KMT2A / MLL1 — the gene that breaks

This is real experimental data — atoms placed by crystallography or cryo-EM, deposited by the people who did the experiment.

• KMT2A itself is under very strong human loss-of-function constraint and heterozygous loss causes a developmental syndrome, so anything that degrades wild-type KMT2A alongside the fusion carries real risk.
• No structure of the intact fusion exists, so every whole-protein figure is a schematic and structure-based design is limited to fragments.
• MLL1 maintains HSC quiescence and promotes progenitor proliferation - opposite signs in adjacent compartments. Any normal-haematopoiesis safety readout pooled across compartments can report 'no effect' by averaging an increase against a decrease, and that is the design most preclinical packages use.
• The catalytic activity everyone names the protein after is dispensable in a SET-deletion allele and required in a WDR5-interface inhibitor. A programme aimed at KMT2A catalysis is aimed at an activity whose necessity the field has not settled.

1. Is there any surface present on the fusion protein but absent from wild-type KMT2A, for example at the junction itself, that could be drugged or targeted immunologically with real selectivity?
2. Since no structure of an intact fusion protein exists, what does the full assembly actually look like on chromatin, and could predicted models of the retained half plus partner guide new interfaces?
3. Does losing the PHD3 CYP33 switch account for the non-terminating program, and can that brake be restored pharmacologically rather than blocking menin?
4. Does the fusion's retained N-terminal region fractionate differently from wild-type N320? The FYRN-FYRC clamp prediction (PMID 12482972) has never been tested on a fusion, and a five-fraction blot in a t(4;11) line would do it - the size dimension attributes the species, which imaging cannot.
5. What IS the subnuclear compartment the N320-C180 heterodimer occupies? It has been named 'a subnuclear compartment' since 2003 and never identified against any known structure.
6. Is the wild-type MLL1 allele still holding HSC quiescence in a cell that also carries a KMT2A fusion? Every quiescence measurement deletes the whole locus in a fusion-free animal, and the fusion is a dominant negative for ATR-mediated stabilisation of wild-type MLL, so the wild-type allele may already be compromised before any drug is given.
7. What fraction of H3K4me3 promoters requires MLL1 in a haematopoietic cell? The under-5% figure is from fibroblasts, and the number this project would actually use has never been measured.
8. Does a menin inhibitor move the HSC quiescence set-point? The class's documented normal-blood liability is impaired recovery under stress, and stress recovery is cell-cycle re-entry. A Ki-67/Hoechst readout on LSK cells after 14 days of drug is the direct experiment and is unreported.
9. SET deletion says the catalysis is dispensable and MM-401 says it is required. Is the difference that the inhibitor also removes WDR5-dependent complex assembly? A WIN-motif point mutant that preserves catalysis while blocking assembly, or the reverse, would separate them, and neither has been reported in a fusion line.
10. Is MLL1-dependent H3K4me3 rhythmic in a blood cell, and at HOXA? Six timepoints over 24 h in a synchronised HSPC culture answers it, and would say whether every single-timepoint occupancy measurement in this project has an uncontrolled covariate.
11. Which of the Mll-dependent HSC network genes are menin-dependent? PMID 23744037 reports the split without enumerating it, and that enumeration decides how much of normal HSC biology a menin inhibitor is expected to touch.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   The 2023 KMT2A recombinome analysed 3,401 acute leukaemia patients diagnosed between 2003 and 2022 and, including published data, identifies **107 in-frame KMT2A fusion partners**, with a further 16 out-of-frame rearrangements, 18 patients with no partner gene fused to 5'-KMT2A, two with a 5'-KMT2A deletion and one KMT2A insertion at an ETV6::RUNX1 breakpoint. Breakpoints concentrate overwhelmingly in the major cluster region BCR1, between KMT2A intron 7 and exon 13: **3,336 patients (98%)**, against 47 (1.4%) in the minor BCR2 between intron 20 and exon 24. The seven most frequent partners together with partial tandem duplications account for more than 90% of all KMT2A recombinations; 37 partners occur recurrently and 63 were seen once.

   The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023
   PMID 37019990DOI
   ▸What this evidence cannot show

   Read in full from PMC10169636 on 2026-09-19; every figure above is from the abstract and the breakpoint-distribution section, not recalled. This item previously said only 'more than 100 documented fusion partners, catalogued across thousands of patients, with breaks concentrated in a major cluster region' and was surfaced by `scripts/verify-claims.py` as one of the graph's lowest-scoring established claims — vagueness scores low against a specific abstract. A referral-based cohort from one diagnostic centre, so partner frequencies reflect which centres send samples rather than a true population distribution, and the 98% BCR1 figure is conditioned on a rearrangement having been detected by that laboratory's assays.

   ▸Evidence that argues against this (2)

   • 'More than 100 documented fusion partners' conflates recurrent drivers with single-patient observations, and the tail is longer than the head. This project's own non-elongation-partner enumeration found partners listed in the recombinome with ZERO supporting publications - SEPTIN3 at n=1 and ITSN1 at n=1 - alongside partners with dozens of papers. A partner count also grows with sequencing depth rather than with biology, so the number tracks method as much as it tracks the locus. Leukemia 2023
   • Concentration in one cluster region is not homogeneity within it, and the same paper says so: outcome has been linked to breakpoint position inside BCR1, with patients breaking in KMT2A intron 11 faring worse than those breaking upstream, the cut point falling between intron 10 and exon 11. The 63 partners seen only once also mean the '107 partners' figure is dominated by singletons whose biology is uncharacterised. Leukemia 2023
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   KMT2A comprises 38 exons; exons 11-16 encode the PHD1-PHD2-ePHD3 module, and breaks at or beyond exon 11 destroy its fold, disabling CYP33 binding and the module's dimerisation.

   The KMT2A/MLL consensus gene structure: a comprehensive update for research and diagnostic implications · Meyer C, Larghero P, Lopes BA, Marschalek R · Leukemia 2024
   PMID 38678092DOI
   ▸What this evidence cannot show

   The structural consequence of a break past exon 11 is inferred from domain boundaries and prior functional work on PHD3, not demonstrated on isogenic fusion proteins differing only in break position.

   ▸Evidence that argues against this (1)

   • Exon numbering for KMT2A has been revised more than once — a 99-bp exon 2 and an exon inside intron 21 were added to the consensus structure — so "exon 11" means different positions in different papers and any breakpoint statement tied to an exon number has to name which numbering it uses. That breaks at or beyond exon 11 destroy the PHD fold is a structural inference from domain boundaries rather than a measured loss of CYP33 binding in patient material. Blood Cells, Molecules and Diseases 2010
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review

   In infant lymphoblastic leukemia the commonest KMT2A partners are AFF1 (49%), MLLT1 (22%), MLLT3 (17%) and MLLT10 (5%); in infant myeloid leukemia MLLT10 (27%), MLLT3 (22%) and ELL (17%) dominate.

   Treatment of infant leukemias: challenge and promise · Brown P · Hematology Am Soc Hematol Educ Program 2013
   PMID 24319237DOI
   ▸What this evidence cannot show

   A 2013 educational review; percentages are of the subset of cases with an identified partner (93% of infant ALL, 66% of infant AML) and come from referral cohorts.

   ▸Evidence that argues against this (1)

   • Partner frequencies are detection-limited. Cohorts characterised by targeted RT-PCR panels built around the common partners will report those partners most, cryptic insertions can be missed entirely, and the proportions shift between eras as methods change. The ordering is robust; the percentages are not. Leukemia 2001
4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   KMT2A is not DepMap common-essential (mean Chronos about -0.21) yet has the strongest loss-of-function constraint among candidate targets examined, with LOEUF 0.070 and pLI approximately 1.0.

   DepMap CRISPR (Chronos) gene effect and gnomAD constraint, retrieved via the Open Targets Platform GraphQL API and the gnomAD GraphQL API · Broad Institute DepMap; gnomAD; Open Targets · Databases 2026
   ▸What this evidence cannot show

   Per-line values were retrieved through the Open Targets API and means computed by us; the DepMap release version is not stated by the API. Cell-line essentiality does not predict tissue tolerance, and constraint reflects selection against heterozygous carriers rather than cellular requirement.

   ▸Evidence that argues against this (1)

   • A gnomAD constraint metric measures intolerance to heterozygous loss over a lifetime from conception, in a population sample. A drug gives partial reversible occupancy in an adult for weeks, so high constraint is not a safety signal for pharmacological inhibition. There is also a tension inside this item: a gene that is not common-essential in DepMap yet carries LOEUF 0.070 is showing exactly the divergence between cellular requirement and organismal constraint that makes constraint a poor proxy for on-target toxicity. gnomAD, Broad Institute 2026
5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   The bacterially expressed MLL AT-hook domain binds cruciform DNA in a gel mobility shift assay, recognising the structure rather than the sequence of the target DNA; the same study maps a transcriptional activation domain 3' and a repression domain 5' of the MLL breakpoint region, so 11q23 translocations retain the repression domain and the AT-hook DNA-binding region while discarding the activation domain.

