fusions
66. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB entry P52948 (NUP98_HUMAN): reviewed record, length, gene names and subcellular-location comments

    UniProt Consortium · UniProtKB/Swiss-Prot 2026
    used by NUP98 - the mobile nucleoporin that leaves the pore to put H3K4me3 on haematopoietic promoters, and the shared 5' donor of seven fusions in this graph, NUP98 — domain architecture and residue coordinates (with the 1817/920 numbering trap resolved), NUP98::HOXA9, t(7;11)(p15;p15) — the founding NUP98 fusion, and an incomplete oncogene, NUP98::KMT2A, inv(11)(p15q23) — the NUP98 fusion that contains KMT2A and is menin-inhibitor RESISTANT
67. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    HGNC record for NUP98 (HGNC:8068)

    HUGO Gene Nomenclature Committee · HGNC REST 2026
    used by NUP98 - the mobile nucleoporin that leaves the pore to put H3K4me3 on haematopoietic promoters, and the shared 5' donor of seven fusions in this graph
68. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    InterPro/Pfam/CATH-Gene3D/SUPERFAMILY domain matches for UniProt P52948

    InterPro Consortium; EMBL-EBI · InterPro 2026
    used by NUP98 — domain architecture and residue coordinates (with the 1817/920 numbering trap resolved)
69. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    Pairwise comparison of UniProt isoform sequences P52948-1 (1817 aa) and P52948-4 (920 aa), cross-checked against RefSeq NP_624357 (Nup98 isoform 2 precursor, 920 aa)

    Computed by us from UniProt and NCBI RefSeq sequence records · UniProt / NCBI RefSeq 2026
    used by NUP98 — domain architecture and residue coordinates (with the 1817/920 numbering trap resolved)
70. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    AlphaFold Protein Structure Database entry for P52948 (model AF-P52948-F1)

    Jumper J, Varadi M, et al.; EMBL-EBI/DeepMind · AlphaFold DB 2026
    used by NUP98 — domain architecture and residue coordinates (with the 1817/920 numbering trap resolved)
71. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB P52948 (NUP98_HUMAN) feature table - MOD_RES, CROSSLNK, CARBOHYD

    · UniProt 2026
    used by NUP98 fusions retain a 469-residue FG repeat region that carries ZERO annotated modification sites - the cleanest retained-portion PTM absence in the four fusion families checked
72. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    Europe PMC searches for NUP98 O-GlcNAcylation, 2026-09-24 (89, 27, 13 and 21 hits, none on topic)

    · Europe PMC REST 2026
    used by NUP98 fusions retain a 469-residue FG repeat region that carries ZERO annotated modification sites - the cleanest retained-portion PTM absence in the four fusion families checked
73. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 42047313
    Clinicogenetic characteristics and outcomes of children with NUP98-rearranged acute myeloid leukaemia: A single-centre cohort study of 32 cases

    Zhang ZX, et al. · British Journal of Haematology 2026
    used by NUP98 rearrangements are 7.2% of paediatric AML and 1.6% of adult, with 35% five-year overall survival — and the partner changes everything
74. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB entry P31269, Homeobox protein Hox-A9 (HOXA9)

    UniProt Consortium · UniProtKB 2026
    used by NUP98::HOXA9, t(7;11)(p15;p15) — the founding NUP98 fusion, and an incomplete oncogene
75. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 42120929
    Comprehensive mutational profiling and clinical outcome of adult acute myeloid leukemia patients with NUP98 rearrangement

    Ducourneau B, et al. · Leukemia 40(7):1494-1501 2026
    used by NUP98::HOXA9, t(7;11)(p15;p15) — the founding NUP98 fusion, and an incomplete oncogene, NUP98::NSD1 without FLT3-ITD loses its prognostic significance — the confound that undermines the fusion's own reputation
76. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB entry P11387, DNA topoisomerase 1 (TOP1)

    UniProt Consortium · UniProtKB 2026
    used by NUP98::TOP1, t(11;20)(p15;q11~q12) — the most precisely mapped NUP98 breakpoint, and the reciprocal that was wrong twice, TOP1 domain architecture (UniProt P11387)
77. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB Q86Y26 (NUTM1_HUMAN) - NUT family member 1, Homo sapiens

    · UniProtKB 2026
    used by NUTM1 (NUT) - a testis-restricted, fully disordered protein that writes no mark and instead switches ON a writer, and the 3' partner of the KMT2A-wild-type infant B-ALL with the opposite prognosis
78. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    IntAct molecular interaction database, REST findInteractions for Q86Y26 and for interactor EBI-8101903

    · IntAct (EMBL-EBI) 2026
    used by NUTM1 (NUT) - a testis-restricted, fully disordered protein that writes no mark and instead switches ON a writer, and the 3' partner of the KMT2A-wild-type infant B-ALL with the opposite prognosis
79. weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paperPMID 41878999
    A Rare Case of Infant B-ALL With SIN3A::NUTM1 Fusion: Durable Molecular Remission Achieved With Standard Acute Lymphoblastic Leukemia Chemotherapy

    Peng D, Luo M, Lin Z, et al. · Pediatric Blood & Cancer 2026
    used by NUTM1-rearranged B-ALL - the KMT2A-wild-type arm of the same infant trials, with 4-year EFS of 100% against 36-37% next door, SIN3A (Q96ST3) - the scaffold of the OTHER HDAC1/2 corepressor complex, a 5' fusion partner in infant B-ALL and in myelodysplasia, and present in three MLL-fusion AP-MS columns but not the index fusion's
80. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    analyses/q36-menin-residue-axis/run.py - per-chain-pair ligand contact footprints on RCSB 7UJ4, 8VA6, 9WKU and 8IG0, run 2026-09-22

    · RCSB PDB, computed 2026
    used by Occupancy-matched three-arm comparison of ziftomenib, revumenib and bleximenib on gamma-H2A.X resolution and HR efficiency in a KMT2A-rearranged B-lineage line, with the peptide competition as the biochemical pre-test, The menin small-molecule pocket as a coordinated surface: four clinical-stage compounds, 21-26 contact residues each, and bleximenib sitting on a different part of it from the other three, The menin-inhibitor class splits at Val371: three of four clinical compounds contact the first residue of the gamma-H2A.X motif and bleximenib does not, so DNA-repair effects should separate from KMT2A-displacement effects within the class
81. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    PubMed E-utilities esearch via scripts/pubmed_safe.py with query translations read back, 2026-09-22

    · NCBI PubMed 2026
    used by Occupancy-matched three-arm comparison of ziftomenib, revumenib and bleximenib on gamma-H2A.X resolution and HR efficiency in a KMT2A-rearranged B-lineage line, with the peptide competition as the biochemical pre-test
82. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
    DepMap CRISPR gene effect via Open Targets, gnomAD GRCh38 loss-of-function constraint, and Human Protein Atlas RNA tissue distribution; joined analysis in analyses/target-safety-triage/

    Broad Institute DepMap; gnomAD; Human Protein Atlas; Open Targets Platform; analysis run by this project · Databases and derived analysis 2026
    used by Of the menin co-dependencies, which are actually safe to hit? (run), Phase-2 dossier: the LEDGF/PSIP1 IBD - the best therapeutic index in the interactome, a solved structure, two hotspot residues, and three compounds in the world
83. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    analyses/q203-cofolding-the-retained-interfaces/score.py and scores.json, run 2026-10-12 to 2026-10-14

    · this repository 2026
    used by Of three co-folding predictors run without alignments, only the single-sequence one (ESMFold2) recovers menin-MBM1 above shuffled-peptide controls, and its real peptides sit within 0.6 A (core) to 2 A of the deposited structures while shuffled peptides do not - on interfaces that were in its training data
84. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.web
    ESMFold2 model card (biohub/ESMFold2)

    · Hugging Face 2026
    used by Of three co-folding predictors run without alignments, only the single-sequence one (ESMFold2) recovers menin-MBM1 above shuffled-peptide controls, and its real peptides sit within 0.6 A (core) to 2 A of the deposited structures while shuffled peptides do not - on interfaces that were in its training data
85. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    IntAct molecular interaction database (EMBL-EBI), PSICQUIC tab27 record for id:O15294 AND id:Q03164

    2026
    used by OGT - the O-GlcNAc transferase, a Polycomb-complexome bait that pulls down KMT2A, and already the metabolic dial on DOT1L in this graph
86. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    mechanism-dot1l-oglcnacylation-stability (this graph) - see that entry for the primary citation

    2026
    used by OGT - the O-GlcNAc transferase, a Polycomb-complexome bait that pulls down KMT2A, and already the metabolic dial on DOT1L in this graph
87. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    UniProtKB O15294 (OGT_HUMAN)

    2026
    used by OGT - the O-GlcNAc transferase, a Polycomb-complexome bait that pulls down KMT2A, and already the metabolic dial on DOT1L in this graph
88. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    Cellosaurus record CVCL_0093 (RS4;11)

    Bairoch A · Cellosaurus, SIB 2026
    used by On an absolute scale, H3K79me2 in the index fusion falls 52-fold (EPZ5676) to 91-fold (compound 10) by day 9 and NOT ONE of 17,341 measurable genes holds its signal - the first absolute-scale chromatin statement this project has been able to make about KMT2A::AFF1, RS4;11 retains 2.5x more H3K79me2 than MOLM13 under EPZ5676 at every one of ten baseline deciles - and 0.71x under compound 10, with the sign reversing in deciles 5-10: the between-line difference survives the baseline confound and fails the drug replication, so it cannot be attributed to the fusion partner, The deposited RS4;11 H3K79me2 table in GSE149183 is two-thirds MOLM13: 38,580 of its 57,794 rows are byte-for-byte identical to the MOLM13 file, so the index fusion is measurable at only 19,214 TSS entries, The eleven-line panel is nine patients, seven of them myeloid, and none an infant
89. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    PubMed E-utilities esearch, live queries with querytranslation recorded

    · NCBI Entrez 2026
    used by One domain, two names - and `MLL[tiab]` does not match `MLL1`. Two token traps, two missing affinity papers each
90. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    PubMed E-utilities esearch, live queries with querytranslation recorded and UID spike tests

    · NCBI Entrez 2026
    used by One domain, two names - and `MLL[tiab]` does not match `MLL1`. Two token traps, two missing affinity papers each
91. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
    Laboratory-level source independence in the Oncofusion knowledge graph: deflation, reach, concentration and self-opposition

    · Oncofusion knowledge graph (internal analysis) 2026
    used by One laboratory holds up 92 entries, which counting papers could not show
92. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
    Load-bearing citation audit of the Oncofusion knowledge graph: reach, concentration, sole support and confidence spread

    · Oncofusion knowledge graph (internal analysis) 2026
    used by One paper holds up 70 entries, and one dataset holds up 167
93. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
    Full-text and abstract term census over PMC5549451 (PMID 28644774) and PMID 28644774/34048275, run 2026-09-21: KMT2A/MLL/11q23 = 0 in the genotype paper against rs12459419 = 42 in the same text; rs12459419/splic/genotype = 0 in the KMT2A-r paper.

    · PubMed Central and PubMed E-utilities (term census; counts recorded in data) 2026
    used by One trial, two published analyses, never crossed: gemtuzumab clearly helps KMT2A-rearranged children, and a common CD33 genotype abolishes gemtuzumab benefit — nobody has asked whether it abolishes it in them
94. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
    PRIDE Archive REST API v3, project keyword search, queried 2026-09-21: keywords 'MLL-AF4' (3 projects), 'KMT2A' (20), 'MV4;11' (13), 'MLL rearranged leukemia' (5), 'RS4;11' (2).

    · PRIDE Archive / ProteomeXchange (EMBL-EBI) 2026
    used by Only 2 of this graph's 26 bounded absences name a deposited dataset — the corpus asks whether a paper exists and almost never whether the data does
95. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
    Narrowed literature recall: a three-tier query ladder over the four queries the DOT1L correction did not re-check

    · Oncofusion knowledge graph (internal analysis) 2026
    used by Only one query was broken, and it was dragging the whole figure down
96. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry
    ClinicalTrials.gov API v2 search, condition 'KMT2A rearranged acute leukemia OR infant acute lymphoblastic leukemia', status RECRUITING or NOT_YET_RECRUITING, retrieved 2026-09-18

    US National Library of Medicine, ClinicalTrials.gov · ClinicalTrials.gov 2026
    used by Open trial map for KMT2A-rearranged and infant acute leukemia, September 2026 — what a clinician can actually enrol into today
97. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry
    ClinicalTrials.gov API v2 search, intervention 'revumenib OR ziftomenib OR bleximenib OR enzomenib OR icovamenib OR menin inhibitor', retrieved 2026-09-18

    US National Library of Medicine, ClinicalTrials.gov · ClinicalTrials.gov 2026
    used by Open trial map for KMT2A-rearranged and infant acute leukemia, September 2026 — what a clinician can actually enrol into today
98. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    Designating an Orphan Product: Drugs and Biological Products

    US Food and Drug Administration, Office of Orphan Products Development · FDA 2026
    used by Orphan drug designation: what it actually gives a KMT2A programme, and what it bends
99. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
    21 U.S.C. 360cc - Protection for drugs for rare diseases or conditions

    United States Code · Office of the Law Revision Counsel, US House of Representatives 2026
    used by Orphan drug designation: what it actually gives a KMT2A programme, and what it bends
100. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Patent and Exclusivity Information, NDA 218944 (REVUFORJ, revumenib citrate)

     US Food and Drug Administration, Orange Book · Approved Drug Products with Therapeutic Equivalence Evaluations 2026
     used by Orphan drug designation: what it actually gives a KMT2A programme, and what it bends
101. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Patent and Exclusivity Information, NDA 220305 (KOMZIFTI, ziftomenib)

     US Food and Drug Administration, Orange Book · Approved Drug Products with Therapeutic Equivalence Evaluations 2026
     used by Orphan drug designation: what it actually gives a KMT2A programme, and what it bends
102. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.news
     Congress Extends Rare Pediatric Disease Priority Review Voucher Program Through 2029

     Spencer Fane LLP · Spencer Fane client insight 2026
     used by Orphan drug designation: what it actually gives a KMT2A programme, and what it bends, The rare paediatric disease priority review voucher: a $150-200m bearer bond, lapsed for 14 months, now live to 2029
103. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     ENCODE Portal experiment and target searches, type=Experiment&status=released&target.label=<SYMBOL> with no assay filter, and /targets/?label=<SYMBOL>

     ENCODE Project Consortium · ENCODE Portal 2026
     used by Outside ENCODE the resection layer is not one absence but four: 31 BRCA1, 45 CBX3, 22 KDM5A, 14 MRE11 and 8 CtIP ChIP-seq experiments exist in ChIP-Atlas, DNA2 and RAD50 have one each, and only BARD1 and EXO1 have none in any of the three archives
104. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB entries 3BLQ, 3BLR, 3MY1, 3TN8, 3TNH, 3LQ5, 4BCF, 4BCG, 4BCH, 4BCI, 4BCJ, 6GZH, 6W9E, 6Z45, 7NWK, 4IMY, 4OR5, 4OGR, 5L1Z and 6CYT, with per-entity UniProt accessions, ligands, methods, resolutions and modelled residue ranges resolved against the RCSB APIs

     RCSB Protein Data Bank · RCSB PDB 2026
     used by P-TEFb (CDK9 + cyclin T1) — the deepest dependency and the clearest therapeutic-index wall in the interactome
105. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL database version 34, targets CHEMBL3116 (cyclin-dependent kinase 9) and CHEMBL2108 (cyclin-T1): target_search, get_bioactivity and get_mechanism queries

     EMBL-EBI ChEMBL · ChEMBL 2026
     used by P-TEFb (CDK9 + cyclin T1) — the deepest dependency and the clearest therapeutic-index wall in the interactome
106. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     ana-q142-one-patients-two-lineages-share-no-more-fusion-target-promoters-than-two-different-patients-do (this graph)

     Iqbal Lab oncofusion knowledge graph, beat Q152 · this project 2026
     used by p279 and p318 are one patient in two lineage states, and collapsing them removes exactly one gene - NOP58 - from the recurrent core
107. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB - Q9UQ80 (PA2G4_HUMAN), Proliferation-associated protein 2G4

     · UniProt Knowledgebase 2026
     used by PA2G4 / EBP1 (Q9UQ80) - the best-evidenced member of the MLL-fusion nucleolar block in the index fusion's own column (p = 0.014), and the only one with a published AML phenotype
108. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     NCBI PubMed esearch, bounded-absence query (counted negative)

     · NCBI E-utilities 2026
     used by PA2G4 / EBP1 (Q9UQ80) - the best-evidenced member of the MLL-fusion nucleolar block in the index fusion's own column (p = 0.014), and the only one with a published AML phenotype
109. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     q147 analyze147.py locus list and a repository-wide check of promoter-scale occupancy claims

     Iqbal Lab oncofusion knowledge graph, beats Q147 and Q153 · this project 2026
     used by PAF1 and MED29 return the identical occupancy call in all 27 GSE252378 samples, because 86.5% of their promoter windows is the same DNA
110. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8N7H5 (PAF1_HUMAN) - subcellular location with evidence codes, fetched 2026-09-20

     UniProt Consortium · UniProtKB 2026
     used by PAF1 is annotated as punctate everywhere in the nucleus EXCEPT the nucleolus - and a 2024 proximity map puts it in the nucleolus
111. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     DepMap CRISPR gene-effect (Chronos) for PALB2, served by the Open Targets Platform GraphQL API

     Broad Institute DepMap / Open Targets Platform · Open Targets Platform 26.06 (meta.dataVersion year 26 month 06), fetched 2026-09-22 2026
     used by PALB2 (Q86YC2, 1,186 aa) - the physical link that makes BRCA1 and BRCA2 one pathway, named in 4 entries and never written; DepMap mean -0.339, statistically indistinguishable from CBX1 and MEN1, and its entire KMT2A intersection is one Ewing sarcoma germline paper
112. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     paralogue-redundancy-excuse/run.py: does paralogue membership predict a weak DepMap signal across this graph's scored genes?

     · Oncofusion knowledge graph (internal analysis) 2026
     used by Paralogue redundancy does NOT explain anti-selective DepMap results here: clearly selective genes carry paralogues at 30.6% against a 32.7% base rate, with no depletion
113. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     acetyltransferase-paralogue-split: within HBO1/MOZ families, KMT2A-selectivity is carried by one member while its paralogues rank mid-genome

     · Oncofusion knowledge graph (internal analysis) 2026
     used by Paralogue redundancy does NOT explain anti-selective DepMap results here: clearly selective genes carry paralogues at 30.6% against a 32.7% base rate, with no depletion
114. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P09874 (PARP1_HUMAN), Poly [ADP-ribose] polymerase 1, 1014 aa, primary gene name PARP1

     · UniProt Consortium 2026
     used by PARP1 - the target whose inhibitors this disease was specifically shown to resist, and which three later papers combine anyway
115. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8N5Y8 (PAR16_HUMAN), human PARP16, 322 aa

     · UniProt Consortium 2026
     used by PARP16 - graded A- in the DOT1L pulldown, the best of the six, and a tail-anchored ER membrane protein that cannot be in the nucleoplasm
116. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed human entry Q02548 (PAX5_HUMAN), retrieved by gene_exact:PAX5 with fields accession, gene_primary, gene_synonym, protein_name, length and the DNA-binding feature table

     UniProt Consortium · UniProt 2026
     used by PAX5 - the most frequently hit non-KMT2A gene in infant KMT2A-rearranged ALL, and it is a deletion in one allele, not a mutation
117. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     NCBI PubMed E-utilities esearch with querytranslation inspected, and UniProtKB reviewed human record Q6ZW49 (PAXIP1); queries run 2026-09-20 by this project

     NCBI PubMed; UniProt Consortium; queries run by this project · Database 2026
     used by PAXIP1 (PTIP) - the subunit that tells an MLL3/4 complex apart from a menin complex, and the only one with a leukaemic-niche phenotype
118. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P40424 (PBX1_HUMAN): reviewed sequence, length and positional feature annotations

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by PBX1 — domain architecture and residue coordinates
119. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     InterPro/Pfam/CATH-Gene3D/SUPERFAMILY domain matches for UniProt P40424

     InterPro Consortium; EMBL-EBI · InterPro 2026
     used by PBX1 — domain architecture and residue coordinates
120. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     AlphaFold Protein Structure Database entry for P40424 (model AF-P40424-F1)

     Jumper J, Varadi M, et al.; EMBL-EBI/DeepMind · AlphaFold DB 2026
     used by PBX1 — domain architecture and residue coordinates
121. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     IntAct molecular interaction database, PSICQUIC MITAB 2.7, query id:P40425 AND id:Q9UHB7, 2 rows retrieved 2026-09-21

     · IntAct (EMBL-EBI) 2026
     used by PBX2 - an AFF4 co-purifying protein in two rounds of one AP-MS atlas, and the PBX paralogue the leukaemia literature explicitly excluded
122. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed counted-absence searches for PBX2 against AFF4, MLL and KMT2A, run 2026-09-21

     · PubMed (NCBI) 2026
     used by PBX2 - an AFF4 co-purifying protein in two rounds of one AP-MS atlas, and the PBX paralogue the leukaemia literature explicitly excluded
123. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries P40424 (PBX1_HUMAN), P40425 (PBX2_HUMAN) and P40426 (PBX3_HUMAN)

     · UniProtKB/Swiss-Prot 2026
     used by PBX2 - an AFF4 co-purifying protein in two rounds of one AP-MS atlas, and the PBX paralogue the leukaemia literature explicitly excluded, The PBX requirement in KMT2A-rearranged leukaemia is a family property, and every experiment that shaped this graph's view of it removed one paralogue at a time
124. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41882099
     A Perturb-seq map of a differentiation hub reveals synergistic vulnerabilities in KMT2A-rearranged acute myeloid leukemia

     Aryal S, Curtiss BM, Zhou X, Lu R, Cheng C · Leukemia 40(5):996-1008 2026
     used by PCGF1 — the PRC1.1 scaffold whose LOSS causes menin-inhibitor resistance, so it is a restraint on the leukaemia rather than a dependency of it, PRC1.1 — the complex this graph has been leaning on, now enumerated subunit by subunit, and the enumeration changes the conclusion
125. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9BSM1 (PCGF1_HUMAN), Polycomb group RING finger protein 1

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by PCGF1 — the PRC1.1 scaffold whose LOSS causes menin-inhibitor resistance, so it is a restraint on the leukaemia rather than a dependency of it
126. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI PubMed esearch, bounded-absence queries with querytranslation and warninglist recorded

     · NCBI E-utilities 2026
     used by PES1 / pescadillo (O00541) - the missing PeBoW subunit of this graph's nucleolar block, pan-essential in 96.9% of DepMap lines, and the one myeloid experiment on it shows proliferation WITHOUT a differentiation block
127. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB data API entry records for 3I91 and 5EQ0

     · RCSB PDB 2026
     used by Phase-2 dossier: CBX8 - the candidate this review would kill, and the reasons are all numeric
128. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     GTEx Portal API v2, medianGeneExpression for CBX8, datasetId gtex_v8

     · GTEx Consortium 2026
     used by Phase-2 dossier: CBX8 - the candidate this review would kill, and the reasons are all numeric
129. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     gnomAD GraphQL API, gnomad_constraint for CBX8

     · Broad Institute 2026
     used by Phase-2 dossier: CBX8 - the candidate this review would kill, and the reasons are all numeric
130. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL REST API, activities for target CHEMBL3232684 with pchembl_value >= 6

     · EMBL-EBI ChEMBL 2026
     used by Phase-2 dossier: CBX8 - the candidate this review would kill, and the reasons are all numeric
131. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB data API entry record for 4HRA (Crystal Structure of DOT1L in Complex with EPZ-5676)

     · RCSB PDB 2026
     used by Phase-2 dossier: DOT1L - perfect chemistry, a completed paediatric trial, and a target that has already had its answer
132. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     GTEx Portal API v2, medianGeneExpression for DOT1L, datasetId gtex_v8

     · GTEx Consortium 2026
     used by Phase-2 dossier: DOT1L - perfect chemistry, a completed paediatric trial, and a target that has already had its answer
133. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     gnomAD GraphQL API, gnomad_constraint for DOT1L

     · Broad Institute 2026
     used by Phase-2 dossier: DOT1L - perfect chemistry, a completed paediatric trial, and a target that has already had its answer
134. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL REST API, activities for target CHEMBL1795117 with pchembl_value >= 6

     · EMBL-EBI ChEMBL 2026
     used by Phase-2 dossier: DOT1L - perfect chemistry, a completed paediatric trial, and a target that has already had its answer
135. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB data API entry record for 6MAJ (HBO1 is required for the maintenance of leukaemia stem cells)

     · RCSB PDB 2026
     used by Phase-2 dossier: KAT7/HBO1 - the only enzyme here with a real active site, a co-crystal inhibitor and a co-dependency that survives lineage adjustment almost intact
136. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     GTEx Portal API v2, medianGeneExpression for KAT7, datasetId gtex_v8