   11q23 translocations split the "AT-hook" cruciform DNA-binding region and the transcriptional repression domain from the activation domain of the mixed-lineage leukemia (MLL) gene · Zeleznik-Le NJ, Harden AM, Rowley JD · Proceedings of the National Academy of Sciences USA 91(22):10610-4 1994
   PMID 7938000DOI
   ▸What this evidence cannot show

   ABSTRACT ONLY for the binding experiment's details - the wording 'recognizing structure rather than sequence of the target DNA' is the authors' own, quoted from the PubMed abstract; the PMC full text was not read, so the substrate panel, competitor set and affinities behind that conclusion were not independently inspected. 1994 gel-shift biochemistry with a bacterially expressed AT-hook fragment, not the full-length protein and not in cells. This is the ONLY direct binding assay on MLL's own AT-hooks; the much larger literature describing AT-hooks as AT-rich minor-groove binders is about the motif class (Reeves & Nissen consensus peptide, HMGA proteins), not about MLL. Structure preference and AT-richness are also confounded, because AT-rich tracts are the sequences that most readily bend and extrude cruciforms.

   ▸Evidence that argues against this (1)

   • The generic AT-hook motif is defined as an A.T-rich minor-groove binder, so a class-level prior runs against the structure-recognition reading, and the 1994 result has not been independently replicated with modern methods. Reeves R et al. 1990
6. E_kmt2a_structural_coverageestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   Across every PDB entry containing KMT2A (Q03164), only 629 of 3,969 residues - 15.8% - have been solved at atomic resolution, in seven disconnected islands. The breakpoint cluster region (~residues 1100-1445) is 21% covered, and that coverage is entirely the CXXC domain (1143-1214); the residues where the gene is actually broken have never been seen. The two largest unsolved stretches are 1785-2838 and 2870-3753.

   PDBe SIFTS best_structures mapping for UniProt Q03164 · EMBL-EBI PDBe API 2026
   ▸What this evidence cannot show

   Computed by the coordinator from the live SIFTS mapping on 2026-09-19 by merging all unp_start/unp_end spans; the number will drift as new entries are deposited. Counts residues present in a deposited model, which is not the same as residues that are ordered and well-resolved - at 6.2 A the cryo-EM span in particular is backbone-level. Isoform numbering follows canonical Q03164.

   ▸Evidence that argues against this (1)

   • Low coverage is not the same as low knowledge. The unsolved regions are largely predicted to be intrinsically disordered, and AlphaFold declines to model the chain at all at this length, so there may be little fixed structure there to solve. The honest reading is that the breakpoint region has no known fold, not that someone simply has not got round to it. EMBL-EBI AlphaFold DB 2026
7. E_q72_kmt2a_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   UNIPROT MODELS THE TWO-FRAGMENT REALITY AND ATTRIBUTES ONLY ONE THIRD OF IT. Fetched 2026-09-21, Q03164 (KMT2A_HUMAN, primary gene KMT2A, 3,969 aa) carries THREE separate SUBCELLULAR LOCATION comments, not one: the full-length protein, 'Nucleus', with three ECO:0000269 attributions (PMID 12482972, 25593309, 34850113); a comment scoped to the molecule 'MLL cleavage product N320', 'Nucleus', WITH NO EVIDENCE CODE AT ALL; and a comment scoped to 'MLL cleavage product C180', 'Nucleus', ALSO WITH NO EVIDENCE CODE, carrying the uncoded note 'Localizes to a diffuse nuclear pattern when not associated with MLL cleavage product N320'. So any entry that reads the first comment and stops has silently dropped the two fragments the protein actually exists as.

   UniProtKB entry Q03164 (KMT2A_HUMAN), three molecule-scoped SUBCELLULAR LOCATION comments and feature table · UniProt Consortium · UniProtKB 2026
   ▸What this evidence cannot show

   A database state on a date, not a biological finding. The molecule-scoped comments are the ones that matter for this project and they are exactly the ones with no evidence code - they restate PMID 12482972's result without citing it, so they are curator prose rather than an attributed assertion, and a downstream tool that filters on ECO:0000269 will keep the least specific line and discard the two useful ones. That is the same failure shape as brief rule 69b: the field you read decides the finding.

   ▸Evidence that argues against this (1)

   • The uncoded C180 note is not merely unattributed, it is CONTESTED. A 2019 study blocking taspase1 cleavage reported MLL1 left uncleaved, more degradation-resistant, with markedly increased genome-wide chromatin occupancy and SUBCELLULAR LOCALISATION UNCHANGED. Those are different perturbations - a clamp-disrupted fragment pair versus prevention of cleavage - but a reader taking UniProt's uncoded note as settled gets no warning that a peer-reviewed paper reports no localisation change in the same pathway. Zhao Z et al. 2019
8. E_q72_kmt2a_loc2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   TWO OF THE THREE EXPERIMENTAL PAPERS UNIPROT CITES FOR 'KMT2A IS NUCLEAR' ARE NOT KMT2A LOCALISATION STUDIES. PMID 25593309 is a screen of human bromodomain proteins for DNA-damage-induced RELOCALISATION, in which more than a third of BRD proteins change localisation after damage and whose subject is ZMYND8 and the NuRD complex; PMID 34850113 is a KSHV latency paper about LANA recruiting MLL1 to viral terminal repeats. Both are legitimate observations of a nuclear MLL1 and neither was designed to establish where KMT2A is. The one cited paper that is a KMT2A localisation study is PMID 12482972 [E_q72_kmt2a_loc3].

   Screen identifies bromodomain protein ZMYND8 in chromatin recognition of transcription-associated DNA damage that promotes homologous recombination. · Gong F, Chiu LY, Cox B, Aymard F, Clouaire T, Leung JW, Cammarata M, Perez M, Agarwal P, Brodbelt JS, Legube G, Miller KM · Genes Dev 2015
   PMID 25593309DOI
   ▸What this evidence cannot show

   Abstracts read in full; full texts not read, so a localisation panel inside either paper cannot be excluded (brief rule 80). This is a comment on PROVENANCE, not a challenge to the conclusion - KMT2A is nuclear and nobody disputes it. It is recorded because this graph repeatedly reads UniProt attributions as if the cited PMID were the founding experiment, and rule 82's lesson applies in reverse here: a high-profile PMID attached to a claim is not evidence that anyone reasoned about that claim in that paper. NOTE ALSO that 25593309 is a HeLa/U2OS DNA-damage system and 34850113 a herpesvirus latency system - neither is haematopoietic (rule 73).

   ▸Evidence that argues against this (1)

   • A KSHV latency paper is a stranger source for a localisation code than it looks, but it is not an error: it shows MLL1 recruited to a specific chromatinised DNA template, which is a stronger nuclear demonstration than bulk staining. The criticism to keep is narrow - UniProt's citation set for this field tells you which papers a curator happened to process, not which experiment established the fact. Tan M et al. 2021
9. E_q72_kmt2a_loc3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   KMT2A IS NOT ONE PROTEIN IN THE NUCLEUS, IT IS TWO THAT CLIP TOGETHER. MLL is normally cleaved at two conserved sites (D/GADD and D/GVDD) into N-terminal p320 (N320) and C-terminal p180 (C180), which form a stable complex that LOCALISES TO A SUBNUCLEAR COMPARTMENT; the FYRN domain of N320 binds the FYRC and SET domains of C180; and DISRUPTING THAT INTERACTION markedly decreases N320 levels and REDISTRIBUTES C180 TO A DIFFUSE NUCLEAR PATTERN. The same paper predicts that MLL fusion proteins of leukaemia, which cannot complex with C180, have their stability conferred by the fusion partner instead.

   Proteolytic cleavage of MLL generates a complex of N- and C-terminal fragments that confers protein stability and subnuclear localization. · Hsieh JJ, Ernst P, Erdjument-Bromage H, Tempst P, Korsmeyer SJ · Mol Cell Biol 2003
   PMID 12482972DOI
   ▸What this evidence cannot show

   Abstract read in full; full text not read. ALREADY LOAD-BEARING IN THIS GRAPH - cited in 5 entries, and localization-fyrn-fyrc-complex-the-only-module-change-with-a-measured-subnuclear-consequence already reasons about it in detail (brief rule 82: read that entry, do not restate it). The subnuclear compartment is NOT NAMED in the paper and has never been identified, so 'a subnuclear compartment' is the whole of what is known. The experiment is on transfected constructs, not endogenous protein, and not in a leukaemic cell. Crucially for this project: EVERY KMT2A FUSION LACKS C180, so the fusion is permanently in the state this experiment created artificially - which is a prediction, not a measurement, because nobody has fractionated a fusion against wild-type N320.