     · GTEx Consortium 2026
     used by Phase-2 dossier: KAT7/HBO1 - the only enzyme here with a real active site, a co-crystal inhibitor and a co-dependency that survives lineage adjustment almost intact
137. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     gnomAD GraphQL API, gnomad_constraint for KAT7

     · Broad Institute 2026
     used by Phase-2 dossier: KAT7/HBO1 - the only enzyme here with a real active site, a co-crystal inhibitor and a co-dependency that survives lineage adjustment almost intact
138. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL REST API, activities for target CHEMBL3774299 with pchembl_value >= 6

     · EMBL-EBI ChEMBL 2026
     used by Phase-2 dossier: KAT7/HBO1 - the only enzyme here with a real active site, a co-crystal inhibitor and a co-dependency that survives lineage adjustment almost intact
139. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB data API entry records for 2J2S and 4NW3

     · RCSB PDB 2026
     used by Phase-2 dossier: the CXXC-DNA interface - two structures, a genetically validated requirement, and a selectivity problem that decides everything
140. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     ChEMBL REST API, activity counts for targets CHEMBL1293299 and CHEMBL2093861 with pchembl_value >= 6

     · EMBL-EBI ChEMBL 2026
     used by Phase-2 dossier: the CXXC-DNA interface - two structures, a genetically validated requirement, and a selectivity problem that decides everything
141. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB data API entry records for 5J9S and 6HT1

     · RCSB PDB 2026
     used by Phase-2 dossier: the ENL YEATS pocket - two co-crystals, nanomolar chemistry, and a co-dependency that only appears inside leukaemia
142. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     GTEx Portal API v2, medianGeneExpression for MLLT1, datasetId gtex_v8

     · GTEx Consortium 2026
     used by Phase-2 dossier: the ENL YEATS pocket - two co-crystals, nanomolar chemistry, and a co-dependency that only appears inside leukaemia
143. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     gnomAD GraphQL API, gnomad_constraint for MLLT1

     · Broad Institute 2026
     used by Phase-2 dossier: the ENL YEATS pocket - two co-crystals, nanomolar chemistry, and a co-dependency that only appears inside leukaemia
144. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL REST API, activities for target CHEMBL4295801 with pchembl_value >= 6

     · EMBL-EBI ChEMBL 2026
     used by Phase-2 dossier: the ENL YEATS pocket - two co-crystals, nanomolar chemistry, and a co-dependency that only appears inside leukaemia
145. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     gnomAD GraphQL API, gnomad_constraint for AFF1

     · Broad Institute 2026
     used by Phase-2 dossier: the fusion junction - a negative feasibility finding, with the one surviving modality named
146. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     GTEx Portal API v2, medianGeneExpression for PSIP1, datasetId gtex_v8

     · GTEx Consortium 2026
     used by Phase-2 dossier: the LEDGF/PSIP1 IBD - the best therapeutic index in the interactome, a solved structure, two hotspot residues, and three compounds in the world
147. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL REST API, activities for target CHEMBL3988590 with pchembl_value >= 6

     · EMBL-EBI ChEMBL 2026
     used by Phase-2 dossier: the LEDGF/PSIP1 IBD - the best therapeutic index in the interactome, a solved structure, two hotspot residues, and three compounds in the world
148. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     gnomAD GraphQL API, gnomad_constraint for MEN1 (ENSG00000133895)

     · Broad Institute 2026
     used by Phase-2 dossier: the menin MBM1 pocket - the only candidate here that is already de-risked, and therefore the only one whose gate is clinical
149. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB data API entry records for 7UJ4 and 6HT1

     · RCSB PDB 2026
     used by Phase-2 dossier: the non-small-molecule modalities - degraders first, and the honest limits of each
150. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry O43189 (PHF1_HUMAN): reviewed record, length, gene names and subcellular-location comments

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by PHF1 (PCL1) - the PRC2 accessory subunit whose Tudor domain reads H3K36me3, and the 3' acceptor shared by three different NuA4-donor fusions in two different tumour types
151. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     HGNC record for PHF1 (HGNC:8919)

     HUGO Gene Nomenclature Committee · HGNC REST 2026
     used by PHF1 (PCL1) - the PRC2 accessory subunit whose Tudor domain reads H3K36me3, and the 3' acceptor shared by three different NuA4-donor fusions in two different tumour types
152. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry O43175 (SERA_HUMAN), D-3-phosphoglycerate dehydrogenase

     UniProt Consortium · UniProt 2026
     used by PHGDH (O43175) - the largest menin-SELECTIVE fall in the whole serine panel, and the least essential gene in it
153. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
     A Randomized Controlled Trial of Serine/Glycine-Restricted Diet Combined With Chemoradiotherapy and Immunotherapy for Neoadjuvant Treatment of Rectal Cancer

     · ClinicalTrials.gov 2026
     used by PHGDH (O43175) - the largest menin-SELECTIVE fall in the whole serine panel, and the least essential gene in it
154. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI E-utilities esearch against PubMed, 2026-09-22, fusion operand set (KMT2A[tiab] OR MLL[tiab] OR "MLL-AF4"[tiab] OR "KMT2A::AFF1"[tiab] OR "MLL-rearranged"[tiab] OR "t(4;11)"[tiab] OR "MLL-AF9"[tiab] OR "MLL-ENL"[tiab]) with every cytogenetic term quoted and every term separately tagged

     · PubMed 2026
     used by PHGDH (O43175) - the largest menin-SELECTIVE fall in the whole serine panel, and the least essential gene in it
155. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries Q03111, P42568, P55197, P55199, O00472, Q96JC9, Q96CJ1: primary gene name, sequence length and Modified residue features

     · UniProt Consortium 2026
     used by Phosphosite counts for the five 3' partners: the atlas beats UniProt 1.4-6.0x, and the retained/lost split is completely different for each partner, The KMT2A::MLLT1 chimeric protein: exactly what survives the join on each side, in residue numbers
156. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     PhosphoSitePlus homeAction.action - HTTP 403 Forbidden (positive control), 2026-09-22

     · PhosphoSitePlus, Cell Signaling Technology 2026
     used by PhosphoSitePlus returns HTTP 403 to its own homepage from this environment, so the PTM-database absence named by the cycle-65 census cannot be bounded there - and the positive control is what proves the 403 is a block and not an answer
157. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     PRIDE Archive API v2 project search, keyword=MLL-AF4 (returns PXD043920)

     · EMBL-EBI PRIDE Archive 2026
     used by PhosphoSitePlus returns HTTP 403 to its own homepage from this environment, so the PTM-database absence named by the cycle-65 census cannot be bounded there - and the positive control is what proves the 403 is a block and not an answer
158. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Europe PMC RESTful search, BODY: full-text field, MLL fusion names x phospho vocabulary, 2026-09-22

     · Europe PMC 2026
     used by PhosphoSitePlus returns HTTP 403 to its own homepage from this environment, so the PTM-database absence named by the cycle-65 census cannot be bounded there - and the positive control is what proves the 403 is a block and not an answer
159. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q13492 (PICAL_HUMAN) - Phosphatidylinositol-binding clathrin assembly protein

     · UniProt Consortium 2026
     used by PICALM (CALM) - the clathrin-assembly adaptor that is the 5' partner of CALM::AF10, PICALM / CALM - domain architecture, residue coordinates, and which parts the fusion keeps, Where wild-type PICALM actually is: coated pits, the Golgi, coated vesicles - and a nuclear pool nobody has explained
160. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Ensembl REST lookup ENSG00000073921 (PICALM), GRCh38

     · Ensembl 2026
     used by PICALM (CALM) - the clathrin-assembly adaptor that is the 5' partner of CALM::AF10
161. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB free-text and accession-scoped searches for CALM/PICALM, and entry records 3ZYK, 3ZYM, 1HFA, 7JXV

     · RCSB PDB 2026
     used by PICALM (CALM) - the clathrin-assembly adaptor that is the 5' partner of CALM::AF10, PICALM / CALM - domain architecture, residue coordinates, and which parts the fusion keeps
162. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     InterPro/Pfam matches on UniProt Q13492: ANTH domain IPR011417 / PF07651 at residues 23-283

     · InterPro / EMBL-EBI 2026
     used by PICALM / CALM - domain architecture, residue coordinates, and which parts the fusion keeps
163. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed human entry P27986 (P85A_HUMAN), retrieved by gene_exact:PIK3R1 with fields accession, gene_primary, gene_synonym, protein_name, length and the domain feature table

     UniProt Consortium · UniProt 2026
     used by PIK3R1 in KMT2A-rearranged infant ALL - the only cooperating lesion in this disease that was functionally proved to be activating, and it is an ITD
164. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P48426 (PI42A_HUMAN) features, catalytic activity, subcellular location with evidence codes and tissue specificity; HGNC:8997 prev_symbol and prev_name via rest.genenames.org; retrieved 2026-09-22

     · UniProt Consortium / HGNC 2026
     used by PIP4K2A - the lipid kinase at a genome-wide childhood-ALL susceptibility locus, whose product is the ligand a PHD finger reads
165. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     IntAct REST findInteractions/P48426 (320 indexed interactions), filtered to Q03164/O00255/O75475/Q9UHB7/Q8TEK3/P51825/Q03111/P42568/P55199/P55197, with experimentalRole, detectionMethod, hostOrganism and publicationPubmedIdentifier read per row; retrieved 2026-09-22

     · IntAct molecular interaction database (EMBL-EBI) 2026
     used by PIP4K2A - the lipid kinase at a genome-wide childhood-ALL susceptibility locus, whose product is the ligand a PHD finger reads
166. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     analyses/depmap-kmt2a-selective-scan-full/results.json - 17,728 protein-coding genes, 11 Cellosaurus-annotated KMT2A-rearranged versus 43 other leukaemia lines, computed 2026-09-19

     · DepMap (Broad Institute), via this repository's scan 2026
     used by PIP4K2A - the lipid kinase at a genome-wide childhood-ALL susceptibility locus, whose product is the ligand a PHD finger reads
167. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI E-utilities esearch, db=pubmed, querytranslation read back from every response, run 2026-09-22

     · NCBI PubMed 2026
     used by PIP4K2A - the lipid kinase at a genome-wide childhood-ALL susceptibility locus, whose product is the ligand a PHD finger reads
168. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct molecular interaction database, REST findInteractions free-text query on 39251607, all 19,380 rows paginated and role-counted 2026-09-23 by Q129

     · IntAct (EMBL-EBI) 2026
     used by PLANNED: run the size-matched promiscuity correction for PSIP1 in the K562 BioID2 map, the way it was run for KMT2A - fully specified here so the next cycle executes rather than redesigns it, PSIP1-UTP3 is one spoke-expanded BioID2 proximity row in K562 with UTP3 as bait and PSIP1 as prey at miscore 0.27; PSIP1 is recovered by 9 of the 50 baits in that screen and only 2 of the 9 are nucleolar, which is the panel's own base rate - declined, The ABT1/KRR1/UTP3 nucleolar triple: pre-registered, then adjudicated - it is not three proteins on two baits but a seven-member ribosomal block in one screen plus one hit in a neighbourhood that shows no nucleolar enrichment, and two of the four pre-registered discriminators were void exactly as the pre-registration predicted, The size correction that dissolved KMT2A's proximity promiscuity should run the OTHER WAY for PSIP1: at 530 aa, recovery by 9 of 50 BioID2 baits may be MORE than its length predicts, which would make PSIP1 rather than KMT2A the promiscuous prey in this project's core pair
169. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Mozobil (plerixafor) injection - US prescribing information, sections 1 (Indications and Usage) and 8.4 (Pediatric Use), retrieved via openFDA drug/label; NDA022311, Sanofi-Aventis U.S. LLC

     · openFDA / US FDA 2026
     used by Plerixafor (AMD3100, Mozobil) - a licensed CXCR4 antagonist that has never been given to an infant with leukaemia, and whose only paediatric efficacy study failed
170. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry
     A Combined Study in Pediatric Cancer Patients for Dose Ranging and Efficacy/Safety of Plerixafor Plus Standard Regimens for Mobilization Versus Standard Regimens Alone

     · ClinicalTrials.gov 2026
     used by Plerixafor (AMD3100, Mozobil) - a licensed CXCR4 antagonist that has never been given to an infant with leukaemia, and whose only paediatric efficacy study failed
171. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P29590 (PML_HUMAN): reviewed sequence, length and positional feature annotations

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by PML — domain architecture and residue coordinates
172. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     AlphaFold Protein Structure Database entry for P29590 (model AF-P29590-F1)

     Jumper J, Varadi M, et al.; EMBL-EBI/DeepMind · AlphaFold DB 2026
     used by PML — domain architecture and residue coordinates
173. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Ensembl transcript ENST00000268058 (PML): exon structure and coding-sequence coordinates, used to convert exon breakpoints into residue numbers

     Ensembl; EMBL-EBI · Ensembl REST API 2026
     used by PML — domain architecture and residue coordinates
174. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     InterPro member-database domain matches for UniProt P29590 (Pfam, PROSITE, SMART, CDD, CATH-Gene3D, SUPERFAMILY)

     InterPro Consortium; EMBL-EBI · InterPro 2026
     used by PML — domain architecture and residue coordinates
175. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P29590 (PML_HUMAN) canonical sequence

     · UniProt 2026
     used by PML K160 SUMOylation - the retained-portion PTM that became an approved therapy, and the benchmark every other retained site should be measured against
176. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q16633 - POU domain class 2-associating factor 1 (Homo sapiens), reviewed

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by POU2AF1 (OCA-B / OBF-1 / BOB-1) - the B-cell-restricted OCT coactivator, in the Q152 AFF1-only core
177. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     QuickGO annotation service - human protein annotations to GO:0030183 (B cell differentiation) and GO:0030098 (lymphocyte differentiation), including descendant terms

     EMBL-EBI Gene Ontology Annotation · EMBL-EBI 2026
     used by POU2AF1 (OCA-B / OBF-1 / BOB-1) - the B-cell-restricted OCT coactivator, in the Q152 AFF1-only core, The lymphoid look of the core does not survive an activity-matched null inside the same 332, and the pre-registered power check says 17 genes cannot carry the statistic anyway
178. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8WUA2 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (a single nucleus value on a bare ECO:0000305 curator inference)

     UniProt Consortium · UniProtKB 2026
     used by PPIL4 - a spliceosomal cyclophilin whose BioID neighbourhood is the widest overlap with this graph's subject matter of any bait
179. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB - O75807, PR15A_HUMAN, protein phosphatase 1 regulatory subunit 15A

     · UniProt Knowledgebase 2026
     used by PPP1R15A / GADD34 (O75807) - the eIF2alpha phosphatase adaptor, and KMT2A fusion proteins were shown to BIND it and block its apoptosis in 1999, a paper this graph has never cited
180. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed E-utilities esearch: PRDM9 against KMT2A / MLL-rearranged / mixed lineage leukemia

     · NCBI PubMed 2026
     used by PRDM9 - a germline risk mechanism for childhood ALL that does not reach KMT2A
181. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9NQV7 (PRDM9_HUMAN), histone-lysine N-methyltransferase PRDM9

     · UniProt Consortium 2026
     used by PRDM9 - a germline risk mechanism for childhood ALL that does not reach KMT2A
182. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9NQV7 (PRDM9_HUMAN) SUBCELLULAR LOCATION comment and note, fetched 2026-09-21

     UniProt Consortium · UniProtKB 2026
     used by PRDM9 - a germline risk mechanism for childhood ALL that does not reach KMT2A
183. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q96EQ9 (PRDM9_MOUSE) SUBCELLULAR LOCATION comment and note, fetched 2026-09-21

     UniProt Consortium · UniProtKB 2026
     used by PRDM9 - a germline risk mechanism for childhood ALL that does not reach KMT2A
184. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Cell Ontology term lookup for pre-B and precursor B cell terms via the EMBL-EBI Ontology Lookup Service OLS4 API

     EMBL-EBI Ontology Lookup Service; Cell Ontology consortium · Database 2026
     used by Pre-B cell — the stage KMT2A::AFF1 leukaemia never reaches, and why that boundary is informative
185. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Human Protein Atlas v25.1 downloadable data and programmatic access pages, fetched 2026-09-23

     2026
     used by Prey recovery in the two BioID depositions tracks SIZE and not abundance - and the interactome tail was being scored on the wrong side of the pair, because every one of the 22 partners is the BAIT and every core protein is the PREY
186. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P78527 (PRKDC_HUMAN), DNA-dependent protein kinase catalytic subunit, 4128 aa

     · UniProt Consortium 2026
     used by PRKDC / DNA-PKcs - the kinase that phosphorylates AF9 on the one regulatory residue the fusion keeps
187. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O60678 (ANM3_HUMAN), Protein arginine N-methyltransferase 3 - FUNCTION, CATALYTIC ACTIVITY, SUBUNIT, SUBCELLULAR LOCATION, DOMAIN and feature table, fetched 2026-09-22

     · UniProt Consortium 2026
     used by PRMT3 - the mostly-cytoplasmic ribosome methyltransferase that also writes the ACTIVATING mark on the residue PRMT5 represses, has an acute-leukaemia degrader, and is invisible to this project's own mark-query rule
188. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
     ClinicalTrials.gov API v2 records for NCT03614728 and NCT03886831 - overallStatus, whyStopped, phases, enrollmentInfo, conditions and minimumAge, fetched 2026-09-22

     · ClinicalTrials.gov 2026
     used by PRMT5 - the type II arginine methyltransferase this graph has reached for sixteen times, whose mark on H3R2 RECRUITS WDR5 to chromatin
189. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed esearch counts, querytranslations and warninglists for PRMT7 against KMT2A, MLL, menin, leukaemia and the H3R2 marks, run 2026-09-22

     · NCBI PubMed 2026
     used by PRMT7 - the enzyme whose class is still disputed, which represses a KMT2 enzyme by writing H4R3me2s, which is required for leukaemia stem cells in CML, and whose only H3R2 claim is the one this beat was dispatched to check
190. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Cell Ontology term lookup via the EMBL-EBI Ontology Lookup Service OLS4 API

     EMBL-EBI Ontology Lookup Service; Cell Ontology consortium · Database 2026
     used by Pro-B cell — the stage at which KMT2A::AFF1 leukaemia is arrested, and the phenotype the diagnosis is made on
191. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities searches run 2026-09-20: (MLL[tiab] OR KMT2A[tiab]) AND "recombination signal sequence"[tiab] returns 0; the unquoted variant returns 16 with a querytranslation mapping the phrase to protein sorting signals and RSS to webcast[Publication Type]

     Q32-break-machinery · NCBI E-utilities 2026
     used by Programmed DNA breaks - V(D)J and class switching - and the verdict on whether either one breaks KMT2A
192. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O43490 (PROM1_HUMAN), Prominin-1, Antigen AC133, CD antigen CD133, 865 aa, single reviewed human record, primary gene name PROM1

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by PROM1 / CD133 - a direct MLL-AF4 target switched on by an intragenic H3K79me2/3 enhancer, and the one surface target in this disease with a published human-relevant kill
193. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O43490 (PROM1) SUBCELLULAR LOCATION comment, fetched 2026-09-21

     · UniProt Consortium 2026
     used by PROM1 / CD133 - a direct MLL-AF4 target switched on by an intragenic H3K79me2/3 enhancer, and the one surface target in this disease with a published human-relevant kill
194. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch, 2026-09-21. the single hit was retrieved by efetch and title-checked before the absence was recorded

     · NCBI PubMed 2026
     used by PROM1 / CD133 - a direct MLL-AF4 target switched on by an intragenic H3K79me2/3 enhancer, and the one surface target in this disease with a published human-relevant kill
195. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Europe PMC RESTful article API, core record for DOI:10.1182/blood.2025028877 - isOpenAccess N, inEPMC N, inPMC N, single fullTextUrl 'Subscription required'

     2026
     used by PROPOSED: run a second oncogene in LIN28B-null human fetal liver CD34+ cells - the control that decides whether the KMT2A::AFF1 transformation block is fusion-specific or just a broken progenitor, PUBLISHED, DONE: human fetal liver CD34+ cells fail to transform on KMT2A::AFF1 induction without LIN28B (Ling et al., Blood 2026) - a dedicated human index-fusion experiment, read from the abstract only, Targeting LIN28B - the essential-function and wrong-sign risks are established by the SAME paper that makes it interesting, and the sign may reverse with lineage
196. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     PubMed esearch enumeration of proximity-labelling terms against KMT2A/MLL and the super elongation complex, run 2026-09-19 with query translations recorded

     · NCBI PubMed (E-utilities) 2026
     used by Proximity-interactome beat: every query run, every paper rejected, and the one resource not reachable, Re-adjudicating the standing weak spot: proximity labelling HAS sampled AFF1, KMT2A and five leukaemia fusions - just never a KMT2A fusion, The 50-RBP BioID2 proximity map in K562 - the first modern proximity dataset this graph can reach, and AFF1 is in it as prey
197. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Negative PubMed search: (BioID OR TurboID OR "proximity labeling") AND (KMT2A OR MLL) AND leukemia, returning 0 results; and a second query adding interactome and proteomics terms, also 0. Conducted for this graph on 2026-09-18 via the NCBI PubMed E-utilities

     Oncofusion Knowledge Graph project, using the NCBI PubMed E-utilities · Literature search 2026
     used by Proximity-labelling map of KMT2A::AFF1 — the experiment nobody has run in twenty years of interactome work
198. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed E-utilities esearch, proximity-labelling terms against KMT2A/MLL

     · NCBI PubMed 2026
     used by Proximity-labelling map of KMT2A::AFF1 — the experiment nobody has run in twenty years of interactome work
199. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 42044713
     The ATF4/PSAT1/JNK signaling axis suppresses ferroptosis to drive venetoclax resistance in AML.