   ▸Evidence that argues against this (1)

   • CONTESTED, and the graph must not repeat either side as settled. Preventing taspase1 cleavage altogether left MLL1 uncleaved, more resistant to cycloheximide-chase degradation and markedly INCREASED in genome-wide chromatin occupancy, with SUBCELLULAR LOCALISATION UNCHANGED and bulk H3K4 methylation unchanged. The perturbations differ - clamp disruption versus no cleavage - but the 2019 result also inverts the stability logic: cleavage destabilises MLL1 rather than maturing it. Zhao Z et al. 2019
10. E_q72_kmt2a_loc4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    IN U937 AND HeLa UNDER CONDITIONAL OVEREXPRESSION, the MLL AMINO TERMINUS targets the protein to the NUCLEOLUS, the interphase NUCLEAR MATRIX and MITOTIC CHROMOSOMAL SCAFFOLDS where it colocalises with topoisomerase II, and conditionally expressed MLL-AF9 and MLL-AF4 give the same punctate nucleoplasmic-plus-nucleolar pattern as a 50 kDa amino-terminal MLL fragment and as endogenous MLL. THAT IS THE SUBNUCLEAR RECORD FOR KMT2A: a compartment set attributed to the half of the protein every fusion keeps.

    The amino terminus targets the mixed lineage leukemia (MLL) protein to the nucleolus, nuclear matrix and mitotic chromosomal scaffolds. · Caslini C, Alarcòn AS, Hess JL, Tanaka R, Murti KG, Biondi A · Leukemia 2000
    PMID 11069025DOI
    ▸What this evidence cannot show

    Abstract read in full; full text not read. CITED 11 TIMES IN THIS GRAPH already - it is one of the corpus's load-bearing localisation papers (rule 82), and localization-nucleolus-kmt2a-fusion, localization-nuclear-matrix-mitotic-scaffold-kmt2a and localization-kmt2a-mllt3-the-address-is-a-kmt2a-property all reason about it. NOT A LEUKAEMIC B-LINEAGE SYSTEM: U937 is monocytic and HeLa is cervical carcinoma, and the fusions are CONDITIONALLY EXPRESSED constructs, not endogenous t(4;11) protein (rule 73). 'Nuclear matrix' is an operational extraction fraction. Because the pattern is indistinguishable between endogenous MLL, the N-terminal fragment and two different fusions, the experiment cannot attribute any compartment to the fusion as opposed to the KMT2A moiety - which is the paper's own point and is the reason this graph should spend no confidence on 'the fusion is nucleolar'.

    ▸Evidence that argues against this (1)

    • The endogenous experiment says the picture cannot tell the species apart. Butler and colleagues resolved chimeric HRX proteins at the predicted fusion sizes on a blot from t(4;11) and t(11;19) cells - the size dimension doing the attribution - then imaged the SAME reagents and found punctate nuclear staining of wild-type and chimeric HRX alike, 'in cells with and without 11q23 translocations', concluding in their own words that this 'precludes an immunocytochemical approach for diagnosis'. So the punctate-nuclear description of KMT2A and of its fusions rests on images that their own authors said are unattributable. Butler LH et al. 1997
11. E_q72_kmt2a_loc5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    THE NUCLEAR ADDRESS IS CARRIED BY THE PART EVERY FUSION KEEPS, WHICH IS WHY 'THE FUSION IS NUCLEAR' IS NOT A FUSION PHENOTYPE. Asked in the hardest case - MLL-AF6, whose partner AFDN/AF6 is a cytoplasmic Ras-binding protein - immunofluorescence AND cell fractionation showed the chimera is nuclear, and the region of the MLL N-terminus responsible was mapped to A REGION CONTAINING THE AT-HOOK MOTIFS. UniProt Q03164 annotates NO nuclear-localization-signal motif on KMT2A at all (fetched 2026-09-21); the five motifs it does annotate are MBM, IBM1, IBM2, 9aaTAD and WIN.

    Chimeric MLL products with a Ras binding cytoplasmic protein AF6 involved in t(6;11) (q27;q23) leukemia localize in the nucleus. · Joh T, Yamamoto K, Kagami Y, Kakuda H, Sato T, Yamamoto T, Takahashi T, Ueda R, Kaibuchi K, Seto M · Oncogene 1997
    PMID 9349501DOI
    ▸What this evidence cannot show

    Abstract read in full; full text not read. 1997 deletion-construct mapping in transfected cells, not endogenous protein and not a leukaemic B-lineage system (rule 73). The mapping localises a SUFFICIENT region and does not distinguish active nuclear IMPORT from nuclear RETENTION by chromatin binding - localization-nuclear-import-determinant-maps-to-the-at-hook-region already records that ambiguity and is the entry to read. The absent NLS feature in UniProt is consistent with this: the determinant is a region, not a recognisable signal sequence.

    ▸Evidence that argues against this (1)

    • One partner does move something. Expression of MLL-AF4 can ALTER THE SUBNUCLEAR LOCALISATION OF AF9 (PMID 14603337), and MLL-AF6-rearranged cells relocate endogenous AF6 from cytoplasm to nucleus (PMID 24695851, cited elsewhere in this graph). So the KMT2A half dominating the fusion's own address does not mean the fusion has no localisation phenotype - it means the phenotype to look for is in TRANS, on partner proteins, and that is where the only positive results are. Erfurth F et al. 2004
12. Q127_celltype1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Committed lymphoid and myeloid cells no longer require Mll, defining the early multipotent stages of haematopoiesis as Mll-dependent; within those stages Mll maintains quiescence in HSCs, whose loss causes ectopic cell-cycle entry and depletion of the quiescent pool, and promotes proliferation in myelo-erythroid progenitors, whose loss reduces proliferation and blunts the response to cytokine-induced cell-cycle entry.

    Unique and independent roles for MLL in adult hematopoietic stem cells and progenitors. · Jude CD, Climer L, Xu D, Artinger E, Fisher JK, Ernst P · Cell Stem Cell 2007
    PMID 18371366DOI
    ▸What this evidence cannot show

    Mouse, inducible conditional knockout of the whole Mll locus - a statement about wild-type MLL1, not about a fusion allele. ABSTRACT READ ONLY (PMC2234224 returned an empty full-text body on 2026-09-23), so no percentage, assay or n is quoted. The PubMed MCP abstract for this record silently deletes every italicised 'Mll'; the symbols here were restored by hand from the title and the PubMed web record. This paper had ZERO mentions anywhere in this corpus before cycle 89.

    ▸Evidence that argues against this (1)

    • A second conditional Mll allele published in the same issue reports GROSSLY NORMAL steady-state haematopoiesis in bone marrow, thymus and spleen, against this paper's fatal bone marrow failure within 3 weeks. The quiescence direction agrees across the two; the severity does not. See concept-two-founding-mll-conditional-knockouts-disagree-about-steady-state-blood. McMahon KA et al. 2007
13. Q127_cellstate1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Mll-null embryos survive to E16.5 with reduced HSC numbers and a greatly reduced quiescent fraction, and those HSCs cannot compete in transplantation; adult conditionally Mll-deleted bone marrow is highly compromised in competitive reconstitution despite grossly normal steady-state haematopoiesis.

    Mll has a critical role in fetal and adult hematopoietic stem cell self-renewal. · McMahon KA, Hiew SY, Hadjur S, Veiga-Fernandes H, Menzel U, Price AJ, Kioussis D, Williams O, Brady HJM · Cell Stem Cell 2007
    PMID 18371367DOI
    ▸What this evidence cannot show

    Mouse; abstract read, full text not opened. 'Greatly reduced' is the authors' wording and no percentage appears in the abstract. The fetal result is a constitutive null and the adult result a conditional deletion, so the two halves are not the same experiment. ZERO mentions anywhere in this corpus before cycle 89.

    ▸Evidence that argues against this (1)

    • The paper published immediately before this one in the same issue, from a different conditional Mll allele, reports FATAL BONE MARROW FAILURE within three weeks of deletion rather than grossly normal steady-state haematopoiesis. The two agree on loss of quiescence and on competitive failure and disagree on whole-animal severity. Jude CD et al. 2007
14. Q127_cellstate2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Wild-type MLL protein is degraded biphasically through the cell cycle, by SCF(Skp2) at S phase and APC(Cdc20) at late M, giving two expression peaks; abolishing the peaks produces G1/S and M-phase defects and overexpression blocks S-phase progression; degradation initiates within the N-terminal approximately 1,400 amino acids, and the tested prevalent MLL fusions are RESISTANT to that degradation.

    Bimodal degradation of MLL by SCFSkp2 and APCCdc20 assures cell cycle execution: a critical regulatory circuit lost in leukemogenic MLL fusions. · Liu H, Cheng EH-Y, Hsieh JJ-D · Genes Dev 2007
    PMID 17908926DOI
    ▸What this evidence cannot show

    HeLa and leukaemia lines; abstract read, full text not opened. Cited in 24 entries of this corpus already, and in none of them was this entry - the conclusion is not new to the graph, only to gene-kmt2a. The fusion-resistance claim is about 'tested prevalent MLL fusions' and the abstract does not enumerate them; cell-state-unsynchronisable-cell-cycle holds the detail, including that MLL-AF4 protein does not fluctuate through the cycle in SEM.