     Zhu X, Huang W, Zhang M, et al., Zhang H · Exp Cell Res 2026
     used by PSAT1 (Q9Y617) - a DIRECT ATF4 target, and the gene that shows the brief's menin-selectivity claim was PHGDH's and not the pathway's
200. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q9Y617 (SERC_HUMAN), Phosphoserine aminotransferase

     UniProt Consortium · UniProt 2026
     used by PSAT1 (Q9Y617) - a DIRECT ATF4 target, and the gene that shows the brief's menin-selectivity claim was PHGDH's and not the pathway's
201. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI E-utilities esearch against PubMed, 2026-09-22, fusion operand set with each term separately tagged and every cytogenetic term quoted

     · PubMed 2026
     used by PSAT1 (Q9Y617) - a DIRECT ATF4 target, and the gene that shows the brief's menin-selectivity claim was PHGDH's and not the pathway's
202. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB entries 6S01, 8PC5, 8PC6, 8PEO, 8PEP, 8CBN, 8CBQ, 2MSR, 3U88, 2N3A, 2B4J, 3F9K and 3HPG, with per-entity UniProt accessions and modelled residue ranges resolved against the RCSB search and data GraphQL APIs

     RCSB Protein Data Bank · RCSB PDB 2026
     used by PSIP1 / LEDGF p75 — the tether with the best therapeutic index in the whole interactome, The LEDGF IBD–KMT2A grip: the second half of the clamp, and nobody is drugging it
203. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL database version 34, target CHEMBL3988590 (PC4 and SFRS1-interacting protein): target_search, get_bioactivity and get_mechanism queries

     EMBL-EBI ChEMBL · ChEMBL 2026
     used by PSIP1 / LEDGF p75 — the tether with the best therapeutic index in the whole interactome, The LEDGF IBD–KMT2A grip: the second half of the clamp, and nobody is drugging it
204. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry O75475 (PSIP1_HUMAN) and AlphaFold Protein Structure Database entry AF-O75475-F1, both queried directly by API

     UniProt Consortium; AlphaFold Protein Structure Database (EMBL-EBI / Google DeepMind) · Databases 2026
     used by PSIP1 / LEDGF p75 — the tether with the best therapeutic index in the whole interactome
205. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI PubMed esearch via scripts/pubmed_safe.py, run 2026-09-23; querytranslations recorded: "FASTKD2"[Title/Abstract] AND "KMT2A"[Title/Abstract], "FASTKD2"[Title/Abstract] AND "MLL"[Title/Abstract]

     2026
     used by PSIP1-FASTKD2 declined: the bait is a mitochondrial matrix and nucleoid protein with two experimental localisations, the prey is nucleoplasmic LEDGF, and the row is one spoke-expanded BioID2 proximity in K562
206. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q01167, retrieved 2026-09-24

     · UniProt Consortium 2026
     used by PSIP1-FOXK2: a single TAP row from the same TF atlas as the RBPJ decline - kept as a lead rather than declined, because FOXK2 is a chromatin-recruiting forkhead expressed in lymphoid cells and LEDGF is a chromatin-recruiting tether
207. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct molecular interaction database, REST findInteractions on O75475 (PSIP1), 114 records retrieved 2026-09-23 by Q129

     · IntAct (EMBL-EBI) 2026
     used by PSIP1-GDF5 is a single DSSO crosslink in isolated U2OS nuclei with no bait on either side; GDF5 is a secreted cartilage growth factor, and the source paper itself attributes 30% of its cross-compartment links to contamination during nuclei isolation - declined, PSIP1-IGHMBP2 is a single DSSO crosslink in one detergent fraction of U2OS nuclei with no bait; unlike the other Fasci rows the compartments match, so this is declined on evidence depth alone - and IGHMBP2 is a fourth ribosome-biogenesis-adjacent protein on this list, PSIP1-PPP1R13B (ASPP1) is a single DSSO crosslink in unfractionated U2OS nuclei with no bait; both proteins are nuclear and both are stress-response regulators, so this is declined on evidence depth alone and is the most biologically coherent row in the beat, PSIP1-TMEM214 is a DSSO crosslink in two preparations of isolated U2OS nuclei with no bait on either side; TMEM214 is an ER-membrane mediator of CASP4-dependent ER-stress apoptosis - declined, PSIP1-UTP3 is one spoke-expanded BioID2 proximity row in K562 with UTP3 as bait and PSIP1 as prey at miscore 0.27; PSIP1 is recovered by 9 of the 50 baits in that screen and only 2 of the 9 are nucleolar, which is the panel's own base rate - declined, The ABT1/KRR1/UTP3 nucleolar triple: pre-registered, then adjudicated - it is not three proteins on two baits but a seven-member ribosomal block in one screen plus one hit in a neighbourhood that shows no nucleolar enrichment, and two of the four pre-registered discriminators were void exactly as the pre-registration predicted, The REST-wide direction audit Q104 asked for: scaffold-as-bait rows DO exist in the 36% of the record PSICQUIC hides - KMT2A is a bait in 96 of 376 rows across 18 publications and PSIP1 in 11 of 114 across 4, but the most recent is 2015 and 2012, and none of the thirteen never-discussed partners comes from one
208. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries P43026 (GDF5) and O75475 (PSIP1): primary gene name, sequence length, subcellular location and function, retrieved 2026-09-23

     · UniProt Consortium 2026
     used by PSIP1-GDF5 is a single DSSO crosslink in isolated U2OS nuclei with no bait on either side; GDF5 is a secreted cartilage growth factor, and the source paper itself attributes 30% of its cross-compartment links to contamination during nuclei isolation - declined
209. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries P38935 (IGHMBP2) and O75475 (PSIP1): primary gene name, sequence length, subcellular location with evidence codes, function and tissue specificity, retrieved 2026-09-23

     · UniProt Consortium 2026
     used by PSIP1-IGHMBP2 is a single DSSO crosslink in one detergent fraction of U2OS nuclei with no bait; unlike the other Fasci rows the compartments match, so this is declined on evidence depth alone - and IGHMBP2 is a fourth ribosome-biogenesis-adjacent protein on this list
210. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries Q96KQ4 (PPP1R13B) and O75475 (PSIP1): primary gene name, sequence length, subcellular location with evidence codes, function and tissue specificity, retrieved 2026-09-23

     · UniProt Consortium 2026
     used by PSIP1-PPP1R13B (ASPP1) is a single DSSO crosslink in unfractionated U2OS nuclei with no bait; both proteins are nuclear and both are stress-response regulators, so this is declined on evidence depth alone and is the most biologically coherent row in the beat
211. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries Q6NUQ4 (TMEM214) and O75475 (PSIP1): primary gene name, sequence length, subcellular location with evidence codes and function, retrieved 2026-09-23

     · UniProt Consortium 2026
     used by PSIP1-TMEM214 is a DSSO crosslink in two preparations of isolated U2OS nuclei with no bait on either side; TMEM214 is an ER-membrane mediator of CASP4-dependent ER-stress apoptosis - declined
212. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries Q9NQZ2 (UTP3) and O75475 (PSIP1): primary gene name, sequence length, subcellular location with evidence codes and function, retrieved 2026-09-23

     · UniProt Consortium 2026
     used by PSIP1-UTP3 is one spoke-expanded BioID2 proximity row in K562 with UTP3 as bait and PSIP1 as prey at miscore 0.27; PSIP1 is recovered by 9 of the 50 baits in that screen and only 2 of the 9 are nucleolar, which is the panel's own base rate - declined
213. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     ana-kmt2a-proximity-promiscuity-is-explained-by-size-97th-percentile-becomes-91st-among-proteins-its-own-length and ana-rbp-bioid2-proximity-map-k562, Oncofusion Knowledge Graph node files, read 2026-09-23

     · Oncofusion Knowledge Graph (internal) 2026
     used by PSIP1-UTP3 is one spoke-expanded BioID2 proximity row in K562 with UTP3 as bait and PSIP1 as prey at miscore 0.27; PSIP1 is recovered by 9 of the 50 baits in that screen and only 2 of the 9 are nucleolar, which is the panel's own base rate - declined, The size correction that dissolved KMT2A's proximity promiscuity should run the OTHER WAY for PSIP1: at 530 aa, recovery by 9 of 50 BioID2 baits may be MORE than its length predicts, which would make PSIP1 rather than KMT2A the promiscuous prey in this project's core pair
214. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P78330 (SERB_HUMAN), Phosphoserine phosphatase

     UniProt Consortium · UniProt 2026
     used by PSPH (P78330) - the third serine-synthesis enzyme, high in Ewing sarcoma alongside PHGDH and PSAT1, and the one that does NOT fall in leukaemia
215. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paperPMID 42637878
     PTBP1 controls oncogenic transcript processing and maintenance of acute myeloid leukemia

     Cao Z, et al. · Leukemia 2026
     used by PTBP1 — the strongest RNA-layer dependency in KMT2A-rearranged lines, found by a domain-focused screen and missed by this beat's own gene panel, RBM15 — m6A adaptor, AMKL fusion partner, and a named hit in the only KMT2A-biased RNA-protein CRISPR screen, RBM15B — the bridge from the DOT1L mark to the RNA layer: it reads H3K79me2 and steers m6A to start codons, The RNA layer and the menin interface are separable: IGF2BP3 depletion sensitises MLL-AF4 leukaemia to menin inhibition and outperformed it alone, The RNA layer scored for KMT2A-selectivity across 79 genes, lineage-split — PTBP1 is the only menin-shaped hit, and the IGF2BP family shows no co-dependency
216. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P26599 (PTBP1_HUMAN), human PTBP1, 557 aa

     · UniProt Consortium 2026
     used by PTBP1 — the strongest RNA-layer dependency in KMT2A-rearranged lines, found by a domain-focused screen and missed by this beat's own gene panel
217. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB P26599 (PTBP1_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

     · UniProt Consortium 2026
     used by PTBP1 — the strongest RNA-layer dependency in KMT2A-rearranged lines, found by a domain-focused screen and missed by this beat's own gene panel
218. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for PTBP1, fetched 2026-09-21

     · Human Protein Atlas 2026
     used by PTBP1 — the strongest RNA-layer dependency in KMT2A-rearranged lines, found by a domain-focused screen and missed by this beat's own gene panel
219. weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.database
     cBioPortal for Cancer Genomics, study aml_target_2018_pub (Pediatric Acute Myeloid Leukemia, TARGET, 2018) — mutation and structural-variant profiles queried via the public REST API on 2026-09-18

     cBioPortal for Cancer Genomics · Database 2026
     used by PTPN11 in KMT2A-r leukemia — named in every gene list, supported by almost no cases
220. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI E-utilities esearch, db=pubmed, fourteen single-term queries run 2026-09-24 via scripts/pubmed_safe.py

     · NCBI PubMed 2026
     used by PubMed deletes zero-hit terms from an OR-group's querytranslation, so the translation is a free audit of which drug names in your list actually contributed - and five of mine contributed nothing
221. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI E-utilities esearch, db=pubmed, run 2026-09-24; analyses/q159-retest-the-abstract-level-bounded-absences/probe-log.jsonl, record label ledgf/agent-names-any

     · NCBI PubMed 2026
     used by PubMed deletes zero-hit terms from an OR-group's querytranslation, so the translation is a free audit of which drug names in your list actually contributed - and five of mine contributed nothing
222. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     scripts/pubmed_safe.py, oncofusion-knowledge-graph

     · this project 2026
     used by PubMed deletes zero-hit terms from an OR-group's querytranslation, so the translation is a free audit of which drug names in your list actually contributed - and five of mine contributed nothing
223. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI E-utilities esummary against PubMed, 2026-09-22, bibliographic records read back for every identifier cited or rejected in this beat

     · PubMed 2026
     used by Q109 enumeration: the ten panel members that got no entry, the measurement that declined each, the papers read and not used, and four identifier traps measured on this panel, Repairing 85 defective citation blocks, and the finding that the queue's own numbers were wrong in three directions at once
224. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI E-utilities esearch and UniProtKB REST, 2026-09-22; data/symbol-collisions.json read and its structure inspected; docs/QUEEN_BRIEF.md rule 88 command executed as written

     · PubMed / UniProt / this repository 2026
     used by Q109 enumeration: the ten panel members that got no entry, the measurement that declined each, the papers read and not used, and four identifier traps measured on this panel
225. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Committed node entries from beats q142 and q144, this repository, compared 2026-09-24: ana-q142-one-patients-two-lineages-share-no-more-fusion-target-promoters-than-two-different-infants, ana-q144-the-two-lineage-switching-patients-behave-oppositely-and-only-the-menin-inhibited, and ana-q144-the-fusion-target-promoter-sets-of-two-patients-overlap-only-1-6-fold-above-chance

     Oncofusion knowledge graph, coordinator · internal adjudication, reproducible from the repository 2026
     used by Q142 and Q144 measured the same patient's lineage switch and disagreed in direction - 0.117 'less than unrelated' against 0.400 'keeps 1,894 of 2,045 calls' - and the beat with 351 pairs and a chance floor beats the one with three pairs and none, which retracts a headline I put in a king log
226. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Retrieval log 2026-09-19: EPMC search resultType=core, NCBI PMC ID converter, elink dbfrom=pubmed db=pmc, {PMCID}/fullTextXML, efetch db=pmc with a body-element test, and the pmc.ncbi.nlm.nih.gov HTML page

     · Europe PMC / NCBI 2026
     used by Q25 enumeration: what was searched, found, rejected and could not be read across the five 3' partners
227. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB P55197 and O00255 fetched with includeIsoform=true on 2026-09-19; 21- and 22-residue windows around each reported position matched against the canonical sequences

     · UniProt Consortium 2026
     used by Q25 enumeration: what was searched, found, rejected and could not be read across the five 3' partners
228. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct molecular interaction database (EMBL-EBI), PSICQUIC tab27 record for id:Q96PU8 AND id:Q03164

     2026
     used by QKI - a KMT2A proximity row in K562, and a tumour suppressor whose loss concentrates in the HOXA subtype of T-ALL
229. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q96PU8 (QKI_HUMAN)

     2026
     used by QKI - a KMT2A proximity row in K562, and a tumour suppressor whose loss concentrates in the HOXA subtype of T-ALL
230. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paperPMID 41537349
     QKI dysregulation induces extensive splicing changes in T-cell acute lymphoblastic leukemia.

     Palhais B, Sharma ND, Fijalkowski I, et al. · Haematologica 2026
     used by QKI - a KMT2A proximity row in K562, and a tumour suppressor whose loss concentrates in the HOXA subtype of T-ALL
231. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O60216 (RAD21_HUMAN), Double-strand-break repair protein rad21 homolog, 631 aa, primary gene name RAD21

     · UniProt Consortium 2026
     used by RAD21 - the cohesin subunit with the most break-prone sites in KMT2A-rearranged patients, and the degron experiment this graph keeps asking for
232. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q92878 (RAD50_HUMAN), DNA repair protein RAD50 - sequence, feature table and subcellular location comments

     UniProt Consortium · UniProtKB 2026
     used by RAD50 (Q92878, 1312 aa) - the third MRN subunit, and the one that settles whether the complex is incoherent in DepMap: mean -0.114 with 5.2% of 1,257 lines below -0.5, so it sits with NBN and not with MRE11, yet it covaries with MRE11 across cell lines as tightly as BRCA1 does with BARD1
233. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI E-utilities esearch over PubMed via scripts/pubmed_safe.py, six queries with querytranslation read back, 2026-09-22

     National Center for Biotechnology Information · NCBI Entrez Programming Utilities 2026
     used by RAD50 (Q92878, 1312 aa) - the third MRN subunit, and the one that settles whether the complex is incoherent in DepMap: mean -0.114 with 5.2% of 1,257 lines below -0.5, so it sits with NBN and not with MRE11, yet it covaries with MRE11 across cell lines as tightly as BRCA1 does with BARD1
234. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q06609 (RAD51_HUMAN), DNA repair protein RAD51 homolog 1, 339 aa, primary gene name RAD51

     · UniProt Consortium 2026
     used by RAD51 - the one repair protein measured AT the KMT2A translocation hotspot, and ATM controls how much of it loads there
235. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     UniProtKB Q2PPJ7 (RGPA2_HUMAN) - Ral GTPase-activating protein subunit alpha-2 (AS250, 250 kDa substrate of Akt; p220)

     UniProt Consortium · UniProtKB 2026
     used by RALGAPA2 (AS250) - the catalytic brake on RalA that behaves as a tumour suppressor, fused to KMT2A exon 22 in an osteosarcoma and predicted out of frame, which makes allele DISRUPTION the better hypothesis than a chimera
236. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     RefSeq NM_020343.4 - Homo sapiens Ral GTPase activating protein catalytic subunit alpha 2 (RALGAPA2), mRNA - NP_065076.2, 1,873 aa

     NCBI RefSeq · NCBI Nucleotide 2026
     used by RALGAPA2 (AS250) - the catalytic brake on RalA that behaves as a tumour suppressor, fused to KMT2A exon 22 in an osteosarcoma and predicted out of frame, which makes allele DISRUPTION the better hypothesis than a chimera
237. weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.dataset
     DepMap CRISPR (Chronos) gene-effect scan of 11 KMT2A-rearranged versus 43 other leukaemia lines (this project's own scan)

     Broad Institute DepMap; joined locally from analyses/depmap-kmt2a-selective-scan-full · analyses/depmap-kmt2a-selective-scan-full 2026
     used by RALGAPA2 (AS250) - the catalytic brake on RalA that behaves as a tumour suppressor, fused to KMT2A exon 22 in an osteosarcoma and predicted out of frame, which makes allele DISRUPTION the better hypothesis than a chimera
238. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     WDR5 bound to MLL1 peptide (PDB 3EG6, KMT2A 3762-3773) and structural basis of WDR5-MLL interaction (PDB 3EMH, KMT2A 3764-3776); MLL1 core complex on the nucleosome (PDB 6KIU and 6PWV, KMT2A 3754-3969 and 3762-3969) - per-entity UniProt residue ranges resolved against the RCSB PDB data API

     RCSB Protein Data Bank · RCSB PDB 2026
     used by Ranking every KMT2A::AFF1 partner on dependency versus essentiality, WDR5 — the partner that is not a partner: it binds the half of KMT2A the fusion throws away, WDR5 WIN site and KMT2A Arg3765: measured at 1.7 uM, three co-crystals, and absent from every leukaemic fusion - plus the MLL3 peptide that binds it 51-fold tighter
239. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P10276 (RARA_HUMAN): reviewed sequence, length and positional feature annotations

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by RARA — domain architecture and residue coordinates
240. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     AlphaFold Protein Structure Database entry for P10276 (model AF-P10276-F1)

     Jumper J, Varadi M, et al.; EMBL-EBI/DeepMind · AlphaFold DB 2026
     used by RARA — domain architecture and residue coordinates
241. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed esearch count and title check for (RBBP4 OR RBBP7) AND (KMT2A OR MLL) AND leukemia, run 2026-09-20

     · NCBI PubMed 2026
     used by RBBP4 - the shared histone-binding subunit that every leukaemia line needs, and whose identical-length paralogue cannot replace it, RBBP7 - the silent paralogue that is not silent, and the reason this project's paralogue rule finally produced a positive answer
242. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed human records Q09028 (RBBP4_HUMAN) and Q16576 (RBBP7_HUMAN), resolved with gene_primary and length requested in fields

     · UniProt 2026
     used by RBBP4 - the shared histone-binding subunit that every leukaemia line needs, and whose identical-length paralogue cannot replace it
243. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed human record Q16576 (RBBP7_HUMAN), resolved with gene_primary and length requested in fields

     · UniProt 2026
     used by RBBP7 - the silent paralogue that is not silent, and the reason this project's paralogue rule finally produced a positive answer
244. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct REST findInteractions/Q99708 (RBBP8), 104 records, 4 against a core bait, fetched 2026-09-21

     · EMBL-EBI IntAct 2026
     used by RBBP8 / CtIP - the only unwritten core-complex partner with a DIRECT-interaction record, and the graph already held its mechanism without holding the protein, The twelve partners with no entry are three times LESS spoke-expanded than the interactome they came from - 23.1% spoke-only against a 70.6% null, Fisher p = 0.00068
245. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed esearch: (LEDGF OR PSIP1)[tiab] AND (CtIP OR RBBP8)[tiab], 5 records, run 2026-09-21

     · NCBI PubMed 2026
     used by RBBP8 / CtIP - the only unwritten core-complex partner with a DIRECT-interaction record, and the graph already held its mechanism without holding the protein
246. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q99708 (CTIP_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

     · UniProt Consortium 2026
     used by RBBP8 / CtIP - the only unwritten core-complex partner with a DIRECT-interaction record, and the graph already held its mechanism without holding the protein
247. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for RBBP8, fetched 2026-09-21

     · Human Protein Atlas 2026
     used by RBBP8 / CtIP - the only unwritten core-complex partner with a DIRECT-interaction record, and the graph already held its mechanism without holding the protein
248. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q96T37 (RBM15_HUMAN), human RBM15, 977 aa

     · UniProt Consortium 2026
     used by RBM15 — m6A adaptor, AMKL fusion partner, and a named hit in the only KMT2A-biased RNA-protein CRISPR screen
249. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q96T37 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (four nuclear location values, all ECO:0000269, plus a nuclear-pore colocalisation note)

     UniProt Consortium · UniProtKB 2026
     used by RBM15 — m6A adaptor, AMKL fusion partner, and a named hit in the only KMT2A-biased RNA-protein CRISPR screen
250. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41629530
     RBM15B recognizes H3K79me2 to guide selective m(6)A-modification of mRNA and enhance oncoprotein translation in MLL-r leukemia

     Chen TQ, et al. · EMBO J 2026
     used by RBM15B — the bridge from the DOT1L mark to the RNA layer: it reads H3K79me2 and steers m6A to start codons, The RNA layer and the menin interface are separable: IGF2BP3 depletion sensitises MLL-AF4 leukaemia to menin inhibition and outperformed it alone
251. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8NDT2 (RB15B_HUMAN), human RBM15B, 890 aa

     · UniProt Consortium 2026
     used by RBM15B — the bridge from the DOT1L mark to the RNA layer: it reads H3K79me2 and steers m6A to start codons
252. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: WRAD subunits x proximity/degron/tiling methods x KMT2A clause, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text, where methods live: 20 records pair a WRAD subunit with a proximity-labelling method and a KMT2A term, 14 pair one with a tiling-CRISPR term, and none is a proximity, degron or tiling study of WRAD in a KMT2A-fusion line
253. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: MLLT1 aliases x reciprocal/spliced fusion language, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: 100 records pair an MLLT1 alias with reciprocal-fusion language and 8 with spliced-fusion language, and none stratifies a KMT2A::MLLT1 outcome, transcriptome or drug response by reciprocal or spliced status
254. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: MLLT1/YEATS1 x condensate x phospho, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: 40 records pair an ENL alias with condensate and phosphorylation language, 11 of them congress abstract books, and none tests whether YEATS reader binding or ENL condensate behaviour is phosphorylation-gated
255. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 42268713
     The MLLT3 YEATS domain binds histone marks (H3K9/18/27ac/cr) and ncRNA (7SK), linking transcription and RNA signaling to regulate hematopoiesis.

     · Cell Reports 2026
     used by Re-bound at full text: 40 records pair an ENL alias with condensate and phosphorylation language, 11 of them congress abstract books, and none tests whether YEATS reader binding or ENL condensate behaviour is phosphorylation-gated
256. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: CXCR4 x menin-inhibitor name list, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: 65 records name CXCR4 and a menin inhibitor in the same document and the 50 read are all reviews of AML therapy - no study measures the CXCR4 axis under menin inhibition
257. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: AFF4 aliases x cell-state clauses, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: AFF4 still has no haematopoietic cell-state result - the loose intersection is 265 records and the tight one 11, and every title read is a review, a complex-membership mention or a congress abstract book
258. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: ASH2L aliases x cell-state clauses, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: ASH2L still has no haematopoietic cell-state result - the loose intersection is 429 records and the tight one 17, and every title read is a review, a complex-membership mention or a congress abstract book
259. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: ELL aliases x cell-state clauses, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: ELL still has no haematopoietic cell-state result - the loose intersection is 1055 records and the tight one 17, and every title read is a review, a complex-membership mention or a congress abstract book
260. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: KMT2B aliases x cell-state clauses, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: KMT2B still has no haematopoietic cell-state result - the loose intersection is 1065 records and the tight one 13, and every title read is a review, a complex-membership mention or a congress abstract book
261. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: LEDGF aliases x KMT2A clause x established-disease phrases, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: no published experiment removes LEDGF from an established KMT2A-rearranged leukaemia in an animal - the tight intersection returns 12 records of which 6 are congress abstract books and none is a treatment experiment
262. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Europe PMC RESTful search, MEN1 x cerebellum positive control, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: no study of MLLT1/ENL function in cerebellum or developing brain exists, and 10 of the 22 full-text hits are congress abstract books - the weakest-bounded absence in the corpus is now the best-bounded of the four ENL ones
263. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: KDM2B aliases x nucleolar terms x KMT2A clause, 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: still no publication puts KDM2B's rDNA repression together with a KMT2A fusion's nucleolar residence - 27 records, 10 of them congress abstract books, the rest reviews - but KDM2B ChIP-seq in MLL-AF9 cells is now deposited, so the rDNA re-map has a second and better dataset
264. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: full-text field, menin-inhibitor name list x gain vocabulary, run 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: still nobody has asked whether a KMT2A fusion GAINS chromatin sites under a menin inhibitor - but the data to ask it are now deposited, 158 ChIP-seq samples including spike-in-normalised FLAG-MLLAF9 under VTP50469, so the class moves from never-measured to measured-not-analysed
265. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     NCBI GEO series GSE260742 - DOT1L provides transcriptional memory through PRC1.1 antagonism [ChIP-Seq]

     · NCBI Gene Expression Omnibus 2026
     used by Re-bound at full text: still nobody has asked whether a KMT2A fusion GAINS chromatin sites under a menin inhibitor - but the data to ask it are now deposited, 158 ChIP-seq samples including spike-in-normalised FLAG-MLLAF9 under VTP50469, so the class moves from never-measured to measured-not-analysed
266. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: full-text field, queries run 2026-09-22

     · Europe PMC 2026
     used by Re-bound at full text: the H3K79me2 decay half-life is still unmeasured - one paper in 12.4 million full texts contains the phrase 'half-life of H3K79me2', it is an unnumbered assertion citing a 2010 review, and the MLL::AF9 degron time-course that could have measured it uses the word zero times
267. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Antibody Registry full-text search endpoint /api/fts-antibodies, queries run 2026-09-22

     · SciCrunch Antibody Registry 2026
     used by Re-bound in the matching instrument: the Antibody Registry holds no junction-specific antibody for any KMT2A fusion under five spellings, while the same query template returns the synovial sarcoma clone as a positive control and turns up two leukaemia-fusion junction reagents nobody here had found
268. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: antibody phrases x KMT2A clause, 2026-09-22

     · Europe PMC 2026
     used by Re-bound in the matching instrument: the Antibody Registry holds no junction-specific antibody for any KMT2A fusion under five spellings, while the same query template returns the synovial sarcoma clone as a positive control and turns up two leukaemia-fusion junction reagents nobody here had found
269. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     analyses/can-the-corpus-refute-its-own-absences/run.py and results.json, this repository, computed 2026-09-24 against the committed node files and the citation index repaired in cycle 120

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by Re-checking all 60 bounded-absence claims against the repaired citation index returns zero strong candidates - and that null is not reassurance, because the one absence known to be false ranks first for the wrong reason and its actual refutation is a deposit the method cannot see
270. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     concept-a-genome-wide-null-flatters-any-interval-that-belongs-to-an-enriched-class

     Q35-encode-reanalysis · this graph 2026
     used by Read all 76 archived CBX3 and BRCA1 tracks at the KMT2A exon-11 anchor at SIGNAL level against a CTCF-anchor-matched null, and score the anchor's rank among its own class in each - no wet work, no new data, and it decides whether the sub-IDR peak is a foothold or a threshold artefact
271. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     ENCSR541AMF - CTCF ChIP-seq in SK-N-SH; optimal IDR thresholded peaks ENCFF685KTA (GRCh38), the default_analysis preferred_default file

     Bradley Bernstein, Broad Institute · ENCODE Portal 2026
     used by Read at last, against a pre-registered class-matched null: in SU-DIPG13 and SU-DIPG36 the SUV39H1 CUT&RUN mean at the KMT2A exon-11 CTCF anchor is 0.0000 and 1.6660 - the first an EXPLICIT zero record, not missing data - and no one of the seven KMT2A intervals in either line exceeds 1.21x the genome-wide random-tile mean. The pre-registered NEGATIVE prior held., The rarest thing in the deposit, measured: across 59,821 CTCF anchors in the SAME nuclei, SUV39H1 occupancy and H3K9me3 correlate at Spearman 0.148 and 0.078 - and a quarter to two-fifths of the enzyme's own strongest sites carry no more of its mark than the median anchor, Two more independent cell types put the KMT2A exon-11 CTCF summit within 12 bp of the coordinate this graph carries: MCF-7 at chr11:118,488,689 (5 bp away) and SK-N-SH at chr11:118,488,682 (12 bp), both score-1000 optimal-IDR peaks - and in the same tracks the anchor ranks only 72.95th of 56,171 MCF-7 anchors, so it is an unambiguous anchor and an ordinary one
272. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Ensembl REST assembly mapping endpoint map/human/GRCh38/11:.../GRCh37 for three intervals, plus GRCh37 lookup of ENSG00000118058 and ENST00000534358, retrieved 2026-09-22

     Ensembl REST API; Ensembl release GRCh37 archive · Ensembl 2026
     used by Read from deposited files today: the KMT2A exon-11 CTCF/cohesin element is a top-0.05% etoposide break site on chromosome 11 in a human B-cell-precursor line and in primary CD34+ marrow - and is NOT one in HCT116, the only cell line in which cohesin was ever degraded, RUN: the eight unreported H3K9me3 samples of GSE210480, read at the KMT2A CTCF element - H3K9me3 does not move where H3K9me1 moves 3.9-fold, and the me1/me2 positive controls passed, The KMT2A exon-11 CTCF/cohesin summit, measured: GRCh38 chr11:118,488,694 (CTCF) and chr11:118,488,677 (RAD21), 17 bp apart, both inside canonical exon 11
273. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
     ClinicalTrials.gov API v2 query: CXCR4-directed interventions intersected with KMT2A/MLL terms, returning NCT01076270 only

     ClinicalTrials.gov (NIH/NLM) · ClinicalTrials.gov 2026
     used by Registry sweep: is there any CXCR4-directed trial in KMT2A-rearranged or infant leukaemia? There is not, and the one paediatric attempt closed with a single patient
274. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
     ClinicalTrials.gov API v2 queries: CXCR4-directed interventions by condition (paediatric AML, n=2; acute lymphoblastic leukaemia, n=23), with whyStopped read on every non-active record

     ClinicalTrials.gov (NIH/NLM) · ClinicalTrials.gov 2026
     used by Registry sweep: is there any CXCR4-directed trial in KMT2A-rearranged or infant leukaemia? There is not, and the one paediatric attempt closed with a single patient
275. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed E-utilities esearch, db=pubmed, run 2026-09-20

     NCBI · PubMed 2026
     used by REL (c-Rel) — the third canonical subunit, measured once and found unchanged, and named in zero papers in this disease
276. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q04864 (REL_HUMAN), Proto-oncogene c-Rel

     UniProt Consortium · UniProt 2026
     used by REL (c-Rel) — the third canonical subunit, measured once and found unchanged, and named in zero papers in this disease
277. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed counted-absence search for menin and NF-kappaB in leukaemia, run 2026-09-20, every hit title-checked

     · PubMed (NCBI) 2026
     used by RELA (p65) — menin binds it and represses it, and it is required to keep MLL oncoproteins on chromatin
278. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed E-utilities esearch, db=pubmed, run 2026-09-20

     NCBI · PubMed 2026
     used by RELB — the NF-kB subunit that SUPPRESSES MLL-AF9 leukaemia, with a direct ChIP target list and a genetic rescue, which makes it the best-evidenced anti-leukaemic node in this beat
279. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q01201 (RELB_HUMAN), Transcription factor RelB

     UniProt Consortium · UniProt 2026
     used by RELB — the NF-kB subunit that SUPPRESSES MLL-AF9 leukaemia, with a direct ChIP target list and a genetic rescue, which makes it the best-evidenced anti-leukaemic node in this beat
280. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     PRIDE archive REST v3 file listing for PXD009338, retrieved 2026-09-24 by Q180: 100 files, categories RAW 44 / PEAK 44 / SEARCH 12, 33.18 GB total

     · ProteomeXchange / PRIDE, EMBL-EBI 2026
     used by Reopen PXD009338 at peptide level — the search results are already deposited, so no re-search is needed, but the deposit carries no sample-to-file mapping and that is the real cost
281. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     PRIDE archive REST v3, PXD009338: GET /projects/PXD009338/files/sdrf returned HTTP 404 on 2026-09-24, and the file listing's fileName values carry no bait identifier

     · ProteomeXchange / PRIDE, EMBL-EBI 2026
     used by Reopen PXD009338 at peptide level — the search results are already deposited, so no re-search is needed, but the deposit carries no sample-to-file mapping and that is the real cost
282. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     scripts/check-counter-note-support.py, structural_defects.truncated_titles detector

     Oncofusion Knowledge Graph, Q40 counter-evidence quality sweep · This repository 2026
     used by Repairing 85 defective citation blocks, and the finding that the queue's own numbers were wrong in three directions at once
283. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     docs/counter-note-support-report.json, suspect_imported_facts, adjudicated by reading 2026-09-20

     Oncofusion Knowledge Graph, Q42 citation-integrity repair · This repository 2026
     used by Repairing 85 defective citation blocks, and the finding that the queue's own numbers were wrong in three directions at once
284. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     analyses/do-this-corpus-s-dois-resolve/run.py, results.json and doi-index.json, this repository, computed 2026-09-24 against the committed node files with each DOI resolved live against the Crossref REST API and the DOI handle system

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by Resolving all 3,438 cited DOIs, the last identifier class never checked here - 99.8% resolve, every apparent defect is a wrong DOI the corpus quotes in order to correct it, and the three that are genuinely unregistered turn out to carry a working PMID in the same block, because 99.7% of citations here carry two identifiers
285. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q13127 (REST_HUMAN), RE1-silencing transcription factor - FUNCTION, SUBUNIT, SUBCELLULAR LOCATION, TISSUE SPECIFICITY, DOMAIN, DISEASE and feature table, fetched 2026-09-22

     · UniProt Consortium 2026
     used by REST / NRSF - the transcription factor that delivers G9a to chromatin through MED19 and MED26, is a paediatric cancer gene, and is the worst search token in this corpus
286. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.journal-articlePMID 42170936
     Blocking SETD2 Enhances the Therapeutic Efficiency of Menin Inhibitor in MLL-Fusion Leukemia.

     Li A, Zhang Y, Li Z, et al. · Cancer Sci 2026
     used by Revumenib does not move H3K27me3 in MLL-fusion leukaemia, in vitro or in vivo: the absence narrows a THIRD time and the experiment's prior collapses, SETD2 (Q9BYW2) - the H3K36 trimethyltransferase that binds all seven MLL fusions, and the only one of them with a published KMT2A::AFF1 dependency test
287. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB search for RGPD1-RGPD8 and RANBP2 in Homo sapiens, reviewed entries with accession, primary gene name and length, retrieved 2026-09-24

     · UniProt Consortium, EMBL-EBI/SIB/PIR 2026
     used by RGPD8 is the top-ranked MLL-AF4-specific prey on 3 spectra, and 96.5% of its tryptic peptides are shared with seven near-identical paralogues — only 4 peptides in the whole protein could ever have discriminated it — declined
288. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     In-silico tryptic digest of RGPD8 against its paralogues, run 2026-09-24 by Q180. Script analyses/q180-is-af4-specificity-a-depth-artefact/rgpd8_peptide_ambiguity.py, output rgpd8_peptide_ambiguity.json; sequences fetched live from UniProt FASTA and the accession asserted against each header.

     · This repository (script and JSON committed) plus UniProtKB FASTA 2026
     used by RGPD8 is the top-ranked MLL-AF4-specific prey on 3 spectra, and 96.5% of its tryptic peptides are shared with seven near-identical paralogues — only 4 peptides in the whole protein could ever have discriminated it — declined
289. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed esearch via scripts/pubmed_safe.py and UniProtKB O14715, both 2026-09-24 by Q180. querytranslation read back as "RGPD8"[Title/Abstract], non-empty, so the count is a genuine 2 and not an instrument failure.

     · NCBI PubMed E-utilities and UniProt Consortium 2026
     used by RGPD8 is the top-ranked MLL-AF4-specific prey on 3 spectra, and 96.5% of its tryptic peptides are shared with seven near-identical paralogues — only 4 peptides in the whole protein could ever have discriminated it — declined
290. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed human record for RING1 (Q06587), resolved by gene_exact:RING1 AND organism_id:9606 AND reviewed:true with gene_primary and length read back

     UniProt Consortium · UniProtKB 2026
     used by RING1 (RING1A) - the paralogue that is not RING1B, and the reason a substring search for 'RING1' in this graph returns the wrong gene
291. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch, 2026-09-21. alias-aware over RING1, RING1A and RNF1; the operand control was run so the zero is not being read off an empty intersection of two large sets (rule 67 item 2)

     · NCBI PubMed 2026
     used by RING1 (RING1A) - the paralogue that is not RING1B, and the reason a substring search for 'RING1' in this graph returns the wrong gene
292. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RN7SK, RNA component of 7SK nuclear ribonucleoprotein: HGNC:10037 (HGNC REST), ENSG00000283293 chr6:52,995,621-52,995,948 GRCh38 plus strand (Ensembl REST lookup), NCBI Gene 125050, RefSeq NR_001445.2 fetched as 332 nt

     HGNC; Ensembl; NCBI RefSeq · Databases 2026
     used by RN7SK / 7SK snRNA - the 332-nucleotide brake on elongation, and the one molecule in this beat that every dataset this project uses is blind to
293. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q99496 (RING2_HUMAN), E3 ubiquitin-protein ligase RING2

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by RNF2 (RING1B) — the one catalytic subunit shared by every PRC1 flavour, which is why the genetics point two ways at once
294. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41533567
     Structural insights into the Bre1-Lge1 and RNF20/RNF40-WAC interactions critical for H2B ubiquitination.

     Shi M, Wang X, Zhang H, Wen Y, Liu Q, Chen P, Chen X, Xiang S · Nucleic Acids Res 2026
     used by RNF20 (BRE1A) — the H2B ubiquitin ligase that MLL-rearranged leukaemia needs, and that DepMap says every other cancer line needs just as much, WAC — the adaptor that hands RNF20/RNF40 to RNA polymerase II, and the only member of this axis whose knockdown alone killed MLL-rearranged ALL cells
295. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41577693
     In vivo CRISPR screening identifies SAGA complex members as key regulators of hematopoiesis.

     Shankar A, Olender L, Hsu I, Miyauchi M, Pálovics R, Meaker GA, Kaito S, Rizq O, Khoo HM, Bozhilov Y, Igarashi KJ, Bhadury J, Munson C, Mack PK, Tan TK, Rehwinkel J, Iwama A, Wyss-Coray T, Nakauchi H, Haney MS, Wilkinson AC · Nat Commun 2026
     used by RNF20 (BRE1A) — the H2B ubiquitin ligase that MLL-rearranged leukaemia needs, and that DepMap says every other cancer line needs just as much
296. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     UniProtKB O75150 (BRE1B_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

     · UniProt Consortium 2026
     used by RNF40 (BRE1B) — the obligate partner of RNF20, and the one of the pair with a documented link to differentiation and bivalency resolution
297. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for RNF40, fetched 2026-09-21

     · Human Protein Atlas 2026
     used by RNF40 (BRE1B) — the obligate partner of RNF20, and the one of the pair with a documented link to differentiation and bivalency resolution
298. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     Route-5 sweep: resolving the DOI of every paper this graph flags as unreadable or abstract-only that has no PMCID, and measuring the stripped text length returned

     · Oncofusion knowledge graph (internal analysis) 2026
     used by Route 5 rescued 2 of 32, and 21 of the 30 failures returned HTTP 200 with no article - each publisher failing at a constant, recognisable length
299. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P27694 (RFA1_HUMAN), Replication protein A 70 kDa DNA-binding subunit - accession, primary gene name, sequence length, feature table and subcellular-location and subunit comments, fetched live 2026-09-24

     · UniProt Consortium, EMBL-EBI/SIB/PIR 2026
     used by RPA1 (RPA70) - the large subunit of replication protein A, which carries most of the ssDNA-binding folds and the N-terminal fold that hands the lesion to ATR, The replication protein A heterotrimer (RPA1/RPA2/RPA3) - the ssDNA clamp that every resection and replication-stress claim in this corpus silently depends on, and which had no entry until cycle 160
300. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Human Protein Atlas search_download API, query 'RPA1', columns gene/ensembl/synonym/subcellular-location/reliability, fetched 2026-09-24; RPA1 = ENSG00000132383, Reliability (IH) Supported, Subcellular location [Nucleoplasm]

     · Human Protein Atlas, proteinatlas.org 2026
     used by RPA1 (RPA70) - the large subunit of replication protein A, which carries most of the ssDNA-binding folds and the N-terminal fold that hands the lesion to ATR, RPA1 and RPA2 are curated HGNC aliases of POLR1A and POLR1B, and RPA34 - a UniProt synonym of RPA2 - is an alias of POLR1G, so three of the ssDNA clamp's names point at RNA polymerase I subunits; the hazard is in symbol-based DATABASE lookups, not in PubMed, and the measured contamination says which
301. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     HGNC REST API fetch/symbol/{POLR1A,POLR1B,POLR1C,RPA1,RPA2,RPA3}, all six fetched live 2026-09-24 with Accept: application/json; hgnc_id, name, alias_symbol and prev_symbol read back for each

     · HUGO Gene Nomenclature Committee, genenames.org 2026
     used by RPA1 and RPA2 are curated HGNC aliases of POLR1A and POLR1B, and RPA34 - a UniProt synonym of RPA2 - is an alias of POLR1G, so three of the ssDNA clamp's names point at RNA polymerase I subunits; the hazard is in symbol-based DATABASE lookups, not in PubMed, and the measured contamination says which
302. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Human Protein Atlas search_download API, query 'RPA2', fetched 2026-09-24; RPA2 = ENSG00000117748, Reliability (IH) Supported, Subcellular location [Nucleoplasm, Nuclear bodies]; second row POLR1B = ENSG00000125630, gene synonyms including RPA2 and RPA135, Subcellular location [Nucleoli fibrillar center, Cytosol]

     · Human Protein Atlas, proteinatlas.org 2026
     used by RPA2 (RPA32) - the middle subunit of replication protein A, whose disordered N-terminus is the phospho-switch of the replication-stress response and whose 43-171 region binds menin, Where RPA actually is: nucleoplasmic and diffuse until there is single-stranded DNA, then punctate prereplication foci at the G1/S border, damage-induced foci with RAD51, and a PML-body pool - and the subunits differ in how well any of this is measured
303. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P35244 (RFA3_HUMAN), Replication protein A 14 kDa subunit - accession, primary gene name, sequence length, complete feature table and subcellular-location and subunit comments, fetched live 2026-09-24

     · UniProt Consortium, EMBL-EBI/SIB/PIR 2026
     used by RPA3 (RPA14) - the 121-residue subunit of replication protein A, which has no annotated domain, no HPA localisation call and no connection to this disease at all
304. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Human Protein Atlas search_download API, query 'RPA3', columns gene/ensembl/synonym/subcellular-location/reliability, fetched 2026-09-24; RPA3 row returns Reliability (IH) 'Approved' and Subcellular location null

     · Human Protein Atlas, proteinatlas.org 2026
     used by RPA3 (RPA14) - the 121-residue subunit of replication protein A, which has no annotated domain, no HPA localisation call and no connection to this disease at all
305. weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.database
     IntAct molecular interaction database, PSICQUIC MITAB 2.7, queries id:Q9NS64 AND id:O00255 and id:O00255 partitioned by publication, retrieved 2026-09-21

     · IntAct (EMBL-EBI) 2026
     used by RPRM (reprimo) - the other half of menin's Ewing 2007 pair, a protein with no experimental localisation at all, and a promoter methylated in 45% of paediatric AML
306. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries Q9NS64 (RPRM_HUMAN, Protein reprimo) and O00255 (MEN1_HUMAN), subcellular location and function comments with evidence codes

     · UniProtKB/Swiss-Prot 2026
     used by RPRM (reprimo) - the other half of menin's Ewing 2007 pair, a protein with no experimental localisation at all, and a promoter methylated in 45% of paediatric AML
307. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed counted-absence search for RPRM/reprimo against menin and MEN1, run 2026-09-21

     · PubMed (NCBI) 2026
     used by RPRM (reprimo) - the other half of menin's Ewing 2007 pair, a protein with no experimental localisation at all, and a promoter methylated in 45% of paediatric AML
308. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     NCBI PubMed esearch, bounded-absence query (counted negative)

     · NCBI E-utilities 2026
     used by RRP12 (Q5JTH9) - a pre-40S export factor in six of seven MLL-fusion interactomes but NOT in the index fusion's, and the block's only member with a recessive human disease published in 2025
309. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB - O43818 (U3IP2_HUMAN), U3 small nucleolar RNA-interacting protein 2

     · UniProt Knowledgebase 2026
     used by RRP9 / U3-55K (O43818) - the U3 snoRNP-specific core protein in four MLL-fusion interactomes but NOT the index fusion's, and the block's only member with a published acetylation switch
310. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     NCBI PubMed esearch, bounded-absence query (counted negative)

     · NCBI E-utilities 2026
     used by RRP9 / U3-55K (O43818) - the U3 snoRNP-specific core protein in four MLL-fusion interactomes but NOT the index fusion's, and the block's only member with a published acetylation switch
311. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI PubMed esearch, bounded-absence queries with querytranslation recorded (counted negatives)

     · NCBI E-utilities 2026
     used by RSBN1 / KDM9 (Q5VWQ0) - an H4K20me2/me3 demethylase in the AFF4 proximity record, and the only read-queue partner that is a chromatin enzyme at all
312. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UCSC Genome Browser REST API, /search endpoint, genome=mm10, queries for Hoxa9, Meis1 and Six1, run 2026-09-23

     · UCSC Genome Browser 2026
     used by Rule 62 applies to a genome browser: the UCSC search API returns several tracks and many rows per gene, 'take the longest' put Hoxa9 and Meis1 on chr1, only the refGene track carries the SYMBOL in posName, and xenoRefGene returns a human-aligned Six1 on mouse chr17
313. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Ensembl REST API, lookup/symbol/homo_sapiens/KMT2A?expand=1, GRCh38 (rest.ensembl.org) and GRCh37 (grch37.rest.ensembl.org) endpoints, called 2026-09-22

     Ensembl · Ensembl REST 2026
     used by Run in coordinates: the protective CTCF site and the patient break-prone window are ONE element, it is cohesin-bound, and the 'functional cut point' is 3.75 kb away - two of the graph's three exon-11s were the same exon and the third was not
314. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UCSC Genome Browser REST API, getData/track, track cytoBand, genome hg38, chrom chr11, called 2026-09-22

     UCSC Genome Browser · UCSC Genome Browser 2026
     used by Run in coordinates: the protective CTCF site and the patient break-prone window are ONE element, it is cohesin-bound, and the 'functional cut point' is 3.75 kb away - two of the graph's three exon-11s were the same exon and the third was not
315. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ENCODE Portal: released human TF ChIP-seq experiments for BRCA1, CBX3 and CBX1, queried and downloaded 2026-09-22

     ENCODE Project Consortium · ENCODE Portal 2026
     used by RUN, and the absence closes: all 13 released ENCODE BRCA1, CBX3 and CBX1 ChIP-seq tracks read at the KMT2A locus with three nulls - zero called peaks in 13 of 13, the breakpoint cluster at 0.57-1.24x the genome, and the exon-11 CTCF element BELOW the median CTCF anchor of its own cell type in 10 of 13
316. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ENCODE Portal: conservative IDR thresholded CTCF ChIP-seq peaks, GRCh38, in K562, GM12878, H1, HeLa-S3, HepG2 and HCT116, downloaded 2026-09-22

     ENCODE Project Consortium · ENCODE Portal 2026
     used by RUN, and the absence closes: all 13 released ENCODE BRCA1, CBX3 and CBX1 ChIP-seq tracks read at the KMT2A locus with three nulls - zero called peaks in 13 of 13, the breakpoint cluster at 0.57-1.24x the genome, and the exon-11 CTCF element BELOW the median CTCF anchor of its own cell type in 10 of 13
317. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     21.6 percent of this project's 19,404 promoter windows contain another gene's TSS

     Iqbal Lab oncofusion knowledge graph, beat Q153 · this project 2026
     used by RUN, by Q156, within the hour: the whole-table displacement audit - drift is 1.83%, the real blind spot is one row per symbol at 14.69%, and the case that motivated it is not displaced
318. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     ENCODE Portal, conservative IDR thresholded peak sets (GRCh38) of experiments ENCSR000EWD, ENCSR000EWI, ENCSR000AUT (SETDB1), ENCSR272JAT, ENCSR372GIN (CBX5), ENCSR948QLZ (CBX1), ENCSR000ATV, ENCSR549NPZ (CBX3), ENCSR175EOM (EHMT2) and ENCSR474CVP (TRIM28)

     ENCODE Project Consortium · ENCODE Portal 2026
     used by RUN: ten ENCODE peak sets and fourteen signal tracks read at the KMT2A locus in two blood cells - not one called peak for SETDB1, CBX5, CBX1, CBX3 or TRIM28 anywhere in the gene, while G9a/EHMT2 and TRIM28 both sit on the KMT2A PROMOTER and neither touches the breakpoint cluster region
319. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     ENCODE Portal, conservative IDR thresholded peaks (GRCh38): ENCSR175EOM/ENCFF373EMV (EHMT2, K562) and ENCSR474CVP/ENCFF382UDO (TRIM28, K562)

     ENCODE Project Consortium · ENCODE Portal 2026
     used by RUN: ten ENCODE peak sets and fourteen signal tracks read at the KMT2A locus in two blood cells - not one called peak for SETDB1, CBX5, CBX1, CBX3 or TRIM28 anywhere in the gene, while G9a/EHMT2 and TRIM28 both sit on the KMT2A PROMOTER and neither touches the breakpoint cluster region
320. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     analyses/q110-h3k9-writer-and-reader-at-the-kmt2a-locus/run.py and results.json - 14 ENCODE per-replicate fold-change-over-control bigWigs read at the KMT2A locus, complete run 2026-09-22

     Q110-h3k9me3-writer-layer, Oncofusion Knowledge Graph · this repository 2026
     used by RUN: ten ENCODE peak sets and fourteen signal tracks read at the KMT2A locus in two blood cells - not one called peak for SETDB1, CBX5, CBX1, CBX3 or TRIM28 anywhere in the gene, while G9a/EHMT2 and TRIM28 both sit on the KMT2A PROMOTER and neither touches the breakpoint cluster region
321. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     ENCODE BRCA1, CBX3 and CBX1 ChIP-seq fold-change-over-control bigWigs and IDR peak sets, GRCh38, read 2026-09-22: ENCFF690UED, ENCFF620MRE, ENCFF133FPU, ENCFF191NNV, ENCFF493XJG, ENCFF017KXT, ENCFF431CIQ, ENCFF051QOQ, ENCFF845YDA, ENCFF888LWP, ENCFF196UXL, ENCFF058NFI, ENCFF895JUC, ranked against ENCFF041ODC, ENCFF796WRU, ENCFF497ZGV, ENCFF533TKY, ENCFF695QSJ and ENCFF850PKJ

     ENCODE Project Consortium · ENCODE Portal 2026
     used by RUN: the IDR-free, signal-level re-read of the KMT2A exon-11 anchor against a class-matched CTCF-anchor census - the pre-registered prior of NEGATIVE HELD, both arms score FAILS on the pre-registered rule (BRCA1 median percentile 62.8 of 6 tracks, CBX3 55.9 of 3), and the one thing the prior did not predict is that 2 of 6 BRCA1 tracks sit above the 90th percentile and both are the non-blood cells
322. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     ana-q35-brca1-cbx3-and-cbx1-are-not-at-the-kmt2a-breakpoint-cluster-in-any-encode-cell - kmt2a_element_percentile_among_anchors and c1_gate, analyses/q35-brca1-cbx3-encode-at-the-kmt2a-locus/results.json