    ▸Evidence that argues against this (1)

    • Cell-cycle control of MLL abundance is not the only cell-cycle circuit this protein sits in, and the two are sometimes conflated: taspase1 cleavage of MLL and MLL2 is separately required for cell-cycle progression, with MLL targeting E2Fs for cyclin expression and H3K4 trimethylation reduced at Cyclin E1 and E2 in Taspase1-null cells. An entry that records only the degradation circuit understates how much of the fusion's cell-cycle abnormality is attributable to loss of processing rather than loss of degradation. Takeda S et al. 2006
15. Q127_epi1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In Mll1-null versus wild-type mouse embryonic fibroblasts, Mll1 is required for H3K4 trimethylation of LESS THAN 5% of the promoters carrying that modification; the affected genes are enriched for developmental regulators including Hox genes and show matched decreases in RNA polymerase II recruitment and expression, while loss of MLL3/MLL4 and/or the Set1 complexes has little to no effect on H3K4 methylation of Hox loci.

    Global analysis of H3K4 methylation defines MLL family member targets and points to a role for MLL1-mediated H3K4 methylation in the regulation of transcriptional initiation by RNA polymerase II. · Wang P, Lin C, Smith ER, Guo H, Sanderson BW, Wu M, Gogol M, Alexander T, Seidel C, Wiedemann LM, Ge K, Krumlauf R, Shilatifard A · Mol Cell Biol 2009
    PMID 19703992DOI
    ▸What this evidence cannot show

    Mouse embryonic fibroblasts - not haematopoietic, no fusion. The <5% figure is a promoter-level call on 2009-era platforms and its denominator is not given in the abstract; abstract read, full text not opened. This bounds what 'KMT2A is an H3K4 methyltransferase' means in a cell; it does not say the enzyme is unimportant. ZERO mentions anywhere in this corpus before cycle 89.

    ▸Evidence that argues against this (1)

    • A pharmacological experiment points the other way in the cells this project cares about: MM-401, which blocks MLL1's H3K4 methyltransferase activity via the WDR5 interface, blocks MLL-cell proliferation with arrest, apoptosis and myeloid differentiation and no general marrow toxicity. An activity required at under 5% of promoters in fibroblasts is nonetheless a viable target in MLL-rearranged cells, so the bound must not be read as 'the enzyme does not matter'. Cao F et al. 2014
16. Q127_epi2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The SET domain, and thus the histone methyltransferase activity of MLL1, is dispensable for maintaining haematopoietic stem cells and for supporting MLL-AF9-driven leukaemogenesis; on Mll1 deletion it is H4K16 ACETYLATION that is selectively depleted at MLL1 target genes alongside reduced transcription, recruited MOF activity rather than intrinsic MLL1 methyltransferase activity maintains those genes, and SIRT1 inhibition is sufficient to prevent the loss.

    The histone methyltransferase activity of MLL1 is dispensable for hematopoiesis and leukemogenesis. · Mishra BP, Zaffuto KM, Artinger EL, Org T, Mikkola HKA, Cheng C, Djabali M, Ernst P · Cell Reports 2014
    PMID 24813891DOI
    ▸What this evidence cannot show

    Mouse; abstract read, full text not opened. 'Dispensable' means a SET-deletion allele was not limiting in those two assays; a SET deletion can perturb folding as well as catalysis. This PMID is cited in 32 entries of this corpus - including subtraction-kmt2a-set-domain, mark-h4k16ac and mechanism-set-loss-neomorph-not-haploinsufficiency - and was in none of gene-kmt2a. It is the single most-cited source this entry was missing.

    ▸Evidence that argues against this (1)

    • CONTESTED BY A PHARMACOLOGICAL EXPERIMENT. MM-401, which blocks the MLL1-WDR5 interaction and so MLL1 H3K4 methyltransferase activity specifically, blocks proliferation of MLL cells with cell-cycle arrest, apoptosis and myeloid differentiation and no general toxicity to normal bone marrow or non-MLL cells, with a transcriptome change resembling MLL1 deletion. A SET deletion and an assembly blockade are different perturbations, so the two results may both stand, but this graph must not state 'the activity is dispensable' unqualified. Cao F et al. 2014
17. Q127_epi3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    H3K4 trimethylation at circadian promoters is rhythmic and follows the profile of circadian H3 acetylation; MLL1 is essential for circadian transcription and for cyclic H3K4 trimethylation, is in a complex with CLOCK-BMAL1, contributes to its rhythmic recruitment, and fails to interact with CLOCK-delta19.

    The histone methyltransferase MLL1 permits the oscillation of circadian gene expression. · Katada S, Sassone-Corsi P · Nat Struct Mol Biol 2010
    PMID 21113167DOI
    ▸What this evidence cannot show

    Mouse and cultured cells, wild-type MLL1, no leukaemia and no fusion; abstract read, full text not opened; single paper with no replication identified in this beat, hence moderate. Recorded because UniProt and HPA both list 'Biological rhythms' among KMT2A's biological processes and this corpus of 2,383 entries had ZERO mentions of this paper and no circadian content about KMT2A at all.

    ▸Evidence that argues against this (1)

    • Bounded by the same study that limits MLL1's H3K4me3 requirement to under 5% of promoters: a rhythm requiring MLL1 is a rhythm at a small promoter set, and nothing shows that any leukaemia-relevant locus, HOXA included, is in it. The relevance of the circadian result to this project is therefore a hypothesis, not a finding. Wang P et al. 2009
18. Q127_celltype2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    For KMT2A (ENSG00000118058) the Human Protein Atlas records RNA tissue specificity 'Low tissue specificity' with distribution 'Detected in all', blood cell specificity 'Low immune cell specificity', blood lineage specificity 'Low lineage specificity' and single-cell 'Low cell type specificity'; the antibody-based data call it Tissue enriched with bone marrow the top tissue and Cell type enriched with hepatocytes the top cell type, on antibodies CAB017794, CAB024270 and HPA044910 with IH reliability 'Enhanced' and IF reliability 'Supported', subcellular main location Nucleoplasm with additional Cytosol.

    Human Protein Atlas - KMT2A (ENSG00000118058), machine-readable gene record · Human Protein Atlas consortium · proteinatlas.org 2026
    ▸What this evidence cannot show

    Fetched 2026-09-23 from the per-gene JSON endpoint, which does NOT return the HPA release version, so this row can drift silently and must be refetched before any value is requoted. RNA calls are consensus transcriptomics and protein calls are antibody staining scored by eye - different instruments. The hepatocyte call has no support in the haematopoietic literature and points away from the bone-marrow call; two discordant calls from one instrument are an antibody signature, not a biological one. Staining intensities are arbitrary units, not concentrations.

    ▸Evidence that argues against this (1)

    • HPA's own two instruments disagree for this gene. The RNA data say Low tissue specificity and Detected in all; the antibody data say Tissue enriched in bone marrow and Cell type enriched in hepatocytes. The hepatocyte call has no support in the literature this project reads, and the safest reading is that the antibody-based tissue calls should not be used as evidence of abundance at all - including the bone-marrow one that happens to agree with the genetics. Human Protein Atlas consortium 2026
19. Q127_cellstate3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The Mll-dependent transcriptional network in haematopoietic stem cells extends well beyond clustered Hox genes, contains many characterised regulators of self-renewal, and comprises target genes that are both dependent on and independent of the MLL cofactor menin; PRDM16 was uniquely effective at partially rescuing Mll-deficient haematopoietic stem and progenitor cells.

    An MLL-dependent network sustains hematopoiesis. · Artinger EL, Mishra BP, Zaffuto KM, Li BE, Chung EKY, Moore AW, Chen Y, Cheng C, Ernst P · Proc Natl Acad Sci U S A 2013
    PMID 23744037DOI
    ▸What this evidence cannot show

    Mouse; abstract read, full text not opened. The menin-dependent versus menin-independent partition is the load-bearing part for this project and the abstract does not name which genes fall on which side, so no gene-level assignment may be taken from here. Cited once in this corpus before cycle 89 (collab-ernst-colorado) and never in gene-kmt2a.

    ▸Evidence that argues against this (1)

    • The menin-independent half of this network is the part that matters for the drug and this study does not enumerate it. The menin knockout itself reports that HoxA9 expression falls only MODESTLY in menin-deficient HSCs, which suggests the menin-dependent portion of the normal HSC programme is smaller than the fusion-centric account assumes. Maillard I et al. 2009

• ← related toAbsence claims: 93% are unbounded in prose, but the evidence claims are 71% bounded - and the query that tested them was itself broken

  UNBOUNDED AT ESTABLISHED: the claim that the residues where the gene breaks have never been resolved is a PDB-coverage statement with no recorded date. Cheap to bound, and load-bearing for the breakpoint beat.