     Q35-encode-reanalysis · this graph 2026
     used by RUN: the IDR-free, signal-level re-read of the KMT2A exon-11 anchor against a class-matched CTCF-anchor census - the pre-registered prior of NEGATIVE HELD, both arms score FAILS on the pre-registered rule (BRCA1 median percentile 62.8 of 6 tracks, CBX3 55.9 of 3), and the one thing the prior did not predict is that 2 of 6 BRCA1 tracks sit above the 90th percentile and both are the non-blood cells
323. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     analyses/does-the-stored-citation-match-the-pmid/run.py, results.json and pmid-record-index.json, this repository, computed 2026-09-24 with records fetched live from PubMed esummary

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by Running the wrong-paper check over the 17,925 PMID citations - the larger half of the corpus and the half where both documented failures happened - gives 100.0% year agreement and 99.7% on authors, one genuine defect, and shows the DOI side's 92% was an artefact of comparing against a single date field
324. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q01196 (RUNX1_HUMAN): reviewed sequence, length and positional feature annotations

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by RUNX1 — domain architecture and residue coordinates (with the RUNX1b/RUNX1c +27 offset)
325. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     RefSeq protein records NP_001001890 (RUNX1 isoform AML1b, 453 aa), NP_001745 (isoform AML1c, 480 aa) and NP_001116079 (isoform AML1a, 250 aa), with pairwise sequence comparison

     NCBI RefSeq; comparison computed by us · NCBI RefSeq / Ensembl 2026
     used by RUNX1 — domain architecture and residue coordinates (with the RUNX1b/RUNX1c +27 offset)
326. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     AlphaFold Protein Structure Database entry for Q01196 (model AF-Q01196-F1)

     Jumper J, Varadi M, et al.; EMBL-EBI/DeepMind · AlphaFold DB 2026
     used by RUNX1 — domain architecture and residue coordinates (with the RUNX1b/RUNX1c +27 offset)
327. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q01196 (RUNX1_HUMAN) and Q06455 (MTG8_HUMAN) canonical sequences

     · UniProt 2026
     used by RUNX1 K43 acetylation - a retained-portion site that AML1-ETO genuinely requires, in the 5' partner, with the 3' partner supplying the acetyltransferase in trans
328. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Europe PMC EXT_ID:21596992 record retrieval, 2026-09-24

     · Europe PMC REST 2026
     used by RUNX1 K43 acetylation - a retained-portion site that AML1-ETO genuinely requires, in the 5' partner, with the 3' partner supplying the acetyltransferase in trans
329. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     UniProtKB Q06455 (RUNX1T1_HUMAN): sequence length, domain, zinc-finger, DNA-binding, repeat, motif and region features, and PDB cross-references

     UniProt Consortium · UniProtKB 2026
     used by RUNX1T1 (ETO / MTG8) domain architecture
330. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q13950 (RUNX2_HUMAN), Runt-related transcription factor 2

     UniProt Consortium · UniProt 2026
     used by RUNX2 - a direct KMT2A-chimera target and dependency in T-ALL, a patient-specific enhancer target in KMT2A::AFF1 B-ALL, and the gene a menin inhibitor represses to restore steroid sensitivity in t(4;11) infants
331. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     HGNC record HGNC:10472 (RUNX2) and Ensembl gene ENSG00000124813 on GRCh37 and GRCh38

     HUGO Gene Nomenclature Committee; Ensembl · genenames.org / Ensembl 2026
     used by RUNX2 - a direct KMT2A-chimera target and dependency in T-ALL, a patient-specific enhancer target in KMT2A::AFF1 B-ALL, and the gene a menin inhibitor represses to restore steroid sensitivity in t(4;11) infants
332. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Ensembl GRCh37 REST: lookup/id/ENSG00000124813?expand=1 for per-transcript starts, and sequence/id/<transcript>?type=protein for the encoded N-termini

     Ensembl · Ensembl GRCh37 archive 2026
     used by RUNX2 - a direct KMT2A-chimera target and dependency in T-ALL, a patient-specific enhancer target in KMT2A::AFF1 B-ALL, and the gene a menin inhibitor represses to restore steroid sensitivity in t(4;11) infants
333. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8N488 (RYBP_HUMAN), RING1 and YY1-binding protein; gene_primary RYBP, 228 aa

     · UniProtKB 2026
     used by RYBP - the subunit that makes a variant PRC1 complex catalytically active, named by two entries in this graph and never written
334. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     GTEx v8 median gene expression, S1PR4 (ENSG00000125910.5) across 54 tissues: Whole_Blood 252.82 TPM, whole-body median 1.11 TPM. Fetched 2026-09-23

     · GTEx Portal 2026
     used by S1PR4 - the third parenchymally clean surface protein in the screen, a leukocyte-restricted GPCR, and not fusion-specific either
335. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed esearch via scripts/pubmed_safe.py; S1PR4[tiab] AND leukemia[tiab] -> 0, querytranslation "S1PR4"[Title/Abstract] AND "leukemia"[Title/Abstract]; run 2026-09-23

     · NCBI PubMed 2026
     used by S1PR4 - the third parenchymally clean surface protein in the screen, a leukocyte-restricted GPCR, and not fusion-specific either
336. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     The KMT2A-rearranged DepMap arm is not MTAP-deleted, and the gene that looked like the test cannot answer it

     Oncofusion Knowledge Graph coordinator · internal analysis node ana-the-kmt2a-arm-is-not-mtap-deleted-and-chronos-cannot-see-a-deletion 2026
     used by SAM / methionine dependence - the one metabolic state in this graph demonstrated in KMT2A-rearranged lymphoblastic leukaemia itself, with a drug and a PDX behind it, The MTAP-deleted, MTA-high state - a genotype-imposed metabolic state that this project's KMT2A-rearranged arm turns out NOT to have, and that its own screen could never have detected directly
337. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q9Y3Z3 (SAMH1_HUMAN), Deoxynucleoside triphosphate triphosphohydrolase SAMHD1

     UniProt Consortium · UniProt 2026
     used by SAMHD1 - the dNTP hydrolase that is the real biomarker behind the IMPDH story, and the reason the KMT2A framing of that story does not hold
338. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q01826 (SATB1_HUMAN), DNA-binding protein SATB1

     UniProt Consortium · UniProt 2026
     used by SATB1 - a nuclear-matrix and PML-body genome organiser with its own matrix-targeting sequence, the clearest subnuclear-localization entry in this beat, and a T-lineage factor that B cells do not need
339. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     HGNC record HGNC:10541 (SATB1) and Ensembl gene ENSG00000182568 on GRCh37 and GRCh38

     HUGO Gene Nomenclature Committee; Ensembl · genenames.org / Ensembl 2026
     used by SATB1 - a nuclear-matrix and PML-body genome organiser with its own matrix-targeting sequence, the clearest subnuclear-localization entry in this beat, and a T-lineage factor that B cells do not need
340. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Ensembl GRCh37 REST: lookup/id/ENSG00000182568?expand=1, per-transcript start coordinates and strand

     Ensembl · Ensembl GRCh37 archive 2026
     used by SATB1 - a nuclear-matrix and PML-body genome organiser with its own matrix-targeting sequence, the clearest subnuclear-localization entry in this beat, and a T-lineage factor that B cells do not need
341. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paperPMID 42492014
     Erratum: All-Oral Combination of Revumenib, Decitabine, and Venetoclax for Relapsed or Refractory AML (SAVE)

     Issa GC, et al. · Journal of Clinical Oncology 44(24):2361 2026
     used by SAVE (NCT05360160): the first prospective menin-inhibitor trial to report a NUP98-rearranged subgroup outcome
342. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9Y3A5 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (four location values with empty evidence lists and a fully cited, cell-cycle-scoped note)

     UniProt Consortium · UniProtKB 2026
     used by SBDS - a ribosome-maturation factor whose loss causes a leukaemia-predisposition syndrome, and whose BioID neighbourhood contains AFF1, KMT2A, CDK9, PAF1 and LEDGF
343. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Unbiased complex-level sweep: median co-dependency coherence against median KMT2A-selectivity rank across 719 human Complex Portal complexes

     · Oncofusion knowledge graph (internal analysis) 2026
     used by Scored at complex level, coherence predicts nothing (rho -0.016 over 719 complexes) - but the Fanconi core complex is 2nd of 719 and the HBO1/MOZ acetyltransferase modules are 1st
344. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Complex Portal: the manually curated resource of macromolecular complexes, human complex table (9606.tsv)

     · EMBL-EBI Complex Portal 2026
     used by Scored at complex level, coherence predicts nothing (rho -0.016 over 719 complexes) - but the Fanconi core complex is 2nd of 719 and the HBO1/MOZ acetyltransferase modules are 1st
345. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     analyses/re-score-the-marks-against-an-arm-matched-floor/run.py and results.json, this repository, computed 2026-09-23 with zero downloads from the committed results of analyses/do-the-marks-move-in-the-bulk and analyses/the-histone-mark-floor-was-on-disk; arm-level floor spread independently measured by the Q135 beat

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by Scoring cycle 104's mark panel against a floor measured in its OWN arm: 9 of 10 mark-by-arm cells at D1 land inside the floor's own spread, so D1 is not resolvable from this deposit at all - and the same mark ratios give 0, 4 or 4 breaking marks depending only on which floor was imported, so cycle 104's pass, cycle 107's F1 and the GSE299100 beat's failure were all measuring the ruler rather than the marks
346. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     analyses/is-there-a-menin-readout-under-dot1l-inhibition/run.py and results.json, this repository, computed 2026-09-23 against live NCBI GEO; GSE173599 series and sample records and PMID 35245435 fetched and verified the same session

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by Searching 32 GEO series and 2,312 samples for a menin readout under DOT1L inhibition: exactly ONE distinct experiment exists, MOLM13_MEN1-ChIP_EPZ5676_4d at 100 nM, and it is the sample Q136 already used - so the deflationary reading of Q136 is blocked not by a missing assay but by a ~50-fold DOSE GAP, and Q136's drug-class conclusion rests on a single unreplicated sample in a single cell line
347. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
     A Phase 3 Randomized, Double-Blinded Bridging Trial to Evaluate the Efficacy of Uproleselan Administered With Chemotherapy Versus Chemotherapy Alone in Chinese Patients With Relapsed/Refractory Acute Myeloid Leukemia

     · ClinicalTrials.gov 2026
     used by SELE (E-selectin) - the vascular-niche chemoresistance receptor with a clinical antagonist, zero KMT2A papers, and a paediatric trial that a company restructuring closed
348. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q15019 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (nine location values, seven from one paper, none of them nuclear)

     UniProt Consortium · UniProtKB 2026
     used by SEPTIN2 (septin-2 / NEDD5) — KMT2A fusion partner in t(2;11)(q37;q23)
349. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB entries 5CYO, 5CYP and 4YQF — human septin 9 GTPase domain structures

     RCSB Protein Data Bank · RCSB PDB entry API 2026
     used by SEPTIN9 (septin-9 / MSF) — cytoskeletal GTPase, KMT2A fusion partner in t(11;17)(q23;q25)
350. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P14618 (KPYM_HUMAN), Pyruvate kinase PKM

     UniProt Consortium · UniProt 2026
     used by Serine/glycine dependence in MLL-ENL leukaemia, reached through a DIRECT fusion target gene - and the reason a PHGDH counted absence in this graph would have been wrong
351. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB - P58004, SESN2_HUMAN, sestrin-2

     · UniProt Knowledgebase 2026
     used by SESN2 / sestrin-2 (P58004) - the ISR's exit into mTORC1: a leucine sensor that falls 1.4 log2 under either inhibitor, and the one panel member with a real structural record
352. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q01105 (SET_HUMAN), Protein SET / Template-activating factor I, 290 aa

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SET / TAF-I beta - a histone chaperone that binds the retained N-terminus of MLL, co-activates HOXA with it, and is itself a HOXA-activating fusion partner that recruits DOT1L
353. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q01105 (SET_HUMAN), Protein SET / Template-activating factor I, 290 aa

     · UniProt 2026
     used by SET / TAF-Ibeta domain map (Q01105, 290 aa) - dimerisation, earmuff, and the acidic tail that binds MLL
354. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Cellosaurus cell line Loucy (CVCL_1380)

     · Cellosaurus (SIB) 2026
     used by SET::NUP214 (TAF-Ibeta-NUP214, SET-CAN), cryptic del(9)(q34.11q34.13) - a HOXA-driving fusion of a histone chaperone to a nucleoporin, with no KMT2A and no menin-binding motif
355. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch: SET-NUP214 name variants against DOT1L inhibitor vocabularies

     · NCBI PubMed 2026
     used by SET::NUP214 recruits DOT1L by a different surface from CALM-AF10 and KMT2A::MLLT10: the founding paper tested for AF10 in the complex and did not find it
356. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O15047 (SET1A_HUMAN), retrieved via the UniProt REST API

     UniProt Consortium · UniProtKB 2026
     used by SETD1A — a leukemia dependency whose enzyme activity turns out to be dispensable
357. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB entries 8ILY and 4EWR, verified against the RCSB data API for title, experimental method and resolution

     RCSB Protein Data Bank · RCSB PDB 2026
     used by SETD1A — a leukemia dependency whose enzyme activity turns out to be dispensable
358. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     UniProtKB O15047 (SET1A_HUMAN), Histone-lysine N-methyltransferase SETD1A, 1,707 aa - SUBCELLULAR LOCATION comment

     · UniProt Knowledgebase, reviewed human record, fetched 2026-09-21 2026
     used by SETD1A — a leukemia dependency whose enzyme activity turns out to be dispensable
359. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9UPS6 (SET1B_HUMAN), Histone-lysine N-methyltransferase SETD1B

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SETD1B - the COMPASS paralogue round 4 deferred for want of a leukaemia paper, which had already been published: its SET domain sets H3K4me3 breadth and MYC in MLL-rearranged AML
360. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch record counts for the COMPASS remainder (SETD1B, CXXC1) against KMT2A, MLL and leukaemia, run for interactome round 5

     · NCBI PubMed, queried 2026-09-19 2026
     used by SETD1B - the COMPASS paralogue round 4 deferred for want of a leukaemia paper, which had already been published: its SET domain sets H3K4me3 breadth and MYC in MLL-rearranged AML
361. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q15047 (SETB1_HUMAN) - Histone-lysine N-methyltransferase SETDB1, Homo sapiens; with HGNC:10761

     UniProt Consortium; HUGO Gene Nomenclature Committee · UniProt Knowledgebase / HGNC 2026
     used by SETDB1 (Q15047, KMT1E/ESET) - the second H3K9me3 writer: it represses the fusion's own targets via the PAF1 complex, it is required in fetal liver HSPCs and in leukaemic stem cells, and it is a broad dependency in half the CRISPR panel - three published directions that do not agree
362. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB entries 3ME9 and 3MEU, verified against the RCSB data API for title, experimental method and resolution

     RCSB Protein Data Bank · RCSB PDB 2026
     used by SGF29 — a SAGA reader with a Tudor pocket, intermediate essentiality, and the best unexploited profile in this beat
363. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Ensembl REST cross-reference service, xrefs/symbol/homo_sapiens for SGF29, CCDC101 and ERG, queried 2026-09-18

     Ensembl / EMBL-EBI · Ensembl 2026
     used by SGF29 — a SAGA reader with a Tudor pocket, intermediate essentiality, and the best unexploited profile in this beat
364. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q96ES7 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (a single ECO:0000250 nucleus call transferred from the RAT orthologue P0C606)

     UniProt Consortium · UniProtKB 2026
     used by SGF29 — a SAGA reader with a Tudor pocket, intermediate essentiality, and the best unexploited profile in this beat
365. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB - O00141, SGK1_HUMAN, serine/threonine-protein kinase Sgk1

     · UniProt Knowledgebase 2026
     used by SGK1 / SGK (O00141) - a MENIN-SELECTIVE upregulated kinase, a druggable AGC family member, and the panel's worst platform disagreement
366. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P34897 (GLYM_HUMAN), Serine hydroxymethyltransferase, mitochondrial

     UniProt Consortium · UniProt 2026
     used by SHMT2 (P34897) - the mitochondrial one-carbon step, and the paralogue that rises while it falls is why its loss may not matter
367. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI E-utilities esearch against PubMed, 2026-09-22, fusion operand set (KMT2A[tiab] OR MLL[tiab] OR "MLL-AF4"[tiab] OR "KMT2A::AFF1"[tiab] OR "MLL-rearranged"[tiab] OR "t(4;11)"[tiab] OR "MLL-AF9"[tiab] OR "MLL-ENL"[tiab]), each term separately tagged and every cytogenetic term quoted

     · PubMed 2026
     used by SHMT2 (P34897) - the mitochondrial one-carbon step, and the paralogue that rises while it falls is why its loss may not matter
368. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8IUQ4 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (two location values and a note, all three with empty evidence lists)

     UniProt Consortium · UniProtKB 2026
     used by SIAH1 — the E3 ligase that degrades AFF1, and that the reciprocal AF4::KMT2A fusion escapes
369. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry
     Safety and Efficacy of Dual Menin and FLT3 Inhibition in Patients With Relapsed/Refractory FLT3-Mutated Acute Myeloid Leukemia Containing a Concurrent MLL-Rearrangement or NPM1 Mutation: A Phase I Study of SNDX-5613 + Gilteritinib

     Borate U (sponsor-investigator) · ClinicalTrials.gov 2026
     used by Signalling co-mutations (RAS pathway, FLT3-ITD, WT1) under menin-inhibitor pressure
370. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Aligning the published siMA6 21-mer against reconstructed KMT2A::AFF1 chimeric transcripts at two recurrent breakpoints

     · Oncofusion knowledge graph (internal analysis) 2026
     used by siMA6 confirmed junction-spanning by reconstruction - 8 KMT2A bases plus 13 AFF1 bases - and it targets the 1362/1363 breakpoint, not the one this graph modelled as canonical
371. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     DepMap KMT2A-selective dependency scan over 17,728 genes: analyses/depmap-kmt2a-selective-scan-full/results.json, generated 2026-09-19. SIN3A rank 17,096, difference +0.1112, mean_kmt2a_r -0.7580, mean_other -0.8691, pan_essentiality -0.8829.

     · DepMap via this repository 2026
     used by SIN3A (Q96ST3) - the scaffold of the OTHER HDAC1/2 corepressor complex, a 5' fusion partner in infant B-ALL and in myelodysplasia, and present in three MLL-fusion AP-MS columns but not the index fusion's
372. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     NCBI esearch via scripts/pubmed_safe.py on 2026-09-24. SIN3A[tiab] AND MLL[tiab] -> 2 (PMIDs 22777353, 22231400). SIN3A[tiab] AND KMT2A[tiab] -> 2 (PMIDs 31221981, 29158550). SIN3A[tiab] AND leukemia[tiab] -> 21. SIN3A[tiab] AND NUTM1[tiab] -> 1 (41878999). Positive control MEN1[tiab] AND MLL[tiab] -> 55.

     · NCBI PubMed E-utilities 2026
     used by SIN3A (Q96ST3) - the scaffold of the OTHER HDAC1/2 corepressor complex, a 5' fusion partner in infant B-ALL and in myelodysplasia, and present in three MLL-fusion AP-MS columns but not the index fusion's
373. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL database version 34, target CHEMBL4506 (NAD-dependent protein deacetylase sirtuin-1): target_search, get_bioactivity and get_mechanism queries

     EMBL-EBI ChEMBL · ChEMBL 2026
     used by SIRT1 - not a fusion partner, the least required gene in KMT2A-rearranged lines, and the one target here where the direction of the drug is genuinely disputed
374. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB entries 4I5I, 4ZZJ, 4ZZH, 5BTR, 4IF6, 4KXQ and 4IG9, with per-entity UniProt accessions and modelled residue ranges resolved against the RCSB search and data APIs

     RCSB Protein Data Bank · RCSB PDB 2026
     used by SIRT1 - not a fusion partner, the least required gene in KMT2A-rearranged lines, and the one target here where the direction of the drug is genuinely disputed
375. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     GTEx v8 median gene expression, SIT1 (ENSG00000137078.8) across 54 tissues: Cells_EBV-transformed_lymphocytes 104.84 TPM, whole-body median 0.33 TPM. Fetched 2026-09-23

     · GTEx Portal 2026
     used by SIT1 - one of only three surface proteins in the screen that is both expressed in KMT2A::AFF1 lines and absent from normal parenchyma, and it is clean because it is leukocyte-restricted, not because it is leukaemic
376. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     HGNC records: SIT1 (signaling threshold regulating transmembrane adaptor 1, ENSG00000137078, Q9Y3P8) and SLC6A20 (solute carrier family 6 member 20, ENSG00000163817, Q9NP91, alias SIT1). Fetched 2026-09-23.

     · HGNC 2026
     used by SIT1 - one of only three surface proteins in the screen that is both expressed in KMT2A::AFF1 lines and absent from normal parenchyma, and it is clean because it is leukocyte-restricted, not because it is leukaemic
377. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     Q143 - a filter-free promoter ranking recovers forty of the fifty published oncoprotein target genes

     Iqbal Lab oncofusion knowledge graph, beat Q143 · this project 2026
     used by Six gene bodies are AFF1-exclusive and recurrent at gene scale; the authors' own pipeline confirms one of them, refutes three and has never measured two - and IRX1 is the one it confirms
378. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     Two-axis target triage: KMT2A-rearranged selectivity crossed with demonstrated ligandability

     · Oncofusion knowledge graph; DepMap via Open Targets + RCSB PDB 2026
     used by Six of 85 proteins are both selectively required and drugged — and MEN1 is not close to the others
379. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB REST API (gene_exact query with primary gene name read back) and DepMap CRISPR (Chronos) gene effect via the Open Targets Platform GraphQL API, with KMT2A-rearrangement status from Cellosaurus; resolved for this beat on 2026-09-18

     UniProt Consortium; Broad Institute DepMap; Cellosaurus; Open Targets Platform; analysis run by this project · Databases and derived analysis 2026
     used by SKIC8 (formerly WDR61) — the last PAF1C subunit, and a gene-symbol trap worth recording
380. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P63208 (SKP1_HUMAN), S-phase kinase-associated protein 1

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SKP1 — a pan-essential housekeeping adaptor that is also a structural subunit of PRC1.1, which makes it the one PRC1.1 member that can never be a target
381. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch record count for (MLL OR KMT2A) AND (SKP1 OR SKP1A OR OCP2 OR TCEB1L), run for this beat

     · NCBI PubMed, queried 2026-09-19 2026
     used by SKP1 — a pan-essential housekeeping adaptor that is also a structural subunit of PRC1.1, which makes it the one PRC1.1 member that can never be a target
382. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct REST findInteractions/Q13309 (SKP2), 365 records, 2 against a core bait, fetched 2026-09-21

     · EMBL-EBI IntAct 2026
     used by SKP2 - the substrate receptor that grips the retained fragment; two nodes in this graph already pointed at it as a free-text string with nothing to point to
383. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q13309 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (two location values from one paper, plus an NLS motif at 67-73)

     UniProt Consortium · UniProtKB 2026
     used by SKP2 - the substrate receptor that grips the retained fragment; two nodes in this graph already pointed at it as a free-text string with nothing to point to
384. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P51531 (SMCA2_HUMAN)

     UniProt Consortium · UniProtKB 2026
     used by SMARCA2 (BRM) - the paralogue nobody measured, and the reason a DepMap rank cannot predict what a BRG1/BRM inhibitor does
385. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.reviewPMID 40181550
     Therapeutic targeting of chromatin alterations in leukemia and solid tumors.