• ← regulatesHDAC3 - the only HDAC in this disease whose requirement was established genetically rather than by a drug, and the same experiment shows it is not fusion-specific

  class I HDAC inhibition relieves an inhibitory function of wild-type MLL, overriding MLL-AF4 at ALOX5

• ← related toAnalysis (runnable): does product viability leave a signature in the KMT2A breakpoint map once DNA fragility is controlled for?

  maps breakpoints onto UniProt Q03164 domain architecture

• ← fuses withAFF1 / AF4 — domain architecture and residue coordinates

  AFF1 is the commonest 3' partner of KMT2A in infant B-ALL

• ← fuses withAFDN / AF6 — domain architecture and residue coordinates

  AFDN is the commonest cytoplasmic KMT2A fusion partner, t(6;11)(q27;q23)

• ← fuses withEPS15 / AF1p — domain architecture, and the one coordinate nobody has published

  EPS15/AF1p is the 3' partner in KMT2A::EPS15, t(1;11)(p32;q23)

• ← part ofSpliced KMT2A fusions join to partner exon 2 — with the per-partner counts, and a correction

  partner-side junction rule for the KMT2A recombinome: the most 3' KMT2A exon splices to partner exon 2

• ← related toPBX1 — domain architecture and residue coordinates

  both act through HOX-cofactor transcriptional programmes in B-lineage leukaemia

• ← related toBCR — domain architecture and residue coordinates

  comparator: unlike KMT2A, BCR contributes a compact crystallised oligomerisation domain with exact coordinates

• ← related toAnalysis (runnable): is anything actually enriched at KMT2A breakpoints? The test has never been run against a background that could say no

  GRCh38 span chr11:118436456-118526832, 90377 bp, used as the permutation domain

• ← causesTOP2B - the isoform most etoposide-induced KMT2A breaks depend on, resolved genetically and not resolvable pharmacologically

  TOP2B-dependent etoposide-induced double-strand breakage at the KMT2A locus

• ← related toTOP2A - the proliferation isoform, the one the drugs are actually measured against, and the only one with a germline association in this disease

  germline tag-SNP study found MLL and EP300 variants associated with childhood ALL and no TOP2A variant

• ← causesEtoposide (VP-16) - the drug that makes this disease and cures several others, and the only entry in this graph whose main finding is that you must not stop using it

  site-specific cleavage in the KMT2A breakpoint cluster region in primary newborn and adult progenitors

• ← regulatesTDP2 - the protein-adduct remover named in this graph's own mechanism entry and until now having no entry of its own

  protects the KMT2A locus from TOP2-induced breakage; loss increases translocation

• ← bindsRAD51 - the one repair protein measured AT the KMT2A translocation hotspot, and ATM controls how much of it loads there

  RAD51 occupancy measured AT the 11q23 translocation breakpoint hotspot after etoposide; excessive under ATM deficiency

• ← related toCTCF - named in the id of an existing entry in this graph and having no entry of its own, and the protein whose binding sites break preferentially in KMT2A-rearranged patients

  404 CTCF sites preferentially break-prone in KMT2A-rearranged patients; KMT2A region 3.2-fold above non-fusion

• regulates →The non-synchronisable state: five KMT2A-rearranged lines refuse mimosine, double thymidine and nocodazole, and the fusion protein does not fluctuate through the cycle

  wild-type KMT2A is an ATR-S516 S-phase checkpoint effector degraded biphasically by SCF-Skp2 and APC-Cdc20; fusions act as dominant negatives on that step

• ← part ofWhere KMT2A and its fusions sit inside the nucleus — SNL1/SNL2, nuclear speckles and the condensate question

  SNL-1 (aa 388-432) and SNL-2 (aa 1051-1089) lie within the N-terminal region retained by every KMT2A fusion

• ← related toCell type or the fusion? MV4-11 LOOSENS the confound the graph's own lineage split was built on - but its founding paper calls it biphenotypic

  the index fusion's 5' gene; KMT2A-r arm mean gene effect -0.811 against -0.206 across all 1,258 lines

• ← related toKMT2B has no haematopoietic cell-state result, and two of the seven records that look like one are about a different gene in muscle

  the rule-66 pair that makes every MLL2/MLL4 citation ambiguous

• ← related toA 2026 preprint puts XPO1 upstream of the menin interface itself: XPO1 reported to stabilise menin on chromatin and menin's binding to KMT2A and KMT2A::MLLT3

  menin-KMT2A interaction reported destabilised by XPO1 inhibition

• ← related toFANCF — FA core complex; rank 11 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toFANCI — ID2 module; rank 15 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toFANCA — FA core complex; rank 20 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toBRIP1 — downstream effector; rank 22 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toFANCL — FA core complex; rank 26 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toFANCG — FA core complex; rank 27 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toUBE2T — E2; rank 33 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toFAAP100 — FA core complex; rank 36 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  no published record links this gene to KMT2A or MLL: a counted, alias-searched, title-checked absence

• ← related toTranscription–replication conflict as the candidate reason KMT2A-rearranged cells need the Fanconi pathway — a chain with two measured links and three inferred ones

  R-loops are mapped by DRIP-seq in MLL-rearranged AML, but as productive architecture rather than damage

• ← related toMED23 confirmed, and the zero-record claim corrected: it is not literature-free in leukaemia, an shRNA screen already found it

  zero PubMed records for MED23 plus CRSP3, KIAA1216, DRIP130, SUR2 and ARC130 against KMT2A or MLL — the prior zero-record claim survives the alias expansion

• ← related toIn-silico digest of 81 modelled KMT2A fusion junctions: the peptide that would prove the fusion protein exists

  five of eleven candidate 5' donor exon boundaries end in lysine, so trypsin cuts at the junction

• ← related toThe protein nobody can attribute by antibody is the one mass spectrometry finds most easily

  the reciprocal junction retains KMT2A 1407-3969 including both taspase1 sites

• ← related toAn anti-MLL-N antibody sees the wild-type allele and the direct fusion, and nothing else - and cannot rank them

  the epitope both the wild-type protein and the direct fusion retain

• ← related toCorrecting the fusion with genome editing: verdict NOT FEASIBLE for infant leukaemia, though intron-targeted disruption is a real published strategy

  Counter-evidence (Leibowitz 2021, PMID 33846636): Cas9 double-strand breaks initiate chromothripsis via micronuclei and chromosome bridges, shown at a clinically relevant locus in clinically relevant cells -- a two-cut, two-chromosome fusion-disruption design is the configuration most likely to trigger it, in the exact cell compartment the disease arises in

• ← related toHypothesis: KMT2A::AFF1 is continuously required, not a hit-and-run initiator whose job is done once the epigenome is set

  Supporting rebuttal recorded as counter-evidence (Horton 2005, PMID 16230385): Tet-off MLL-ENL immortalised myeloid progenitors terminally differentiate or die on fusion withdrawal with specific loss of Hoxa4-Hoxa11, so an MLL fusion is continuously load-bearing once the state exists -- murine, retroviral, MLL-ENL, which is the gap the human t(4;11) degron experiment closes

• ← bindsSET / 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

  TAF-I beta acidic region binds the MLL N-terminal region retained in every fusion

• ← regulatesThe other arm of the crosstalk: H2Bub licenses H3K4 methylation through ASH2L and relieves a built-in COMPASS autoinhibition — and MLL3 does not respond at all

  trans-regulation by ubiquitin promotes MLL1 activity for all three H3K4 methylation states; MLL3 does not respond

• depends on →MHC class I is a bivalent gene set, and a menin inhibitor derepresses it - shown in K-562, SCLC and neuroblastoma, and measured on HLA-A in primary AML, but never once in a KMT2A-rearranged cell

  combined KMT2A/KMT2B knockout abolishes the menin-inhibitor-induced MHC-I increase

• ← related toHLA-B - the class I heavy chain the bivalency screen was actually read out on, and the locus that presents the one KMT2A epitope this graph has a structure for

  presents KMT2A 747-755 phosphopeptide from the retained portion every fusion keeps

• ← depends onCD274 (PD-L1) - named in zero entries of this graph until now, a direct MLL1-H3K4me3 target gene in another tissue, and a cell-intrinsic dependency of MLL-AF9 leukaemia-initiating cells

  MLL1 binds the cd274 promoter and catalyses H3K4me3 to activate PD-L1 - in pancreatic cancer, by SET-domain activity the fusion has lost

• ← related toSETD1A — a leukemia dependency whose enzyme activity turns out to be dispensable

  Counter-evidence: MLL1 SET-domain catalysis is dispensable for MLL-AF9 leukemogenesis, which undercuts the Win-motif/WDR5 rationale shared by SETD1A and MLL1 complexes

• ← related toRBBP5 — a WRAD subunit that binds the one part of KMT2A every fusion throws away

  Counter-evidence: the MLL1 SET domain that RBBP5/WRAD stimulates is dispensable for MLL-AF9 leukemogenesis, and KMT2A fusions do not retain it

• ← related toASH2L — the WRAD subunit that has DPY30's essentiality profile and none of its co-dependency signal

  Counter-evidence: same objection as RBBP5 — MLL1 methyltransferase activity is not required for leukemogenesis, so the ASH2L catalytic role is not a target rationale here