     Perner F, Berg T, Sasca D, Mersiowsky SL, et al. · Int J Cancer 2026
     used by SMARCA2 (BRM) - the paralogue nobody measured, and the reason a DepMap rank cannot predict what a BRG1/BRM inhibitor does
386. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P51531 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (one location value plus a damage-induced relocalisation note)

     UniProt Consortium · UniProtKB 2026
     used by SMARCA2 (BRM) - the paralogue nobody measured, and the reason a DepMap rank cannot predict what a BRG1/BRM inhibitor does
387. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P51532 (SMCA4_HUMAN)

     UniProt Consortium · UniProtKB 2026
     used by SMARCA4 (BRG1) - a tumour suppressor by mutation and a maintenance requirement by knockdown, and in KMT2A-rearranged lines the dependency data say it is not selective at all
388. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed entry P51532 (SMCA4_HUMAN): sequence length and domain, region and motif feature table with Pfam cross-references, fetched live from the UniProt REST API

     UniProt Consortium · UniProtKB 2026
     used by SMARCA4 (BRG1) - a tumour suppressor by mutation and a maintenance requirement by knockdown, and in KMT2A-rearranged lines the dependency data say it is not selective at all
389. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Complex-level enrichment over the MLL-fusion AP-MS interactome, recomputed with three alternative backgrounds: BAF/SWI-SNF 6 of 17 found, 0.82 expected, p = 9.44e-05, q_BH = 1.89e-04, q at nuclear background (5,754) = 0.0756, q at recovered background (4,600) = 0.1859; per-bait decomposition over seven MLL-fusion baits

     Oncofusion knowledge graph, analyses/mll-fusion-interactome-complexes/run.py, over the AP-MS network of PMID 29777171 (network size 960) · Oncofusion knowledge graph (internal analysis) 2026
     used by SMARCA4 (BRG1) - a tumour suppressor by mutation and a maintenance requirement by knockdown, and in KMT2A-rearranged lines the dependency data say it is not selective at all
390. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q12824 (SNF5_HUMAN)

     UniProt Consortium · UniProtKB 2026
     used by SMARCB1 (BAF47/INI1/SNF5) - zero entries, zero mentions and, on the evidence, correctly so: it is lost in primary AML rather than required
391. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q12824 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (a single location value with an empty evidence list)

     UniProt Consortium · UniProtKB 2026
     used by SMARCB1 (BAF47/INI1/SNF5) - zero entries, zero mentions and, on the evidence, correctly so: it is lost in primary AML rather than required
392. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q96GM5 (SMRD1_HUMAN)

     UniProt Consortium · UniProtKB 2026
     used by SMARCD1 (BAF60A) - rank 59 of 17,728, survives the lineage split, and had never been named in this graph or studied in this disease
393. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q96GM5 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (one location value on one co-purification paper plus a curator inference)

     UniProt Consortium · UniProtKB 2026
     used by SMARCD1 (BAF60A) - rank 59 of 17,728, survives the lineage split, and had never been named in this graph or studied in this disease
394. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q92925 (SMRD2_HUMAN)

     UniProt Consortium · UniProtKB 2026
     used by SMARCD2 (BAF60B) - named in one KMT2A paper in the whole of PubMed, and the dependency it carries is myeloid identity, not the rearrangement
395. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q14683 (SMC1A_HUMAN), Structural maintenance of chromosomes protein 1A, 1233 aa, primary gene name SMC1A

     · UniProt Consortium 2026
     used by SMC1A - a bounded absence written with its query: eleven records, none about this disease, and the whole literature is Cornelia de Lange syndrome
396. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed esearch, db=pubmed, term '(MLL OR KMT2A) AND SMC1A', run 2026-09-20; count 11; all 11 titles read from a full efetch record with itertext and judged off-topic

     · NCBI PubMed 2026
     used by SMC1A - a bounded absence written with its query: eleven records, none about this disease, and the whole literature is Cornelia de Lange syndrome
397. weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paperPMID 42591212
     Case Report: Immune-driven clonal selection underlying lineage switch from B-Precursor acute lymphoblastic leukemia to acute myeloid leukemia following inotuzumab ozogamicin.

     Kondo K, Sadato D, Najima Y, Itabashi T, Ueda T, Hirama C, et al. · Front Immunol 2026
     used by SMC1A - a bounded absence written with its query: eleven records, none about this disease, and the whole literature is Cornelia de Lange syndrome
398. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9UQE7 (SMC3_HUMAN), Structural maintenance of chromosomes protein 3, 1217 aa, primary gene name SMC3

     · UniProt Consortium 2026
     used by SMC3 - the cohesin ATPase arm, the only subunit whose partial loss has been tested directly against a KMT2A fusion, and it was tolerated
399. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9H7B4 (SMYD3_HUMAN) - reviewed human record, length 428 aa

     UniProt Consortium · UniProtKB 2026
     used by SMYD3 - a lysine methyltransferase with three claimed histone marks and one robust non-histone substrate, sitting on an HOXA9 feed-forward loop in MLL-AF9 AML and a fatty-acid-oxidation axis in CML stem cells, with zero interventional trials
400. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry
     ClinicalTrials.gov intervention search, run 2026-09-22

     US National Library of Medicine · ClinicalTrials.gov 2026
     used by SMYD3 - a lysine methyltransferase with three claimed histone marks and one robust non-histone substrate, sitting on an HOXA9 feed-forward loop in MLL-AF9 AML and a fatty-acid-oxidation axis in CML stem cells, with zero interventional trials
401. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB - O14508, SOCS2_HUMAN, suppressor of cytokine signaling 2

     · UniProt Knowledgebase 2026
     used by SOCS2 (O14508) - a MENIN-INHIBITOR-SELECTIVE member of the signature, a cullin-5 E3 substrate receptor, and its likeliest upstream regulator is MEF2C rather than ATF4
402. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P17947 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (a single location value on one sequence rule plus one experimental paper)

     UniProt Consortium · UniProtKB 2026
     used by SPI1 / PU.1 — a myeloid dependency that a KMT2A panel mistakes for its own
403. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P10451 (OSTP_HUMAN) - Osteopontin

     · UniProtKB 2026
     used by SPP1 (osteopontin) - the endosteal matrix protein whose knockdown prolongs survival in MLL-AF9 mice, and the one KMT2A::AFF1-specific paper about it cannot be read
404. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q7Z699 - Sprouty-related, EVH1 domain-containing protein 1 (Homo sapiens), reviewed

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by SPRED1 - the RAS/MAPK brake whose deletion marks relapse in paediatric B-ALL, third-widest margin in the Q152 AFF1-only core
405. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     UniProtKB Q6ZRS2 (SRCAP_HUMAN), Chromatin remodeling protein SRCAP, 3230 aa

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SRCAP - the other H2A.Z-loading motor, whose entire PubMed intersection with MLL is five neurodevelopmental gene panels
406. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P36956 (SRBP1_HUMAN), primary gene name SREBF1, length 1,147 aa, processing and subcellular location, fetched 2026-09-21

     · UniProtKB 2026
     used by SREBF1 / SREBP-1 - the weakest of the three MEN1 sterol partners, because it is simultaneously the pathway's readout
407. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P36956 (SRBP1_HUMAN), SUBCELLULAR LOCATION comments for the precursor, the processed form and two constitutively active isoforms, plus the topology feature table with evidence codes and sources, fetched live 2026-09-22

     UniProt Consortium · Databases 2026
     used by SREBF1 / SREBP-1 - the weakest of the three MEN1 sterol partners, because it is simultaneously the pathway's readout
408. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q15532 (SSXT_HUMAN)

     UniProt Consortium · UniProtKB 2026
     used by SS18 - a SWI/SNF subunit that is itself a fusion partner in another cancer, and a lineage-independent dependency in KMT2A-rearranged lines
409. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q08945 (SSRP1_HUMAN), FACT complex subunit SSRP1

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SSRP1 - the small FACT subunit, and the only protein in this beat that binds a confirmed KMT2A-fusion interactor directly (LEDGF/p75, via HMG box to PWWP)
410. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch record counts for SSRP1 and the FACT complex against KMT2A, MLL and leukaemia, run for interactome round 5

     · NCBI PubMed, queried 2026-09-19 2026
     used by SSRP1 - the small FACT subunit, and the only protein in this beat that binds a confirmed KMT2A-fusion interactor directly (LEDGF/p75, via HMG box to PWWP)
411. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     St. Jude Cloud Genomics Platform — Getting Started / Overview documentation

     St. Jude Children's Research Hospital · docs.stjude.cloud 2026
     used by St. Jude Cloud and the Pediatric Cancer Genome Project — raw paediatric leukaemia genomes a small lab can get without a collaboration
412. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8N3U4 (STAG2_HUMAN), cohesin subunit SA-2, 1,231 aa; and Q8WVM7 (STAG1_HUMAN), cohesin subunit SA-1, 1,258 aa

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by STAG2 - the cohesin subunit that defines secondary AML, and the one whose loss a KMT2A fusion does not care about
413. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8N3U4 (STAG2_HUMAN) - SUBCELLULAR LOCATION comment block with evidence codes

     · UniProt Consortium 2026
     used by STAG2 - the cohesin subunit that defines secondary AML, and the one whose loss a KMT2A fusion does not care about
414. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB - O76061, STC2_HUMAN, stanniocalcin-2

     · UniProt Knowledgebase 2026
     used by STC2 / stanniocalcin-2 (O76061) - the second largest effect in the signature (-4.40 log2), a SECRETED glycoprotein, and its localisation rests on curator inference alone
415. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     XL MLL plus - Break Apart Probe (product documentation, D-5060-100-OG)

     MetaSystems Probes · MetaSystems Probes product catalogue 2026
     used by Step 4: the split-signal FISH nucleus - one fusion, one green, one red
416. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
     A Phase 1, Open-label Study of the Absorption, Metabolism, and Excretion of DSP-5336 Following a Single Oral Dose in Patients With Refractory or Relapsed Acute Leukemia, Myelodysplastic Syndrome, Multiple Myeloma, or Myeloproliferative Neoplasms

     Sumitomo Pharma America, Inc. · ClinicalTrials.gov 2026
     used by Sumitomo Pharma — enzomenib (DSP-5336): one trial, adolescents at 12 years and 40 kg, no paediatric programme
417. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct REST findInteractions/P63165 (SUMO1), 542 records, 4 against a core bait, fetched 2026-09-21

     · EMBL-EBI IntAct 2026
     used by SUMO1 - not a complex partner of LEDGF but a modifier of it, and the interactome row cannot tell you which
418. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P63165 SUBCELLULAR LOCATION comments, evidence codes and feature table, fetched live 2026-09-21 (six location values, one of them ECO:0000250 transferred from SUMO2)

     UniProt Consortium · UniProtKB 2026
     used by SUMO1 - not a complex partner of LEDGF but a modifier of it, and the interactome row cannot tell you which
419. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch record counts for the FACT subunits against KMT2A, MLL and the super elongation complex, run for interactome round 5

     · NCBI PubMed, queried 2026-09-19 2026
     used by SUPT16H (SPT16) - the large FACT subunit: pan-essential, an anti-dependency in KMT2A-rearranged lines, and drugged in MLL-r leukaemia by a compound that works through p53
420. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9Y5B9 (SP16H_HUMAN), FACT complex subunit SPT16

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SUPT16H (SPT16) - the large FACT subunit: pan-essential, an anti-dependency in KMT2A-rearranged lines, and drugged in MLL-r leukaemia by a compound that works through p53
421. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P63272 (SPT4H_HUMAN), Transcription elongation factor SPT4

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SUPT4H1 (SPT4) — zero papers linking it to KMT2A, and the 138th most KMT2A-selective gene out of 17,728
422. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch record count for (MLL OR KMT2A) AND (SUPT4H1 OR SPT4 OR SUPT4H), run for this beat

     · NCBI PubMed, queried 2026-09-19 2026
     used by SUPT4H1 (SPT4) — zero papers linking it to KMT2A, and the 138th most KMT2A-selective gene out of 17,728
423. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O00267 (SPT5H_HUMAN), Transcription elongation factor SPT5

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by SUPT5H (SPT5) — the pausing factor the KMT2A field has never studied, and the counted absence that undercuts 'the fusion hijacks pause release'
424. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch record counts for SUPT5H/SPT5/DSIF co-occurrence with KMT2A, MLL and MLL-fusion terms, run for this beat

     · NCBI PubMed, queried 2026-09-19 2026
     used by SUPT5H (SPT5) — the pausing factor the KMT2A field has never studied, and the counted absence that undercuts 'the fusion hijacks pause release'
425. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q7KZ85 (SPT6H_HUMAN), SUBCELLULAR LOCATION comments with evidence codes and sources, fetched live 2026-09-22

     UniProt Consortium · Databases 2026
     used by SUPT6H (SPT6) - the histone chaperone that rides with RNA polymerase II, which puts it closer to the fusion's proposed mechanism than anything else in this beat
426. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O43463 (SUV91_HUMAN) - Histone-lysine N-methyltransferase SUV39H1, Homo sapiens; with HGNC:11479

     UniProt Consortium; HUGO Gene Nomenclature Committee · UniProt Knowledgebase / HGNC 2026
     used by SUV39H1 (O43463, KMT1A) - the H3K9me3 writer: a menin effector whose binding region OVERLAPS the drug pocket at its N-terminus, a tumour suppressor in MLL-AF9 AML, recruited to double-strand breaks by CTCF, and never localised at the KMT2A locus in any cell type
427. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.reviewPMID 41443386
     Polycomb repressive complexes as therapeutic targets in hematologic malignancies.

     Yamagishi M, Iwama A, Kitabayashi I · Exp Hematol 2026
     used by SUZ12 - the scaffold that holds PRC2 together and decides which PRC2 it is, required in this disease and studied in it exactly once, Tazemetostat - an approved EZH2 inhibitor, and the drug this disease's own biology predicts will not work, Valemetostat - an approved dual EZH1/EZH2 inhibitor that has already worked in an MLL-AF4 xenograft, in a disease nobody has enrolled it in
428. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed esearch count and title check for SUZ12 AND (KMT2A OR MLL) AND leukemia, run 2026-09-20

     · NCBI PubMed 2026
     used by SUZ12 - the scaffold that holds PRC2 together and decides which PRC2 it is, required in this disease and studied in it exactly once
429. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed human record Q15022 (SUZ12_HUMAN), resolved with gene_primary and length requested in fields

     · UniProt 2026
     used by SUZ12 - the scaffold that holds PRC2 together and decides which PRC2 it is, required in this disease and studied in it exactly once
430. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O60506 (HNRPQ_HUMAN), human SYNCRIP, 623 aa

     · UniProt Consortium 2026
     used by SYNCRIP (hnRNP Q) — required for MLL leukaemia stem cells in one paper and a tumour suppressor deleted in T-ALL in another
431. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     REVUFORJ (revumenib) US prescribing information, INDICATIONS AND USAGE, retrieved via the openFDA drug label API

     US Food and Drug Administration; Syndax Pharmaceuticals, Inc. · openFDA drug label API 2026
     used by Syndax Pharmaceuticals — revumenib (SNDX-5613), the only menin inhibitor licensed down to infancy
432. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry
     ClinicalTrials.gov registry records for NCT06226571, NCT07211958, NCT06177067, NCT03013998 and NCT05360160, retrieved via the ClinicalTrials.gov API v2 (fields: MinimumAge, MaximumAge, OverallStatus, Phase, LeadSponsorName, CollaboratorName)

     Syndax Pharmaceuticals; St. Jude Children's Research Hospital; Beat AML, LLC; MD Anderson Cancer Center · ClinicalTrials.gov 2026
     used by Syndax Pharmaceuticals — revumenib (SNDX-5613), the only menin inhibitor licensed down to infancy
433. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
     DS-1594b With or Without Azacitidine, Venetoclax, or Mini-HCVD for the Treatment of Relapsed or Refractory Acute Myeloid Leukemia or Acute Lymphoblastic Leukemia

     · ClinicalTrials.gov 2026
     used by Systematic enumeration of menin-inhibitor resistance mechanisms: seven entries written, thirteen candidates rejected with reasons
434. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q92804 (RBP56_HUMAN) - sequence, feature table and compositional bias; retrieved live 2026-09-24 by q173

     · UniProt Consortium 2026
     used by TAF15 (RBP56, TAF2N) - the FET-family, prion-like partner that donates a low-complexity N-terminus to ZNF384
435. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB REST search executed live 2026-09-24 by q173: query=gene_exact:TAF15 AND organism_id:9606 AND reviewed:true, fields=accession,id,length,protein_name,gene_names

     · UniProt Consortium 2026
     used by TAF15 (RBP56, TAF2N) - the FET-family, prion-like partner that donates a low-complexity N-terminus to ZNF384
436. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     gnomAD v4 loss-of-function constraint (pLI, LOEUF, observed and expected LoF counts) via the gnomAD GraphQL API, joined to DepMap CRISPR gene effect via Open Targets; analysis in analyses/target-safety-triage-v2/

     Genome Aggregation Database; Broad Institute DepMap; Open Targets Platform; analysis run by this project · Databases and derived analysis 2026
     used by Target safety triage v2: constraint with its uncertainty attached (run)
437. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     In-silico junction-peptide detectability analysis for KMT2A fusions (this graph)

     · Oncofusion Knowledge Graph 2026
     used by Targeted PRM for the KMT2A::AFF1 junction peptide - the experiment that makes every other attribution claim testable, The protein nobody can attribute by antibody is the one mass spectrometry finds most easily
438. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed esearch count and title check for tazemetostat AND (KMT2A OR MLL), run 2026-09-20

     · NCBI PubMed 2026
     used by Tazemetostat - an approved EZH2 inhibitor, and the drug this disease's own biology predicts will not work
439. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9UL17 - T-box transcription factor TBX21 (Homo sapiens), reviewed

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by TBX21 (T-bet) - the Th1/NK master regulator, the third lymphoid gene in the Q152 AFF1-only core and the one that makes the B-lineage reading fail
440. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct molecular interaction database (EMBL-EBI), PSICQUIC MITAB 2.7, pubid:35140242, retrieved 2026-09-23

     2026
     used by TBXT / brachyury - the notochord master regulator, and one of only two of 102 transcription-factor baits whose neighbourhood contained DOT1L
441. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI PubMed esearch via scripts/pubmed_safe.py, run 2026-09-23; querytranslations recorded: ("TBXT"[Title/Abstract] OR "brachyury"[Title/Abstract]) AND "KMT2A"[Title/Abstract] and the same with "MLL"[Title/Abstract]

     2026
     used by TBXT / brachyury - the notochord master regulator, and one of only two of 102 transcription-factor baits whose neighbourhood contained DOT1L
442. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Human Protein Atlas v25.1, proteinatlas.json.gz, records for TBXT (ENSG00000164458) and DOT1L (ENSG00000104885), retrieved 2026-09-23

     2026
     used by TBXT / brachyury - the notochord master regulator, and one of only two of 102 transcription-factor baits whose neighbourhood contained DOT1L
443. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P15923 (TFE2_HUMAN): reviewed sequence, length and positional feature annotations

     UniProt Consortium · UniProtKB/Swiss-Prot 2026
     used by TCF3 (E2A) — domain architecture and residue coordinates, TCF3::ZNF384 - the partner whose contribution is now actually demonstrated: steroid resistance, and it is not shared with EP300::ZNF384
444. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     InterPro/Pfam/CATH-Gene3D/SUPERFAMILY domain matches for UniProt P15923

     InterPro Consortium; EMBL-EBI · InterPro 2026
     used by TCF3 (E2A) — domain architecture and residue coordinates
445. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     PDB 2MH0 'Solution NMR structure of the p300 Taz2:ETAD1 complex' and 6MGN 'mouse Id1 (51-104) - human hE47 (348-399) complex' (titles retrieved and verified at RCSB)

     RCSB Protein Data Bank (entry metadata retrieved and title verified) · RCSB PDB 2026
     used by TCF3 (E2A) — domain architecture and residue coordinates
446. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Ensembl transcript ENST00000262965 (TCF3): exon structure and coding-sequence coordinates, used to convert exon breakpoints into residue numbers

     Ensembl; EMBL-EBI · Ensembl REST API 2026
     used by TCF3 (E2A) — domain architecture and residue coordinates
447. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     AlphaFold Protein Structure Database entry for P15923 (model AF-P15923-F1)

     Jumper J, Varadi M, et al.; EMBL-EBI/DeepMind · AlphaFold DB 2026
     used by TCF3 (E2A) — domain architecture and residue coordinates
448. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     SIFTS UniProt residue-level mappings for PDB entries 6MGN, 3U5V, 2MH0 (construct ranges read back from the PDBe mappings API)

     PDBe / SIFTS; EMBL-EBI · PDBe SIFTS 2026
     used by TCF3 (E2A) — domain architecture and residue coordinates
449. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     RCSB PDB search by UniProt accession: Q16534 (HLF) returns zero entries; P15923 (TCF3) returns 2MH0, 2YPA, 2YPB, 3U5V, 6MGN

     RCSB Protein Data Bank · RCSB PDB 2026
     used by TCF3::HLF (E2A-HLF), t(17;19)(q22;p13)
450. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct REST findInteractions/21103407 (Mahmoudi et al. 2010), 42 records across mouse and human arms, fetched 2026-09-21

     · EMBL-EBI IntAct 2026
     used by TCF7L2 - the protein IntAct's 'DOT1L-TCF4' row is actually about, and the Wnt effector that DOT1L and AF10 coactivate, The DOT1L partner IntAct calls TCF4 is TCF7L2 - the second UTP18/WDR5-class symbol mis-assignment found in this interactome, and gene-tcf4 is declined
451. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q9NQB0, P15884, Q924A0 and Q60722 - primary gene names, synonyms and lengths, fetched 2026-09-21

     · UniProtKB 2026
     used by TCF7L2 - the protein IntAct's 'DOT1L-TCF4' row is actually about, and the Wnt effector that DOT1L and AF10 coactivate, The DOT1L partner IntAct calls TCF4 is TCF7L2 - the second UTP18/WDR5-class symbol mis-assignment found in this interactome, and gene-tcf4 is declined
452. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct REST findInteractions/Q9NQB0 (TCF7L2), 410 records, 1 against a core bait, fetched 2026-09-21

     · EMBL-EBI IntAct 2026
     used by TCF7L2 - the protein IntAct's 'DOT1L-TCF4' row is actually about, and the Wnt effector that DOT1L and AF10 coactivate
453. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB Q9NQB0 (TF7L2_HUMAN) SUBCELLULAR LOCATION comment, fetched 2026-09-21

     · UniProt Consortium 2026
     used by TCF7L2 - the protein IntAct's 'DOT1L-TCF4' row is actually about, and the Wnt effector that DOT1L and AF10 coactivate
454. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     Human Protein Atlas v25.1, subcellular (immunofluorescence) and Blood Atlas records for TCF7L2, fetched 2026-09-21

     · Human Protein Atlas 2026
     used by TCF7L2 - the protein IntAct's 'DOT1L-TCF4' row is actually about, and the Wnt effector that DOT1L and AF10 coactivate
455. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     NCBI PubMed esearch, bounded-absence query (counted negative)

     · NCBI E-utilities 2026
     used by TCOF1 / treacle (Q13428) - the Treacher Collins nucleolar phosphoprotein in the MLL-fusion nucleolar block, the block's only link to rDNA damage signalling through NBS1, and the only member whose loss causes a named human developmental syndrome
456. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     UniProtKB - Q13428 (TCOF_HUMAN), Treacle protein

     · UniProt Knowledgebase 2026
     used by TCOF1 / treacle (Q13428) - the Treacher Collins nucleolar phosphoprotein in the MLL-fusion nucleolar block, the block's only link to rDNA damage signalling through NBS1, and the only member whose loss causes a named human developmental syndrome
457. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O95551 (TYDP2_HUMAN), Tyrosyl-DNA phosphodiesterase 2, 362 aa, primary gene name TDP2

     · UniProt Consortium 2026
     used by TDP2 - the protein-adduct remover named in this graph's own mechanism entry and until now having no entry of its own
458. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed esearch, db=pubmed, term '(MLL OR KMT2A) AND TDP2', run 2026-09-20; count 1; querytranslation contains no [UID] artefact

     · NCBI PubMed 2026
     used by TDP2 - the protein-adduct remover named in this graph's own mechanism entry and until now having no entry of its own
459. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     PubMed E-utilities esearch, live queries with querytranslation recorded and operand controls run through the identical query shape

     · NCBI Entrez 2026
     used by Ten affinity measurements for this contact and not one uses a fusion protein, full-length KMT2A or a nucleosome
460. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI ClinVar esummary and dbSNP Variation Services refsnp records, resolved 2026-09-20

     · NCBI ClinVar / dbSNP 2026
     used by Ten of menin's 33 IntAct partners come from one BRET screen for MUTATION-created interactions — and five are gains caused by a named allele
461. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     NCBI GEO GSE260742, GSM8123359 / GSM8123361 (MLLAF9 FLAG ChIP-seq, DMSO and VTP50469 500 nM 48 h), re-binned to 2 kb from the deposited igvtools .tdf

     · NCBI Gene Expression Omnibus 2026
     used by Tested in an independent deposit with an epitope-tagged fusion: of the pivotal paper's three named losers under a menin inhibitor, MEIS1 and JMJD1C reproduce and MEF2C does not; of its three named retained loci, the HOXA cluster reproduces and MYB does not - and GAPDH, the negative control, loses half its share too, The residual fraction this graph says is absent from the entire literature, measured: FLAG-MLLAF9 enrichment at its own top 1,228 sites falls from 15.9x the genome median to 3.0x under VTP50469 - about 18% residual - and 12.8% of those sites stay in the top 0.1%
462. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     pattern-audit.py - every toolchain pattern tested against a case it must NOT match

     · Oncofusion knowledge graph (internal tooling) 2026
     used by Testing every pattern against the sentence it must not match
463. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     analyses/are-measurements-in-fields-or-in-prose/run.py and results.json, this repository, computed 2026-09-24 against the committed node files

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by Testing my own diagnosis from last cycle: 809 of 1,052 entries that state a quantity in prose already carry it as typed data - 76.9% - so the corpus was machine-readable, the detector was simply weak, and the reporting field I imposed on every future beat rests on a different and narrower justification than the one I gave
464. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     analyses/does-every-citation-point-at-the-paper-it-claims/run.py and results.json, this repository, computed 2026-09-23 against the committed node files and the live-title cache in analyses/what-this-corpus-actually-cites/citation-index.json

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by Testing whether every citation points at the paper it claims: 314 scoreable blocks, 242 scoring below 0.3 on title overlap, and every low scorer read is CORRECT - because this project's citation title is a descriptive sentence by house style, which means a PMID pointing at the wrong paper is undetectable by any automated check over the corpus alone
465. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q8NFU7 (TET1_HUMAN), Methylcytosine dioxygenase TET1 / CXXC-type zinc finger protein 6 / Leukemia-associated protein with a CXXC domain

     UniProt Consortium · UniProtKB 2026
     used by TET1 (LCX / CXXC6) - a CXXC reader that is also a KMT2A fusion partner, and whose leukaemic fusion throws its own CXXC domain away
466. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB Q6N021 (TET2_HUMAN), Methylcytosine dioxygenase TET2

     UniProt Consortium · UniProtKB 2026
     used by TET2 is NOT a CXXC protein - its CXXC domain is a separate gene, and it is DEPLETED in KMT2A-rearranged AML rather than enriched
467. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 42187108
     Characteristics and Prognostic Implications in Newly Diagnosed KMT2Ar AML: A Multicenter Study of the ECLA Group.