• ← part ofDPY30 — the smallest COMPASS subunit, and the only one that behaves like a selective dependency

  WRAD core subunit of the MLL1/KMT2A methyltransferase complex

• ← bindsPAF1 — pan-essential subunit of a complex the fusion needs, where the window is in the interface and not in the protein

  PAF1C binds MLL via CxxC-flanking sequences retained in all fusions

• ← bindsRBBP5 — a WRAD subunit that binds the one part of KMT2A every fusion throws away

  binds the MLL1 SET domain - which every KMT2A fusion loses, so this is a wild-type contact only

• ← bindsPRMT1 — an MLL-fusion cofactor whose first-in-human inhibitor was terminated for lack of efficacy and toxicity

  essential component of the MLL-EEN complex via Sam68; not demonstrated for KMT2A::AFF1

• ← depends onKMT2A::MLLT3 (MLL-AF9), t(9;11)(p21.3;q23.3)

  the retained wild-type KMT2A allele is required for MLL-AF9 leukemogenesis and maintenance, and the fusion needs a prebound wild-type MLL1 complex for stable HOXA9 association

• supports →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)

  the intact allele is not a bystander: MLL1 fusion proteins lack the PHD fingers and require a prebound wild-type MLL1 complex for stable HOXA9 association

• ← related toPPIE (CyP33) — the off-switch the fusion throws away, and the reason PHD3 reinsertion kills transformation

  corrective: H3K4me3/2 and the CyP33 RRM bind different PHD3 surfaces and can occupy them simultaneously in a ternary complex (Wang 2010), so the switch is not simply mutually exclusive

• ← depends onProximity-labelling map of KMT2A::AFF1 — the experiment nobody has run in twenty years of interactome work

  the wild-type-allele control is confounded: the fusion requires a prebound wild-type MLL1 complex at the same loci, so both tagged lines label one neighbourhood

• ← fuses withCREBBP (CBP) — binds the KMT2A transactivation domain, which no fusion keeps

  t(11;16)(q23;p13.3) KMT2A::CREBBP - the CBP bromodomain and HAT domain fused to the MLL N-terminus are both required for transformation

• ← related toPSIP1 / LEDGF p75 — the tether with the best therapeutic index in the whole interactome

  the LEDGF IBD binds a conserved disordered IBM shared with PogZ, JPO2, IWS1 and MED1, and MLL1 IBM phosphorylation switches the affinity - a selectivity problem and a regulated, multivalent contact

• ← bindsHCFC1 (HCF-1) — a founding member of the wild-type KMT2A complex that the fusion cannot bind

  (D/E)HxY HCF-binding motif at KMT2A aa 1799-1802 — wild-type KMT2A only

• ← bindsPPIE (CyP33) — the off-switch the fusion throws away, and the reason PHD3 reinsertion kills transformation

  CyP33 RRM binds KMT2A PHD3 at ~2 uM (ITC), competitively with H3K4me3

• ← bindsKAT8 (MOF) — a genuine KMT2A partner that binds the half of KMT2A the fusion discards

  MOF zinc finger to the MLL1 C-terminal region; reciprocal co-IP, cosedimentation, cotransfection

• ← bindsCREBBP (CBP) — binds the KMT2A transactivation domain, which no fusion keeps

  CBP KIX (581-687) to the KMT2A TAD (2829-2883), five hydrophobic residues critical

• ← bindsRUNX1 — activated by KMT2A::AFF1 and then bound by the reciprocal fusion, in the one published two-oncoprotein circuit in t(4;11)

  RUNX1 protein interacts with the reciprocal AF4-MLL complex, not with the der(11) product

• ← fuses withAFF3 (LAF4) — the AFF-family scaffold of SEC-L3, and a KMT2A fusion partner in its own right

  KMT2A::AFF3, t(2;11)(q11;q23) — AFF3 is itself a recurrent 3' partner

• ← fuses withMLLT6 (AF17) — the fourth KMT2A fusion partner inside DotCom, and the only one no cell needs

  KMT2A::MLLT6, t(11;17)(q23;q21) — a fusion partner, but not in t(4;11)

• ← bindsSKIC8 (formerly WDR61) — the last PAF1C subunit, and a gene-symbol trap worth recording

  PAF1C-KMT2A contact at the retained CxxC-flanking region; not attributed to SKIC8 specifically

• ← related toMYB — rank 17 of 17,728 for KMT2A selectivity, and the only node in this beat where partial, transient suppression eradicated leukaemia without stopping normal blood production

  reverse direction: an enhancer RNA (Myrlin) from the -81 kb Myb enhancer recruits wild-type KMT2A to the Myb locus; its loss reduces CDK9 and Pol II and causes Pol II pausing in Myb exon 1

• ← bindsNPM1 (nucleophosmin) — the nucleolar shuttle whose mutant form binds chromatin with KMT2A and buys the second menin-inhibitor indication

  NPM1c binds directly to specific chromatin targets co-occupied by KMT2A (MLL1).

• ← related toKANSL1 - the NSL scaffold, and the protein through which WDR5 binding to NSL and to MLL/COMPASS is mutually exclusive

  NSL depletion lowers H3K4me2 as well as H4K16ac at a shared promoter, placing NSL upstream of MLL/SET complexes.

• ← bindsNSL complex - the MOF assembly that five NUP98 fusions bind alongside MLL1, and that competes with MLL/COMPASS for WDR5

  Five NUP98 fusions interact with both MLL1 and NSL, and MLL1 is required for NUP98-HOXA9 proliferation.

• ← bindsCXXC1 (CFP1) - a CpG-reading COMPASS subunit that binds MLL1 and recruits it to HoxA7, and whose CXXC domain cannot substitute for KMT2A's in a fusion

  CGBP/CXXC1 co-purifies with MLL1, co-localises with it, and its knockdown reduces MLL1 recruitment and H3K4me3 at HoxA7.

• regulates →METTL3 — the m6A writer with a clinical-grade inhibitor, an MLL-rearranged AML mechanism, and a B-lymphoid signal that goes the wrong way

  METTL3 is transcriptionally activated by MLL in MLL-rearranged AML

• ← related toThe 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

  KMT2A is a positive control at the 0.0th null percentile alongside MEN1

• ← bindsLEDGF IBD to KMT2A: TWO motifs, not one - IBM1 at 124-133 and IBM2 at 147-152, and the Kd this graph recorded as missing is 14.7 uM

  KMT2A IBM1 124-133 and IBM2 147-152

• ← bindsWDR5 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

  KMT2A WIN motif 3762-3767, Arg3765

• ← bindsMBM2, the second menin-binding motif of KMT2A (residues 23-40): 26-fold weaker than MBM1, never separately co-crystallised, and the only part of the menin grip a drug is not aimed at

  KMT2A residues 23-40

• ← bindsTaspase1 and its two KMT2A cleavage sites: an enzyme-substrate interface with four apo structures, one co-crystal with a small molecule, and no structure with MLL at all

  cleavage sites after D2666 and D2718

• ← bindsCXXC to unmethylated CpG DNA: nine mapped contact residues, and an affinity that spans 44 uM to 33 nM depending on the construct, the DNA and the salt

  CXXC 1147-1203 on unmethylated CpG DNA, Kd 44 uM

• ← bindsPHD3 reading H3K4me3: Kd 4.3 uM for the PHD3-bromo cassette, 30 uM for PHD3 alone, two co-crystals - and the whole module is thrown away by the breakpoint

  PHD3 1564-1628 reads H3K4me3, Kd 4.3 uM with the bromodomain

• ← bindsPHD3 and the CyP33 RRM: Kd 7-15 uM, and a 2023 ternary structure that settles the 2010 disagreement - H3K4me3 and the RRM CAN be held at once, and the real switch is RNA

  PHD3 1586-1628 binds the CyP33 RRM, Kd 14.7 uM

• ← bindsThe KMT2A transactivation domain on the CBP KIX domain: a cooperative allosteric site at 2842-2858, solved three times, and thrown away by every breakpoint

  KMT2A TAD 2842-2858 on the CBP KIX domain

• ← bindsKMT2A 747-755 on HLA-B*07:02: a phosphoneoantigen from the half of KMT2A every fusion keeps, with seven crystal structures and a cognate TCR

  KMT2A 747-755 phosphopeptide presented on HLA-B*07:02

• ← related toTwo-thirds of this graph's core-complex interaction evidence comes from experiments where our protein was not the bait, and LEDGF has one bait-supported partner out of thirty-six

  16 OF 91 PARTNERS BAIT-SUPPORTED: most of what the graph records as KMT2A's interactome is other people's by-products.

• ← related toCARM1 / PRMT4 (Q86X55) - the H3R17me2a writer, and the enzyme whose product is read by the KMT2A C-terminus that every fusion deletes

  CARM1 makes the mark that the C-TERMINAL MLL region reads - Kawabe 2012 defined that region as a reader domain for arginine-methylated proteins such as Pax7, and every KMT2A fusion deletes it

• ← related toHalf this graph's claims are unopposed; these are the ones that matter

  Hub entry whose partner-count claim conflates recurrent drivers with single-patient observations.