     Pan J, Fu C, et al. · Am J Hematol 2026
     used by TET2 is NOT a CXXC protein - its CXXC domain is a separate gene, and it is DEPLETED in KMT2A-rearranged AML rather than enriched
468. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O43151 (TET3_HUMAN), Methylcytosine dioxygenase TET3

     UniProt Consortium · UniProtKB 2026
     used by TET3 - the CXXC reader whose domain prefers 5-carboxylcytosine, and the only CXXC-family gene ever found mutated in a KMT2A-rearranged cohort
469. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     entry-novelty-audit/run.py: pairwise IDF-weighted overlap over every entry title, expert summary and evidence claim in the graph

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The 'never duplicate an entry' rule has held across 1,761 entries - and the audit that checked it would not have caught the near-duplicate that prompted it
470. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Compartment audit of the Skucha 2018 MLL-AF4 prey classes against live UniProt, run 2026-09-24 by Q180. Script analyses/q180-is-af4-specificity-a-depth-artefact/compartment_of_the_tail.py, output compartment_of_the_tail.json; 152 accessions resolved of 152 requested.

     · This repository (script and JSON committed) plus UniProtKB REST 2026
     used by The 110 MLL-AF4-specific preys are 14.5% mitochondrial and 34% nuclear; the 42 pan-fusion preys are 2.4% mitochondrial and 88% nuclear — Fisher p = 0.042 on a nuclear-extract pulldown
471. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     Q152 pre-registration, written before computation

     Iqbal Lab oncofusion knowledge graph, beat Q152 · this project 2026
     used by The 17-promoter AFF1-only recurrent core is not more recurrence than a label permutation of the same matrix gives
472. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Cellosaurus CVCL_0064 (MV4-11)

     · Cellosaurus, SIB 2026
     used by The 2015 paper did not discover the KMT2A acetylation site - it inherited it from the 2009 Choudhary acetylome, which was run in a cell panel that includes MV4-11, a KMT2A::AFF1 line
473. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     science.org access attempt, 2026-09-24

     · this project 2026
     used by The 2015 paper did not discover the KMT2A acetylation site - it inherited it from the 2009 Choudhary acetylome, which was run in a cell panel that includes MV4-11, a KMT2A::AFF1 line
474. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     DOT1L/H3K79 literature recall, re-measured after narrowing the query to leukaemia

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The 3.7% was mostly the query: DOT1L recall is 19.6% once you ask about leukaemia
475. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     finding-q151-the-two-antibody-2x-occupancy-rule-calls-a-tenth-of-all-promoters-occupied (this graph)

     Iqbal Lab oncofusion knowledge graph, beat Q152 · this project 2026
     used by The 332-promoter AFF1-only set is BELOW the count a random eight-sample arm produces (null 425)
476. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Scoring same-source and different-source counter-notes against their cited abstracts, to ask whether the independence metric predicts support

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The 82.9% population scored twice as many defect-shaped notes as the same-source one, and reading six of them says the comparison is confounded by our own DepMap numbers
477. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     Counter-evidence quality sweep: adjudicating same-source counter-notes into author's-own-caveat, legitimate-by-necessity, and gap

     · Oncofusion knowledge graph (internal) 2026
     used by The 82.9% population scored twice as many defect-shaped notes as the same-source one, and reading six of them says the comparison is confounded by our own DepMap numbers
478. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     Q129 pre-registration, session scratchpad q129_prereg.md, written 2026-09-23 before data collection; text quoted in full in the 'The pre-registration, quoted' section of this entry

     · Oncofusion Knowledge Graph (internal) 2026
     used by The ABT1/KRR1/UTP3 nucleolar triple: pre-registered, then adjudicated - it is not three proteins on two baits but a seven-member ribosomal block in one screen plus one hit in a neighbourhood that shows no nucleolar enrichment, and two of the four pre-registered discriminators were void exactly as the pre-registration predicted
479. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     ana-utp18-is-not-in-the-mll-complex-it-is-wdr5-miscurated, Oncofusion Knowledge Graph node file, read 2026-09-23

     · Oncofusion Knowledge Graph (internal) 2026
     used by The ABT1/KRR1/UTP3 nucleolar triple: pre-registered, then adjudicated - it is not three proteins on two baits but a seven-member ribosomal block in one screen plus one hit in a neighbourhood that shows no nucleolar enrichment, and two of the four pre-registered discriminators were void exactly as the pre-registration predicted
480. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UCSC Genome Browser hg38 refGene, ncbiRefSeq and knownGene tracks over chr4:86,915,000-86,945,000, plus Ensembl REST gene and transcript overlaps for 4:86,915,000-86,945,000

     UCSC Genome Browser api.genome.ucsc.edu; Ensembl REST rest.ensembl.org · UCSC Genome Browser / Ensembl 2026
     used by The AFF1 promoter window holds three genes' TSSs, not two, and the third is protein-coding - Q153's 92.5% is confirmed and its two-gene framing is not
481. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     Q155 50 bp base-resolution profile across chr4:86,932,500-86,938,500 over all 133 GSE252378 libraries

     Iqbal Lab oncofusion knowledge graph, beat Q155 · this project 2026
     used by The AFF1 promoter window holds three genes' TSSs, not two, and the third is protein-coding - Q153's 92.5% is confirmed and its two-gene framing is not, The KMT2A-N signal at the AFF1 locus is bimodal, neither mode is on AFF1's TSS, and the flank test I ran first said the opposite
482. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI elink pubmed to gds, sra, biosample and bioproject for PMID 34373735

     · NCBI Entrez Programming Utilities 2026
     used by The AHD inhibitor's deposited arm is MOLM-13 and its in vivo arm is nowhere: a half-resolution that names the limit exactly
483. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     Control-locus reads of GEO GSM3714939/GSM3714940/GSM3714941 and a header read of GEO GSM6430988, with pybigtools 0.3.0 over HTTP range requests, 2026-09-22

     Q107 beat; data from Canela et al. 2019 (GSE129526) and Gray et al. 2023 (GSE210480) · NCBI GEO 2026
     used by The bigWig refusal of cycle 57 is overturned: pybigtools installs on this box under Python 3.12, reads GEO tracks by HTTP range request without downloading them, and passes the control loci the refusal itself demanded
484. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entries P51825 (AFF1_HUMAN) and Q9UHB7 (AFF4_HUMAN), accession, primary gene name, length and organism read back 2026-09-21

     · UniProtKB 2026
     used by The binary ladder for AFF1: does the index fusion's partner touch P-TEFb without AFF4, and does it touch AFF4 at all?
485. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC full-text search: BODY:"K1130" AND BODY:"leukemia" (hitCount 4), with fullTextXML retrieval of each hit

     · Europe PMC REST 2026
     used by The blind spot is closed: all four K1130-and-leukaemia full-text hits are reviews restating the source paper, and the citation-network bound survives 26 nomenclatures rather than five
486. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     analyses/q165-the-retained-acetylation-site/citation-network-nomenclature-bound.json - 26 Europe PMC CITES queries with per-term hit counts and PMIDs

     · this project; Europe PMC REST 2026
     used by The blind spot is closed: all four K1130-and-leukaemia full-text hits are reviews restating the source paper, and the citation-network bound survives 26 nomenclatures rather than five
487. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     ana-q113-every-bounded-absence-classified-by-the-instrument-that-bounded-it (this corpus)

     · Oncofusion Knowledge Graph 2026
     used by The bounded-absence population went 39 to 73 in two days, and the abstract-level defect class went from 26 of 39 to 2 of 34 (Fisher p = 1.3e-5)
488. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     scripts/check-absence-class.py, oncofusion-knowledge-graph

     · this project 2026
     used by The bounded-absence population went 39 to 73 in two days, and the abstract-level defect class went from 26 of 39 to 2 of 34 (Fisher p = 1.3e-5)
489. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     Q147 region table - per-library shares for 21 named regions including the eight width-matched negative controls

     Iqbal Lab oncofusion knowledge graph, beat Q147 · this project 2026
     used by The cannot-be-occupied floor never binds at gene-body scale in any of the 27 samples, so q147's gene-scale rule is a bare 2.0x gate - and the max-versus-95th-percentile difference between q147 and q152 is immaterial there
490. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Evidence census over the seventeen chain-member entries: 110 evidence items, 85 distinct PMIDs, fusion attribution attempted from citation titles

     · Oncofusion knowledge graph (internal measurement) 2026
     used by The chain's biggest risk is generalising a myeloid mouse result to infant B-ALL, and the graph cannot measure how much of it does
491. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Whole-pathway test of haem biosynthesis selectivity in KMT2A-rearranged lines, with ALAS1/ALAS2 as an internal isoform control

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The check I called decisive was not: the haem signal is still open (run)
492. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Europe PMC RESTful search, BODY: IDH1/IDH2 x KMT2A clause x infant/paediatric, 2026-09-22

     · Europe PMC 2026
     used by The cohort paper this graph said had certainly measured IDH status was opened: 96,537 characters of full text contain IDH1 zero times, IDH2 zero times, TET2 zero times and DNMT3A zero times, so the measured-not-reported class is confirmed by reading rather than by inference
493. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     analyses/computed-but-unread/run.py and results.json, this repository; run 2026-09-23 against the working tree

     Oncofusion knowledge graph, coordinator · internal analysis, reproducible from the repository 2026
     used by The computed-but-unread scan ran, passed its own declared informativeness control at 60.3% cited, and produced a headline nobody can act on: 36,110 uncited values across 182 results files, dominated by ranks, taxids and chromosome extents. The useful population is not 'uncited numbers' but 'PRE-REGISTERED quantities that went uncited', which is what the one real instance actually was
494. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.trial_registry
     Open trial map for KMT2A-rearranged acute leukemia, 2026

     Oncofusion Knowledge Graph, therapeutics beat · ClinicalTrials.gov sweep recorded in this graph 2026
     used by The counter-argument: the menin field is an adult field, and the infant studies were withdrawn
495. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     different-source-defect-triage: classification and reading of the 31 flagged different-source counter-notes

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The counter-note defect detector mistakes cross-referencing for unsupported assertion, which is why the different-source rate looked twice as bad
496. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     counter-note-support-by-population: support scores for same-source and different-source counter-notes under one instrument

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The counter-note defect detector mistakes cross-referencing for unsupported assertion, which is why the different-source rate looked twice as bad
497. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     cross-entry-contradiction-screen: three successive thresholds for a class-D contradiction detector, and their candidate counts

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The cross-entry contradiction screen does not work, and the reason is that this graph writes carefully
498. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41545699
     CSF1R marks a subset of foetal haematopoietic multipotent progenitor cells with acute myeloid leukaemia propagation properties

     Camiolo G, González Silvera D, Leah T, et al. (senior author Ottersbach K) · Leukemia 2026
     used by The CSF1R-positive fetal lymphoid-primed multipotent progenitor — a fetal-restricted subset of LMPPs that makes AML rather than ALL when KMT2A::MLLT3 is switched on, and exists in human embryos
499. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     breakpoint-kmt2a-major-bcr, this graph's existing entry for the KMT2A major breakpoint cluster region, read 2026-09-22; its own evidence is Meyer et al.'s recombinome series as recorded there

     · Iqbal Lab oncofusion knowledge graph 2026
     used by The CTCF site in KMT2A exon 11 - a conserved insulator element sitting exactly on the functional cut point of the breakpoint cluster region, whose occupancy decides whether the locus rearranges
500. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct molecular interaction database, PSICQUIC query for UniProt P51825 (AFF1)

     · EMBL-EBI IntAct 2026
     used by The curated interaction database knows less about AFF1 than this graph does
501. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     Bimodality of DAXX Chronos across 13 B-lineage ALL DepMap lines, with donor age and fusion identity tested as rival explanations

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The DAXX requirement in B-lineage KMT2A-rearranged lines is a telomere phenotype, not a fusion-target phenotype - and the two non-KMT2A lines in the deep group are the test, The lymphoid-restricted DAXX dependency survives three nulls and two comparator restrictions - and ATRX, MEN1 and DOT1L do not survive the same test in the same lines
502. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ENCFF251DYL, RSEM gene quantifications, preferred_default of ENCSR000CWQ (polyA plus RNA-seq, MCF-7, GRCh38, GENCODE V29), with gencode.v29.annotation.gtf.gz

     ENCODE Project Consortium; GENCODE · ENCODE Portal / EBI 2026
     used by THE DEFLATING NULL, RUN: matching on gene length AND MCF-7 expression roughly HALVES the H3K9me3 depletion over the KMT2A gene body - from the 7.6th-15.0th percentile of class-matched CTCF anchors to the 18.2nd-23.4th of 385 matched transcribed gene bodies - so the methyl-poor breakpoint domain is partly a fact about transcription and does not dissolve into one
503. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     Q147 gene-scale results - IRX1 and HOXA occupancy across six KMT2A fusion partners

     Iqbal Lab oncofusion knowledge graph, beat Q147 · this project 2026
     used by The deposit paper's own supplementary table calls IRX1 an oncoprotein target in exactly p279, p318 and p454 - the same three samples this project reached by a different instrument, Verdict: q147's IRX1 headline survives as a measurement and not as an inference - the pre-registered branch is NEITHER, on U5 and on a FALLS condition that fired
504. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     ana-the-six-new-dot1l-partners-are-nine-spokes-of-one-pulldown, this graph, cycle 42

     · EMBL-EBI IntAct 2026
     used by The DOT1L partner IntAct calls TCF4 is TCF7L2 - the second UTP18/WDR5-class symbol mis-assignment found in this interactome, and gene-tcf4 is declined, The interactome instrument is migrated to IntAct REST: 277 partners, 118 written, 42.6% - the denominator grew by 60 and coverage fell 3.5 points, which is the instrument getting honest
505. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct REST findInteractions/O60477 (BRINP1), 35 records, 1 against a core bait, fetched 2026-09-21

     · EMBL-EBI IntAct 2026
     used by The DOT1L partner IntAct calls TCF4 is TCF7L2 - the second UTP18/WDR5-class symbol mis-assignment found in this interactome, and gene-tcf4 is declined
506. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB Protein Data Bank entries 6NJ9, 6NQA, 6NOG, 6O96, 6JMA, 6JM9; identifiers, resolutions and deposited titles retrieved from the RCSB data API on 2026-09-19

     RCSB Protein Data Bank; retrieval performed by this project · Database 2026
     used by The DOT1L–ubiquitin and DOT1L–acidic-patch interfaces: two non-catalytic surfaces that MV4;11 cells cannot do without
507. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paperPMID 41142648
     A chemical-genetic interaction between PAF1 and ENL/AF9 YEATS inhibition.

     Barta PA, Garnar-Wortzel L, Bishop TR, et al. · RSC Chemical Biology 2026
     used by The ENL YEATS acetyl-reader inside KMT2A::MLLT1: what it reads, what it does, and why both control arms carry it, What this beat searched, read and set aside - including one cell line the accession database has not caught up with, and one query that returns a clean zero
508. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Paediatric investigation plans

     European Medicines Agency · EMA human regulatory overview 2026
     used by The EU paediatric investigation plan: an obligation, not an incentive - and the deferral is the loophole
509. weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.database
     Medicines for human use under evaluation / EMA medicine search (retrieval attempts, 2026-09-18)

     European Medicines Agency · EMA 2026
     used by The EU paediatric investigation plan: an obligation, not an incentive - and the deferral is the loophole
510. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     GEO sample GSM8123359, SOFT record: !Sample_title = 'MLLAF9 Control - DMSO FLAG', Mus musculus, taxid 10090, tissue Blood, treatment 3 days

     · NCBI Gene Expression Omnibus 2026
     used by The eviction quartiles are defined in the 'MLLAF9 Control' FLAG culture and the H3K79me2 they stratify is measured in the 'NTvKO Control' and 'NTvKO PCGF1 KO' cultures - zero shared libraries - so the commonest way a stratified analysis fools itself is structurally unavailable here
511. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry Q9NPD8 (UBE2T)

     · UniProtKB/Swiss-Prot 2026
     used by The FA core complex does not move together, and that is the strongest argument against reading this as a Fanconi-pathway dependency at all, UBE2T — E2; rank 33 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all
512. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Fanconi module coherence: which members behave as one unit across DepMap, and which are only annotated into it

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The Fanconi core complex is 2nd of 719 complexes for KMT2A-selectivity - but coherence does not explain it, and the explanation this entry was built on is dead
513. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Control pathways for the coherence-selectivity relationship: NOTCH, cholesterol biosynthesis and a size-matched random set

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The Fanconi core complex is 2nd of 719 complexes for KMT2A-selectivity - but coherence does not explain it, and the explanation this entry was built on is dead
514. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     Null distribution of the coherence-selectivity rho over 500 random 32-gene sets

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The Fanconi core complex is 2nd of 719 complexes for KMT2A-selectivity - but coherence does not explain it, and the explanation this entry was built on is dead
515. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Ribosome control for the coherence-selectivity relationship, run from an explicit RPL*/RPS* symbol list

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The Fanconi core complex is 2nd of 719 complexes for KMT2A-selectivity - but coherence does not explain it, and the explanation this entry was built on is dead
516. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Systematic sweep of the coherence-selectivity relationship across 14 Reactome pathways of comparable size

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The Fanconi core complex is 2nd of 719 complexes for KMT2A-selectivity - but coherence does not explain it, and the explanation this entry was built on is dead
517. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Unbiased KMT2A-rearranged selectivity scan, 17,728 protein-coding genes

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The Fanconi core complex is 2nd of 719 complexes for KMT2A-selectivity - but coherence does not explain it, and the explanation this entry was built on is dead
518. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Fanconi pathway enrichment among KMT2A-rearranged selective dependencies, 17,728-gene scan against Reactome R-HSA-6783310

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The Fanconi signal is real, and it sits in the ubiquitination module
519. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     PSICQUIC IntAct provider, current/search/query: brca2=189, Q03164=218, id:Q03164=161, mll_af9=0, EBI-2890930=32, id:EBI-2890930=32

     · EMBL-EBI PSICQUIC 2026
     used by The federated sweep this graph prescribes for the missing index-fusion interactor cannot be run: PSICQUIC free text returns ZERO for `mll_af9`, an entity that provably exists and returns 32 records under its accession - so every provider's zero in that sweep is uninterpretable
520. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     PSICQUIC registry ACTIVE list, 14 providers; Reactome provider returns 0 for brca2 while BioGrid returns 322 and IMEx 341

     · EMBL-EBI PSICQUIC Registry 2026
     used by The federated sweep this graph prescribes for the missing index-fusion interactor cannot be run: PSICQUIC free text returns ZERO for `mll_af9`, an entity that provably exists and returns 32 records under its accession - so every provider's zero in that sweep is uninterpretable
521. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     EMBL-EBI Complex Portal complex-ws search: MLL-AF4 = 0, KMT2A = 15, AFF1 = 7, MLL = 36

     · EMBL-EBI Complex Portal 2026
     used by The federated sweep this graph prescribes for the missing index-fusion interactor cannot be run: PSICQUIC free text returns ZERO for `mll_af9`, an entity that provably exists and returns 32 records under its accession - so every provider's zero in that sweep is uninterpretable
522. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct molecular interaction database, PSICQUIC MITAB 2.7, pairwise queries id:Q9UHB7 AND id:{P63010,Q08379,Q5JR59,Q12933}, retrieved 2026-09-21

     · IntAct (EMBL-EBI) 2026
     used by The four AFF4 two-hybrid partners are three rows of one screen, not three experiments - and in that same screen GOLGA2 has 249 partners while AFF4 has five
523. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     IntAct molecular interaction database, PSICQUIC MITAB 2.7, per-interactor queries with publication-level partner deduplication, retrieved 2026-09-21

     · IntAct (EMBL-EBI) 2026
     used by The four AFF4 two-hybrid partners are three rows of one screen, not three experiments - and in that same screen GOLGA2 has 249 partners while AFF4 has five
524. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     IntAct molecular interaction database, PSICQUIC MITAB 2.7, per-interactor queries partitioned by source publication, retrieved 2026-09-21

     · IntAct (EMBL-EBI) 2026
     used by The four AFF4 two-hybrid partners are three rows of one screen, not three experiments - and in that same screen GOLGA2 has 249 partners while AFF4 has five
525. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     Junction peptides for the three UniProt-annotated recurrent KMT2A breakpoints, computed in this repository from NM_001197104.2 (KMT2A) and NM_005935.4 (AFF1): all three architectures use the same AFF1 acceptor at nt 1175 and the same AFF1 head peptide QTYSNEVH

     Oncofusion Knowledge Graph, analyses/junction-peptides-all-breakpoints · Internal analysis 2026
     used by The fusion begins at AFF1 residue 347 and the cyclin T1-binding sequence ends at 308, so KMT2A::AFF1 is the one AFF-family protein that cannot enter the reservoir, The two t(4;11) products split AFF1 exactly at the P-TEFb-binding module, and the elongation half goes to the reciprocal
526. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     RCSB PDB - full-text search for KMT2A/MLL entries (verified 2026-09-17)

     · RCSB Protein Data Bank 2026
     used by The fusion junction itself as a small-molecule target: a candid verdict
527. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     The bidirectional-promoter enrichment is mostly accessibility, partly KMT2A, and not fusion-specific at all

     Iqbal Lab oncofusion knowledge graph, coordinator control run · this project 2026
     used by The fusion-biased peak at the AFF1 locus fires in 23 of 27 samples across all six partners, so it is not the fusion visiting its own partner
528. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UCSC Genome Browser API, chromosome sizes for hg19 and hg38, fetched 2026-09-23 - hg19 chr1 249,250,621 / chr2 243,199,373; hg38 chr1 248,956,422 / chr2 242,193,529

     · UCSC Genome Browser 2026
     used by The genome build of GSE127507 is hg19, verified from the chromosome extents written inside each of the nine downloaded .tdf files - 9 bp from hg19 and 2,818,723 bp from hg38 on chr1, in every file - and independently from the deposited TSS table's own coordinates
529. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     Measuring confidence-grade consistency per cited paper across the graph

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The grading rule applied to twenty citations, and ORCID tested as a clustering key and rejected: zero coverage before 2010, The grading rule: confidence grades the claim, not the source - so one paper carrying several grades is correct, and only a three-step spread is worth reading
530. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed efetch author records for 238 papers from the graph's fifteen largest research groups, tested for ORCID identifier coverage by era

     · NCBI E-utilities 2026
     used by The grading rule applied to twenty citations, and ORCID tested as a clustering key and rejected: zero coverage before 2010
531. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     NCBI efetch db=pmc for PMC4119175 and elink pubmed to gds, sra, biosample, bioproject for PMID 21726816

     · NCBI Entrez Programming Utilities 2026
     used by The H2BK34ub HOXA9/MEIS1 result is HeLa: the full text contains the word leukaemia zero times
532. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     mark-h3k9me3, this graph's existing entry, read 2026-09-22; its primary sources are recorded there as PMIDs 29774127, 33054052, 33037410, 31243293 and 25822366

     · Iqbal Lab oncofusion knowledge graph 2026
     used by The H3K9me3 null that appears to exonerate prophylactic G9a inhibition was measured in retinal pigment epithelium, at 1 kb resolution, with n = 2 - and the warning it appears to lift lives in a leukaemic blast, The one experiment that would separate the H3K9me1/2 rearrangement brake from H3K9me3 was run, deposited as eight public samples, and reported nowhere in the paper
533. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Per-gene test of the HBO1/MOZ acetyltransferase modules against a pan-essentiality control, and the paralogue split within each family