• ← analogous toHMGA2 (P52926) - a direct KMT2A::AFF1 target through the let-7b axis, and a recurrent fusion partner whose own fusions work by escaping a microRNA rather than by making a chimeric protein

  the cleanest mechanistic CONTRAST in the analogous-families set: KMT2A fusions matter because the chimera gains an activity, HMGA2 rearrangements because the transcript loses its 3' UTR and escapes let-7 - same cytogenetics, opposite kind of mechanism

• ← bindsWDR5 — the partner that is not a partner: it binds the half of KMT2A the fusion throws away

  binds the WIN motif, which sits in the half of KMT2A the fusion throws away

• ← regulatesZFP64 — rank 4 on selectivity, and there is not even a ChEMBL entry for it
• ← regulatesSIRT1 - 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

  opposes KMT2A function rather than partnering with it

• causes →KMT2A-rearranged acute myeloid leukaemia in children and infants

  the same gene, rearranged, drives the myeloid form as well as the lymphoblastic one

• ← part ofThe lost SET domain, and the verdict: the fusion is a gain-of-function neomorph, not a KMT2A loss-of-function

  the verdict that the fusion is a gain of function, not a loss

• proposed by →Patricia Ernst (University of Colorado) — what normal KMT2A and menin actually do, which is the therapeutic-window question

  what normal KMT2A and menin actually do, which is the safety argument

• ← supportsCo-dependency scan: what else does a menin-dependent cell need? (run, not planned)

  top correlate with MEN1 at r=0.58, the method's positive control

• ← related toOf the menin co-dependencies, which are actually safe to hit? (run)

  strongest human loss-of-function constraint in the set (LOEUF 0.07)

• ← fuses withLPP (lipoma-preferred partner) - a KMT2A fusion partner in acute monoblastic leukaemia whose day job is adhesion mechanotransduction, and which is barely expressed in leukocytes

  t(3;11)(q28;q23) in acute monoblastic leukaemia; breakpoint in intron 8 of both genes, MLL exon 8 to LPP exon 9, contributing LIM 2 (474-534) and LIM 3 (535-603). MLL exons 1-8 retain the CXXC/MT domain; the SET domain (exons 34-36) is lost.

• ← related toDepMap agrees with the mouse genetics: ARID1B is not a FUSION-SELECTIVE dependency (+0.0145, p=0.80) and reads -0.08 in both groups - but CYCLE 81 SHOWED THAT FLATNESS IS PARALOGUE BUFFERING, not dispensability. ARID1A and ARID2 ARE dependencies and are equally so in both groups - so no BAF subunit tested is fusion-selective. And MEN1 (-1.26) is a harder and more selective dependency than KMT2A itself (-0.81)

  KMT2A scores -0.8113 with a -0.5571 group difference - real and selective, but weaker than MEN1 in the same lines; note CRISPR KMT2A knockout hits the wild-type allele and the fusion's 5' portion together

• ← related toRe-measured against IntAct the interactome is 103/217 (47.5%) covered, up from 41.0%, and the read queue is now ZERO - no partner with two publications or two baits lacks an entry, so the instrument's own ranking criterion has stopped selecting anything. It was also persisting the list of partners that DO have entries while counting and discarding the 105 that do not, which is the exact defect cycle 57 fixed for the other list

  KMT2A own partner list is the worst covered bait at 49/101, across all three measurements

• ← related toDepMap co-dependency cannot rank the 97 unwritten interactome partners either, pan-cancer OR restricted to leukaemia lines: mean r 0.008 rising to 0.023 under restriction while the leukaemia-matched null rises with it, giving 1.08x the chance-expected count. Recomputing the null on the restricted lines is the only thing that stopped this being reported as a discovery, and my first pass criterion called a below-chance result a PASS

  43 of the unrankable partners are KMT2A's, including OGT at r=0.138 in leukaemia lines - inside the noise

• ← related toThe interactome tail is orderable after all, just not by evidence: 62 of the 96 unwritten partners are already DISCUSSED somewhere in this corpus and need only promoting to entries, while 34 have never been read at all. Two cycles looked for a priority order in evidence strength and found none, because the ordering that exists is over how much work is left, not over how good the evidence is

  KMT2A remains the worst-covered bait and contributes 8 of the 34 never-read partners

• ← related to43 of the 62 discussed-but-unwritten interactome partners (69%) are genuinely draftable from the committed corpus, so cycle 93's decision to work them before the 34 never-read ones has a real cost basis - but the clause that decided it was the citation clause I thought was safe, not the proxy I flagged as weak, and the 12 RNA-processing failures turn out to be a promiscuity census that should NOT be promoted partner by partner

  KMT2A is recovered by 25 of 50 proximity baits, which bounds what any of its individual proximity partners can support

• ← related toNOT YET REPLICATED (corrected cycle 97). Within the K562 map KMT2A's 25-of-50 promiscuity is explained by size (97.1st percentile falling to 90.7th among proteins its own length) and menin is under-reached at the 23.5th, surviving an abundance control. The replication attempt was UNINFORMATIVE, not negative: eleven of twelve other proximity panels do not sample the nucleus, and the one that does shows the same shape

  KMT2A is 3,969 aa and reached by 25 of 50 proximity baits, which is what its size predicts

• ← related toCORRECTED SAME CYCLE: the replication test was UNINFORMATIVE, not negative. Eleven of the twelve proximity panels do not sample the nucleus - they are centrosome, RTK, kinase, ion-channel, CFTR and mitochondrial baits - so KMT2A's 0th percentile in three of them measures panel composition, not promiscuity. In the one transcription-factor panel KMT2A is at the 71.7th and menin is absent, which is DIRECTIONALLY CONSISTENT with the K562 result at n=1

  KMT2A is not a promiscuous proximity prey in general - 0th percentile in three of five scoreable publications; the K562 map is the outlier

• ← related toSalvaged from a session-limit death: the DMSO-versus-DMSO selection control cannot be run cleanly on GSE260742, because every same-condition vehicle pair is confounded - two differ about three-fold in aggregate signal and the only admissible same-mark pair crosses processing states (filt.bam.tdf versus tdf). Its automated verdict of MANUFACTURED is NOT supported, and the control remains UNRUN rather than passed or failed

  the eviction quartiles this control tests are defined on KMT2A-fusion FLAG occupancy

• ← related toGSE299100 supplies the vehicle-versus-vehicle pair that neither GSE260742 nor GSE127507 could - two independent control hairpins, both DMSO, same date, uniformly processed to hg38_rmdup.tdf, for MEN1, MLL1 and input - in MV4-11, which Cellosaurus confirms carries KMT2A::AFF1 though it presents as monocytic rather than the infant B-ALL this project studies. It also holds MEN1 and MLL1 ChIP under five different menin inhibitors in that line, and the six control files total 1.87 GB

  MLL1 ChIP paired with MEN1 across the same five-drug arm

• ← related toSix files titled H2AK119ub rep1 are FLAG ChIPs, so prediction 5's eviction clause - declared UNREACHABLE by Q124 - was computable from bins already on disk. It is now scored and it CONFIRMS: VTP-50469 removes the same share of fusion occupancy in the PCGF1 knockout as in the control, 0.7655 against 0.7554, a 1.3% difference against 22.8% genotype variation on a control drug

  FLAG occupancy of the KMT2A fusion is the measured channel

• ← related toMeasured against a technical floor of 4-10% from two independent input pairs, PCGF1 loss lowers baseline fusion occupancy at the top-FLAG sites by 16.3% and changes the DOT1L inhibitor's effect on that occupancy by 22.8% - while leaving the menin inhibitor's effect unchanged at 1.3%. In the knockout SGC0946 displaces more fusion than VTP-50469 does, which is the opposite of the control

  FLAG occupancy of the KMT2A fusion is the measured channel

• ← related toPCGF1 loss flattens the fusion's genomic distribution - the knockout-over-control share ratio rises monotonically from 0.810 at the highest-signal decile to 1.136 at the lowest while both input pairs stay flat - and the drug does something spatially different: VTP-50469 evicts almost only at the top 0.1% (0.7554 against 0.9937 at the decile below). The drug acts at the peaks, the genotype acts in the bulk

  the genomic distribution of KMT2A-fusion occupancy, stratified by parental signal

• ← related toThe first measurement of any histone mark outside the top 0.1% of fusion sites: by ratio the drug's mark response is confined to the peaks as pre-registered, while in ABSOLUTE share the movement at the decile below is 5 to 33 times larger - though that is partly guaranteed by the stratum being 100x bigger. H3K79me2 falls at the peaks (0.8065) and is not measurably changed at D1, and H3K4me3 and H3K79me2 both break the floor at the lowest decile

  strata are defined by parental KMT2A-fusion occupancy

• ← related toExcluding spleen, lymph node and thymus moves 9 of 29 normal-tissue safety ratios and not one of the worst cases - the artefact is real, concentrated and confined to the middle of the table, and it takes the ACTB reference down with it

  KMT2A 0.97 to 0.92 under the exclusion; its top tissue was thymus

• fuses with →AFF1 / AF4 — the 3' half of the fusion, and a target with almost no structural footing

  t(4;11)(q21;q23), the commonest infant fusion

• part of →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)

  contributes the 5' half: DNA targeting plus the menin grip

• ← part ofWhat is lost: the PHD3/CYP33 switch that lets wild-type KMT2A turn itself off

  the brake the break throws away

• ← related to49 of 91 proteins in this graph have never had a drug-like molecule bound to them

  The naive ligand count for KMT2A is 14 and is wrong: DNA from the CXXC complex, the SAH cofactor, PEG and a detergent. The real count is 4, the MLL1 SET-domain inhibitor series.