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The HBO1/MOZ modules are not selective as modules: one paralogue per family carries the signal, and no member clears both a dependency and a pan-essentiality bar
534. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.trial_registry
     ClinicalTrials.gov API v2 free-text query for HTL001 and for HXR9

     ClinicalTrials.gov (NIH/NLM); query run by this project 2026-09-20 · Trial registry 2026
     used by The HOX-PBX interface, and its one peptide: HXR9/HTL001 has been tried in eight tumour types and in KMT2A-rearranged leukaemia exactly once
535. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.reviewPMID 41436318
     HSPC-like blasts in acute lymphoblastic leukemia: biology and therapeutic opportunities

     Xu J, Kennedy VA, Ma C, et al. (senior author Tan K) · Trends in Cancer 2026
     used by The HSPC-like blast fraction — a minority of blasts that transcriptionally resemble haematopoietic stem and progenitor cells, found in the blood of the youngest KMT2A-rearranged patients and predicting drug response and outcome across leukaemia subtypes
536. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41631723
     Identification of cellular hierarchy in paediatric acute myeloid leukaemia: The Japan Children's Cancer Group trial (JCCG AML-12)

     Komori T, Kawai T, Okuno Y, et al. (senior author Shiba N) · British Journal of Haematology 2026
     used by The HSPC-like blast fraction — a minority of blasts that transcriptionally resemble haematopoietic stem and progenitor cells, found in the blood of the youngest KMT2A-rearranged patients and predicting drug response and outcome across leukaemia subtypes
537. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB reviewed human proteins carrying the zf-CXXC domain: Pfam PF02008, InterPro IPR002857, and the ft_zn_fing:CXXC-type feature query

     UniProt Consortium · UniProt REST (rest.uniprot.org) 2026
     used by The human CXXC/CpG-reading family is exactly twelve proteins, and TET2 is not one of them
538. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB REST search, gene_exact queries over 19 human surface-marker symbols with gene_primary and length requested, run 2026-09-20

     · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
     used by The immunophenotype beat's own ledger: what was enumerated, what was written, what was dropped and why, and what could not be read
539. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProt O88573 (AFF1_MOUSE), canonical sequence, 1,216 aa

     · UniProtKB 2026
     used by The index fusion DISCARDS the SIAH degron: the junction is at AFF1 residue 347, the degron is at human 282-304, and three entries had it the other way because they measured a murine motif against an exon-numbering junction
540. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB entry P51825 (AFF1_HUMAN), accession, primary gene name, length and PDB cross-references read back 2026-09-21

     · UniProtKB 2026
     used by The index fusion's 3' partner has ONE observed binary interaction in IntAct and forty-two residues of solved structure - every SEC contact attributed to AFF1 is an AFF4 result wearing AFF1's name
541. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     IntAct REST findInteractions totalElements versus PSICQUIC MITAB row counts for the six scaffold accessions

     · EMBL-EBI IntAct 2026
     used by The interactome coverage figure is computed on 651 of the 1,015 IntAct rows available - 64% - and the gap is worst for menin and LEDGF, the two baits the menin-interface hypothesis rests on
542. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI Gene esummary for GeneID 153571 (IRX2-DT), 153572 (IRX2) and 79192 (IRX1), assembly NC_000005.10

     NCBI Gene / E-utilities · NCBI Gene 2026
     used by The IRX2 / IRX2-DT geometry re-derived from three annotations: 215 bp and 89.2% under RefSeq, 437 bp and 78.2% under Ensembl, and the verdict does not move
543. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Ensembl REST API, lookup/symbol/homo_sapiens for IRX2, IRX2-DT and IRX1, GRCh38

     Ensembl REST, rest.ensembl.org · Ensembl 2026
     used by The IRX2 / IRX2-DT geometry re-derived from three annotations: 215 bp and 89.2% under RefSeq, 437 bp and 78.2% under Ensembl, and the verdict does not move
544. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB REST search: gene:C5orf38 AND organism_id:9606, and gene_exact:IRX2 AND organism_id:9606 AND reviewed:true

     UniProt Consortium · UniProtKB 2026
     used by The IRX2 / IRX2-DT geometry re-derived from three annotations: 215 bp and 89.2% under RefSeq, 437 bp and 78.2% under Ensembl, and the verdict does not move
545. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     ENCODE experiment ENCSR000BRT (CBX3 ChIP-seq, K562, Richard Myers HAIB) - file ENCFF271ZAZ, IDR ranked peaks, GRCh38

     ENCODE Project Consortium · ENCODE Portal 2026
     used by The KMT2A exon-11 CTCF anchor carries a CBX3 peak and a BRCA1 peak in three of the very ENCODE experiments Q35 called empty - the peaks are in ENCODE's own pre-IDR files, present in both replicates, and deleted by the IDR threshold at global IDR 0.227 and 0.059
546. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     ENCODE files ENCFF843VHC (K562 CTCF optimal IDR thresholded peaks) and ENCFF154APE (HeLa-S3 CTCF optimal IDR thresholded peaks), GRCh38

     ENCODE Project Consortium · ENCODE Portal 2026
     used by The KMT2A exon-11 CTCF anchor carries a CBX3 peak and a BRCA1 peak in three of the very ENCODE experiments Q35 called empty - the peaks are in ENCODE's own pre-IDR files, present in both replicates, and deleted by the IDR threshold at global IDR 0.227 and 0.059
547. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB P55197 with includeIsoform=true: canonical 1068 aa, isoform P55197-1 1027 aa, isoform-2 126 aa, isoform-3 179 aa; sequence windows compared 2026-09-19

     · UniProt Consortium 2026
     used by The KMT2A-r AML patient table numbers MLLT10 on isoform P55197-1, not the canonical sequence: S674 is S658 and S705 is S689
548. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.
     mtap-deletion-in-the-kmt2a-arm: are the 11 KMT2A-rearranged DepMap lines enriched for MTAP deletion, which would make MAT2A's selectivity rank a property of the panel?

     · This repository (analysis, reproducible from the cached scan) 2026
     used by The KMT2A-rearranged DepMap arm is not MTAP-deleted, and the gene that looked like the test cannot answer it
549. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed E-utilities esearch, db=pubmed, batch run 2026-09-20 with querytranslation inspected for every query

     NCBI Entrez Programming Utilities · NCBI E-utilities 2026
     used by The KMT2A::MLLT1 mechanism sweep: what was searched, what was read and not used, and the three retrieval traps that would have produced confident wrong answers
550. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     RCSB PDB entries 6S01, 8PC5, 8PC6, 8PEO, 8PEP, 8CBN, 8CBQ and 9XYC, with per-entity UniProt accessions, methods, resolutions and modelled residue ranges resolved against the RCSB search and data GraphQL APIs

     RCSB Protein Data Bank · RCSB PDB 2026
     used by The LEDGF PWWP–H3K36me nucleosome contact: how the fusion learns its address
551. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     ChEMBL database version 34, targets CHEMBL3132741 (Peregrin/BRPF1) and CHEMBL3988590 (PC4 and SFRS1-interacting protein): target_search, get_bioactivity and get_mechanism queries

     EMBL-EBI ChEMBL · ChEMBL 2026
     used by The LEDGF PWWP–H3K36me nucleosome contact: how the fusion learns its address
552. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     Does the lymphoid-restricted DAXX/ATRX dependency in KMT2A-rearranged leukaemia lines survive a permutation null, a stratified interaction null, a background gene set and a B-lineage-restricted comparator?

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The lymphoid-restricted DAXX dependency survives three nulls and two comparator restrictions - and ATRX, MEN1 and DOT1L do not survive the same test in the same lines
553. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     Null C: the lymphoid index-versus-comparator difference for 277 randomly drawn protein-coding genes, seed 20260924

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The lymphoid-restricted DAXX dependency survives three nulls and two comparator restrictions - and ATRX, MEN1 and DOT1L do not survive the same test in the same lines
554. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Cellosaurus cell line records CVCL_0095 (SEM), CVCL_S906 (SEM-K2), CVCL_0093 (RS4;11), CVCL_8227 (HB11;19)

     · Cellosaurus, SIB Swiss Institute of Bioinformatics 2026
     used by The lymphoid-restricted DAXX dependency survives three nulls and two comparator restrictions - and ATRX, MEN1 and DOT1L do not survive the same test in the same lines
555. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     Per-line Chronos values and complete-separation counts for DAXX, ATRX and MEN1 in the lymphoid stratum

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The lymphoid-restricted DAXX dependency survives three nulls and two comparator restrictions - and ATRX, MEN1 and DOT1L do not survive the same test in the same lines
556. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     HGNC REST fetch/symbol/KMT2A and Ensembl REST lookup/id/ENSG00000118058, both fetched 2026-09-22

     HUGO Gene Nomenclature Committee; Ensembl · HGNC / Ensembl 2026
     used by The menin arm and the CTCF arm are not two stories: menin/SETD2/H3K36me3/LEDGF licenses CtIP short-range resection and CTCF/H3K9me/HP1-gamma licenses BRCA1/BARD1-EXO1/DNA2 long-range resection, which are the two consecutive halves of ONE reaction - so a menin inhibitor and an H3K9-axis inhibitor are acting on the same pathway at different steps
557. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     analyses/q34-menin-surface-geometry/run.py - menin surface geometry over all 69 O00255 PDB deposits

     Q34-menin-surface-map · Oncofusion Knowledge Graph 2026
     used by The menin-gamma-H2A.X motif, menin Val371-Gln400: necessary, sufficient as a peptide, and the nine residues at its centre have never been seen in any of the 69 menin structures
558. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     Bounded absence query with controls for the menin-inhibitor plus PARP-inhibitor combination, PubMed esearch 2026-09-20

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The menin-inhibitor plus PARP-inhibitor experiment is genuinely unrun, but its rationale should be DNA repair, not HOXA9 - and the evidence comes from a disease where menin is a tumour suppressor
559. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     antigen-presentation-menin-polycomb: GO class I and class II gene sets from the EBI QuickGO annotation API tested against Zhou supplemental Table 4 by permutation

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The menin/Polycomb split between NLRC5 and CIITA does not extend to their pathways, and the MHC locus is structurally invisible to the method that found it
560. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     mhc-rescued-by-hgnc-primary-id: HGNC primary Ensembl ids used to join GO antigen-presentation sets against Zhou supplemental Table 4

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The MHC was never invisible - the resolver was wrong - and once the classical chains are joined the class I machinery is bivalent, not unmarked
561. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Multi-mapping exposure of gene symbols to Ensembl gene ids, this graph versus a random protein-coding background

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The MHC was never invisible - the resolver was wrong - and once the classical chains are joined the class I machinery is bivalent, not unmarked
562. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Cellosaurus records CVCL_0007 (U-937), CVCL_1844 (OCI-AML-3), CVCL_2130 (MUTZ-8) and CVCL_1425 (Mono-Mac-1), DI/OI/CC fields, retrieved via the Cellosaurus API on 2026-09-19

     Bairoch A; SIB Swiss Institute of Bioinformatics · Cellosaurus 2026
     used by The monocytic confound is worse than recorded: two of the six rearranged lines are one patient, and the four-line control group contains a HOXA-driven fusion and two NPM1c/myelomonocytic lines
563. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Cellosaurus records CVCL_1425 (Mono-Mac-1) and CVCL_1426 (Mono-Mac-6): same-origin relation and gene-fusion annotation, retrieved by accession

     Bairoch A; SIB Swiss Institute of Bioinformatics · Cellosaurus 2026
     used by The monocytic confound is worse than recorded: two of the six rearranged lines are one patient, and the four-line control group contains a HOXA-driven fusion and two NPM1c/myelomonocytic lines
564. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     Cellosaurus records CVCL_1425 (Mono-Mac-1) and CVCL_1426 (Mono-Mac-6): sequence-variation lists, omics inventory and same-origin relation

     Bairoch A; SIB Swiss Institute of Bioinformatics · Cellosaurus 2026
     used by The monocytic confound is worse than recorded: two of the six rearranged lines are one patient, and the four-line control group contains a HOXA-driven fusion and two NPM1c/myelomonocytic lines
565. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41261957
     Clinical Impact of MTAP Rearrangement in Pediatric Acute Lymphoblastic Leukemia: Insights From Whole Transcriptome Analysis.

     Zhang S, Wu Z, Liu Y, Li B, Song Y · Pediatr Blood Cancer 2026
     used by The MTAP-deleted, MTA-high state - a genotype-imposed metabolic state that this project's KMT2A-rearranged arm turns out NOT to have, and that its own screen could never have detected directly
566. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
     Supplementary key-resources table of PMID 41634341, 41556_2025_1859_MOESM6_ESM.xlsx, sheet 'Antibody', 27 rows, retrieved 2026-09-23 from the Europe PMC supplementaryFiles endpoint for PMC12904788

     · Nature Cell Biology supplementary information 2026
     used by The murine MLL1 ChIP that sits beside a FLAG-tagged fusion in GSE260742 uses Bethyl A300-086A - the identical antibody to the human MLL1n track in GSE127507 - which no GEO record in either series would tell you, and which is what makes a masking calibration against a tag possible at all
567. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProt Q13492 (PICALM_HUMAN), the reference for the CALM NES coordinate

     · UniProtKB 2026
     used by The murine-coordinate sweep is a NEGATIVE result - zero new defects in 3,086 entries - and its actual yield was a fourth entry that had the degron right before Q191 did
568. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     non-gene-antigen-blind-spot/run.py: antigens named in the graph that a HGNC-symbol queue cannot reach, by failure class

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The named-but-unwritten queue is keyed on gene symbols, so CD15 is invisible to it at 23 entries
569. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     data/symbol-collisions.json and the regenerated docs/named-but-unwritten-report.json

     · Oncofusion knowledge graph (internal) 2026
     used by The named-but-unwritten queue's top two entries were a web API and a survival endpoint, and MYC was the real answer all along
570. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 41574310
     SKIDA1 sustains MLL::ENL-expressing hematopoietic progenitors during neonatal stages and promotes B-lineage priming.

     Mendoza-Castrejon J, Yang W, et al. · Blood neoplasia 2026
     used by The neonatal window — the state in which an MLL fusion can transform, bounded at both ends
571. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paperPMID 42679029
     A Hox-dependent anchoring mechanism mediates transcriptional repression of autophagy-related genes at the nuclear periphery

     Vanderperre S, Ajuria L, Duffraisse M, Da Silva T, Vonau S, Brocard J, Forcet C, Merabet S · Proceedings of the National Academy of Sciences USA 2026
     used by The nuclear lamina and lamina-associated domains - evidence grade for KMT2A: weak and indirect, not none
572. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     ana-q147-irx1-fusion-occupancy-is-confined-to-kmt2a-aff1 (this graph)

     Iqbal Lab oncofusion knowledge graph, beat Q147 · this project 2026
     used by The occupancy rule this project has used since q145 calls a tenth of ALL promoters occupied, and no entry using it records that
573. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Ensembl REST lookup, genes ENSG00000172493 (AFF1) and ENSG00000118058 (KMT2A), assembly GRCh38, retrieved 2026-09-20

     Ensembl · Ensembl REST API 2026
     used by The other half of the break: what is actually known about the AFF1 breakpoint region, and the negative-control gene nobody cites
574. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Europe PMC supplementaryFiles endpoint for PMC6880934, and three alternative routes to NIHMS1542708-supplement-5.xlsx

     · EMBL-EBI / NCBI PMC 2026
     used by The PPTC childhood PDX panel — 90 leukaemia models, an MLL-ALL group, and no published partner or age breakdown
575. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI esearch db=gds field query for 37679565[PMID], run against the Entrez Programming Utilities

     · NCBI Entrez Programming Utilities 2026
     used by The prescribed GEO lookup returns zero for every paper this project has ever resolved, including the one it was written from
576. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI elink dbfrom=pubmed&db=gds for PMID 37679565 and elink to sra, biosample and bioproject across the worklist

     · NCBI Entrez Programming Utilities 2026
     used by The prescribed GEO lookup returns zero for every paper this project has ever resolved, including the one it was written from
577. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     analyses/interactome-bait-versus-prey/results.json (PSICQUIC, generated 2026-09-21) compared against analyses/q160-the-prey-only-interactome/rest-roles.json (IntAct REST, 2026-09-24)

     · this project 2026
     used by The prey-only interactome is 36.3% written, not 0% - the census, by identity, of all 212 human prey-only partners of the six core proteins
578. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB REST search, accession/organism/gene_names/protein_name/length for 245 accessions, retrieved 2026-09-24; cached in analyses/q160-the-prey-only-interactome/uniprot-cache.json

     · UniProt Consortium 2026
     used by The prey-only interactome is 36.3% written, not 0% - the census, by identity, of all 212 human prey-only partners of the six core proteins
579. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PRIDE Archive project metadata for PXD005252, PXD009839, PXD041171, PXD006201, PXD023218

     · PRIDE Archive, EMBL-EBI 2026
     used by The PTM-database route runs: K1130 and K1133 acetylation are independently re-observed in deposited mass spectrometry, K1133 in the CBP/p300 acetylome itself - and K1130 is also a ubiquitination site nobody has mentioned
580. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     phosphosite.org access attempt, 2026-09-24

     · this project 2026
     used by The PTM-database route runs: K1130 and K1133 acetylation are independently re-observed in deposited mass spectrometry, K1133 in the CBP/p300 acetylome itself - and K1130 is also a ubiquitination site nobody has mentioned
581. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Rare Pediatric Disease Designation and Priority Review Voucher Programs

     US Food and Drug Administration · FDA 2026
     used by The rare paediatric disease priority review voucher: a $150-200m bearer bond, lapsed for 14 months, now live to 2029
582. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     21 U.S.C. 360ff - Priority review to encourage treatments for rare pediatric diseases

     United States Code · Office of the Law Revision Counsel, US House of Representatives 2026
     used by The rare paediatric disease priority review voucher: a $150-200m bearer bond, lapsed for 14 months, now live to 2029
583. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.news
     JPM26: Jazz inks $200M priority review voucher sale

     Fierce Pharma · Fierce Pharma 2026
     used by The rare paediatric disease priority review voucher: a $150-200m bearer bond, lapsed for 14 months, now live to 2029
584. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paperPMID 41710508
     Priority without progress: the FDA's neglected tropical disease voucher program after 18 years

     Goh M, Outterson K, Kesselheim AS · Health Affairs Scholar 4(2):qxag024 2026
     used by The rare paediatric disease priority review voucher: a $150-200m bearer bond, lapsed for 14 months, now live to 2029
585. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     Commissioner's National Priority Voucher (CNPV) Pilot Program

     US Food and Drug Administration · FDA 2026
     used by The rare paediatric disease priority review voucher: a $150-200m bearer bond, lapsed for 14 months, now live to 2029
586. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed E-utilities esearch, counted-absence query set for XPO1 chromatin occupancy in KMT2A-rearranged leukaemia and under menin inhibition

     Oncofusion Knowledge Graph project (Q61-export-axis) · NCBI E-utilities 2026
     used by The reciprocal ChIP: is chromatin-bound XPO1 present in a KMT2A-rearranged cell, and does a menin inhibitor move it?, Three measurements the export axis needs and nobody has published: XPO1 occupancy in KMT2A-r, H3K79 under an XPO1 inhibitor, and CRM1 occupancy under a menin inhibitor
587. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     referral-denominator-audit: evidence-item level scoring of frequency claims sourced to the KMT2A recombinome series

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The referral-denominator audit: 97 frequency claims, 63 qualified, and one real defect where the ledger predicted seven
588. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
     Re-checking every patient phosphosite against canonical UniProt numbering by residue type, UniProt annotation and constant-offset search

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The residue-type numbering check passed a site that is in the wrong place
589. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     NCBI GEO GSM6430988 full record and the 60-sample series listing, retrieved 2026-09-22 (acc.cgi form=text), giving !Sample_type, !Sample_relation SRA/BioSample, !Sample_extract_protocol_ch1 and !Sample_supplementary_file_1

     · NCBI Gene Expression Omnibus 2026
     used by The resolution floor: every deposited H3K9 track in GSE210480 is 1,000 bp binned, so the 147 bp exon-11 question cannot be answered from the processed files - and the planned analysis that asked for 500 bp windows was asking for something that is not there
590. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     NCBI esummary for PMIDs 25751424, 24735925, 23870121, 22872640, 22773103, 22748127 and 18598942, fetched 2026-09-23 by Q129 to date the bait publications

     · NCBI PubMed 2026
     used by The REST-wide direction audit Q104 asked for: scaffold-as-bait rows DO exist in the 36% of the record PSICQUIC hides - KMT2A is a bait in 96 of 376 rows across 18 publications and PSIP1 in 11 of 114 across 4, but the most recent is 2015 and 2012, and none of the thirteen never-discussed partners comes from one
591. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IntAct molecular interaction database, REST findInteractions endpoint behaviour on field-scoped versus bare publication queries, measured 2026-09-23 by Q129

     · IntAct (EMBL-EBI) 2026
     used by The REST-wide direction audit Q104 asked for: scaffold-as-bait rows DO exist in the 36% of the record PSICQUIC hides - KMT2A is a bait in 96 of 376 rows across 18 publications and PSIP1 in 11 of 114 across 4, but the most recent is 2015 and 2012, and none of the thirteen never-discussed partners comes from one
592. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     IHEC Data Portal metadata API (getReleases.py and getDataHub.py, all 50 releases) and the BLUEPRINT DCC current_release file index 20160816.data.index, retrieved 2026-09-22

     · International Human Epigenome Consortium Data Portal; BLUEPRINT Data Coordination Centre, EMBL-EBI 2026
     used by The resting-state H3K9me2 depletion over KMT2A DOES replicate in a human CD34+ haematopoietic progenitor - measured in 3 of 3 progenitor ChIPs, where the locus sits in the 7th, 11th and 37th longest H3K9me2-free gap on chr11 - but it is not a progenitor property: granulocytes, K562 and 6 of 10 primary AML samples show the same gap
593. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
     cell-state-innate-immune-lamp5, Oncofusion Knowledge Graph, data/graph/nodes/cell-state/, read 2026-09-23

     · Oncofusion Knowledge Graph 2026
     used by The screen this beat was dispatched to run 'because it has never been run' was published in Haematologica in 2022, used the same instrument, and is cited in this graph already - in an entry filed under innate immunity
594. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     scripts/beat-inventory.py --keywords surface antigen immunotherapy surfaceome membrane, run 2026-09-23 over 2,565 entries

     · Oncofusion Knowledge Graph 2026
     used by The screen this beat was dispatched to run 'because it has never been run' was published in Haematologica in 2022, used the same instrument, and is cited in this graph already - in an entry filed under innate immunity
595. strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
     PubMed esearch via scripts/pubmed_safe.py, querytranslations recorded verbatim, 2026-09-23

     · NCBI PubMed 2026
     used by The screen this beat was dispatched to run 'because it has never been run' was published in Haematologica in 2022, used the same instrument, and is cited in this graph already - in an entry filed under innate immunity
596. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
     Is 'bait-supported' a quality filter? Hand-verified source labels for 56 bait-supported partners across 21 publications

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The six unwritten DOT1L partners are not six findings - they are nine spokes of ONE GFP-DOT1L pulldown, and the source paper already graded them
597. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
     UniProtKB O60477 (BRNP1_HUMAN), human BRINP1, 761 aa

     · UniProt Consortium 2026
     used by The six unwritten DOT1L partners are not six findings - they are nine spokes of ONE GFP-DOT1L pulldown, and the source paper already graded them
598. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
     interface-psip1-utp3-declined-psip1-is-prey-to-nine-of-fifty-rbp-baits, Oncofusion Knowledge Graph node file, written by Q129 2026-09-23

     · Oncofusion Knowledge Graph (internal) 2026
     used by The size correction that dissolved KMT2A's proximity promiscuity should run the OTHER WAY for PSIP1: at 530 aa, recovery by 9 of 50 BioID2 baits may be MORE than its length predicts, which would make PSIP1 rather than KMT2A the promiscuous prey in this project's core pair
599. moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
     PubMed E-utilities esearch, db=pubmed: (MLL-AF4 OR MLL-AF9 OR KMT2A-AFF1 OR MLL fusion protein) AND (chromatin fractionation OR chromatin-bound fraction OR nucleoplasmic fraction), run 2026-09-20

     Not applicable - database query · NCBI E-utilities 2026
     used by The soluble-nucleoplasm versus chromatin-bound axis - the one localisation readout that keeps a size dimension, and nobody has run it on a KMT2A fusion
600. established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.analysis
     pi-dimension-coverage: structured-field and edge audit of cell type, cell state, subcellular and subnuclear localization across gene entries

     · Oncofusion knowledge graph (internal analysis) 2026
     used by The standing priority asks for localization on every interactor; the graph has it on 14% of genes and 24% of its most central ones, and DOT1L has none at all