• ← related toMLL1 stays on condensed mitotic chromosomes and tethers menin there — a menin-dependent function absent from this graph until cycle 34

  WILD-TYPE MLL1 BOOKMARKS MITOTIC CHROMATIN, a function separable from its H3K4 methylation output.

• ← bindsNUMA1 / NuMA (Q14980) - a nuclear-matrix protein that becomes the spindle pole, and the 5' partner of a variant APL fusion whose coiled-coil is required for oncogenicity

  anti-tag coIP in HEK293T, IntAct MI:0914 association, MI-score 0.35 - below IntAct's own 0.6 threshold, one publication (OpenCell, PMID 35271311), and not resolved to either side of the breakpoint

• ← related toNobody has ever run proximity labelling on a KMT2A fusion, and the interaction records that look like they did are RBP-atlas preys from HEK293

  THE MOST DIRECT INSTRUMENT FOR THE STANDING PRIORITY, never applied to the fusion: no proximity-labelling experiment on any KMT2A fusion is retrievable.

• ← related toThe 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

  KMT2A's own processed chain Q03164-PRO_0000390949 appeared as a partner of another bait and was discarded by my first parser; recovered by accession readback

• ← bindsMEN1 / menin — the adaptor the fusion cannot work without

  menin clamps the KMT2A N-terminal MBM1 motif (Kd ~53 nM)

• ← part ofThe menin–KMT2A grip (MBM1), the interface the drugs actually block

  KMT2A residues 4-15, retained in every fusion

• part of →KMT2A major breakpoint cluster region (BCR1): intron 7 - exon 13, ~8.3 kb

  98% of breaks fall in this region

• part of →KMT2A minor breakpoint cluster region (BCR2): intron 20 - exon 24

  1.4% by PCR, ~6% by targeted sequencing

• ← related toHow many children this affects

  KMT2A rearrangement in 70-80% of infant ALL

• ← related toWhat is a fusion gene?

  worked example of a fusion parent gene

• ← related toHow a chromosome translocation actually forms: proximity + breakage + repair

  how the break and rejoin happen at this locus

• ← part ofThe fusion junction itself as a small-molecule target: a candid verdict

  the neo-surface created where the two genes meet

• ← part ofCXXC-DNA interface of the KMT2A fusion: a validated but hard target

  CXXC domain, retained in every fusion

• ← supportsLooking for a second handle on the retained fragment, and not finding one
• ← related toKMT2B (MLL2/MLL4) - the other protein menin binds, and the one this graph never wrote

  closest paralogue; both carry an N-terminal menin-binding motif and no other KMT2 family member does

• ← related toKMT2E (MLL5) - the family member with a dead SET domain, and the exclusion from COMPASS is justified

  KMT2-family member with a structurally dead SET domain; excluded from COMPASS on catalysis, not on leukaemia relevance

• ← related toOpen ground: the questions in this graph that nobody in the landscape appears to be working on

  Counter-evidence (Yan 2026, PMID 41501053): open-ground area 3 is not untested -- AKAP95 condensates co-condense with the MLL1 translocated fragment, are required for MLL-rearranged leukemogenesis and are already drugged with a phase-separation-disrupting peptide; and the proposed endogenous-tagging FRAP experiment cannot label the fusion without labelling wild-type KMT2A or AFF1

• ← related toPatricia Ernst (University of Colorado) — what normal KMT2A and menin actually do, which is the therapeutic-window question

  mapped the MLL-dependent transcriptional network sustaining normal haematopoietic stem cells beyond HOX genes

• ← related toLEDGF does NOT hold the fusion on chromatin: knocking it down INCREASES fusion occupancy, and an IBD-binding-dead fusion still binds its targets

  LEDGF is required for wild-type KMT2A retention, not the fusion's

• ← related toDoes the fusion need the PWWP arm? An artificial CXXC + H3K36me2-reader + AF4-interaction protein substitutes for a whole MLL fusion - but LEDGF itself is not required for the fusion to reach chromatin

  CXXC-driven targeting is the fusion's own, LEDGF's H3K36 arm belongs to wild-type KMT2A

• ← related toHypothesis: menin inhibition does not merely displace KMT2A::AFF1 from chromatin, it destabilises it — the evicted fusion is degraded, not parked

  Counter-evidence (Liu 2007, PMID 17908926): MLL degradation initiates in the N-terminal ~1400 aa retained by every fusion, yet prevalent MLL fusions are resistant to that degradation — the prior on drug-induced fusion turnover should fall to ~0.1

• ← part ofThe three AT-hooks (aa 169-180, 217-227, 301-309) as a localisation determinant — and the correction that MLL's hooks read DNA *structure*, not AT-richness

  three AT-hooks at 169-180, 217-227, 301-309 of UniProt Q03164

• ← part ofThe CXXC domain (aa 1147-1195, PDB 2KKF) is a chromatin address, not a compartment address — CpG mutants lose transformation but keep their subnuclear position

  CXXC zinc finger at 1147-1195; PDB 2KKF covers 1147-1203

• ← related toWDR5 and KMT2A are at the centrosome and the midbody - a non-chromatin address for both, with a patient-cell phenotype

  MLL/KMT2A localises to the pericentriolar material and the midbody, with a Wiedemann-Steiner patient-cell phenotype

• causes →H4K16ac - the one mark rule 80 actually rescued in this beat: elevated at MLL-AF9 targets in a supplementary figure, and the reason MLL1's own enzyme turns out to be dispensable

  wild-type MLL1 maintains H4K16ac at its target promoters via recruited MOF; its own SET domain is dispensable

• ← related toH3K4me2 - the LSD1 substrate, and the only histone mark ever compared head-to-head between two KMT2A fusion partners in human cells

  written by the KMT2/COMPASS family, whose MLL1 SET domain is dispensable for MLL-AF9 leukaemogenesis

• ← related toH3.3 - the graph's most-named epigenetic object with no entry, 325 mentions, and not one measurement in a KMT2A-rearranged cell

  no H3.3 measurement exists in any KMT2A-rearranged system; bounded by 1 abstract record and 19 full texts

• ← related toHypothesis: the cell-of-origin methylome, not the breakpoint, decides whether a KMT2A::AFF1 leukemia is HOXA-high or HOXA-low

  Counter-evidence (Okuda 2014, PMID 24465000): fusion targeting needs non-methylated CpG AND H3K36me2/3, so target choice tracks genes that are already transcriptionally active, not only an inherited methylome.

• ← related toEnhancer usage in KMT2A-rearranged leukemia — occupancy and amplification, not hijacking, and not super-enhancers

  Counter-account (Okuda 2014, PMID 24465000): KMT2A fusions recruit coactivators to previously active CpG-rich promoters via two chromatin-reading modules, which explains much of the 'enhancer' signal as a consequence of prior promoter activity.

• ← related toDataset: PDB and AlphaFold - the structural census, including the two absences that shape every plan

  AlphaFold returns HTTP 404 for Q03164 and local fragment prediction is not a clean substitute: AlphaFold's confidence collapses on intrinsically disordered regions (Ruff & Pappu, PMID 34418423), which is most of KMT2A outside its folded modules.

• ← related toThe LEDGF IBD–KMT2A grip: the second half of the clamp, and nobody is drugging it

  Counter-evidence (Van Belle 2021, PMID 33477970): the PSIP1 paralogue HRP-2/HDGFL2 carries both PWWP and IBD and binds MLL through a conserved interface that is LESS menin-dependent than MLL-LEDGF, leaving a menin-independent IBD-family route to MLL that an IBD-selective ligand would not close.

• ← related toMAT2A - the enzyme that makes the methyl donor, and the one metabolic dependency in this disease with two independent papers and a lymphoid PDX result

  supplies the methyl donor for the SET domain and for the fusion-driven hypermethylated state

• ← related toIs the metabolic phenotype a consequence of the fusion programme or an independent dependency? The answer is that BOTH directions are published, and one of them points the other way entirely

  frames whether the fusion programme causes or merely coexists with the metabolic phenotype

• ← part ofThe KMT2A::MLLT1 chimeric protein: exactly what survives the join on each side, in residue numbers

  5' contribution: MBM, IBM1/2, AT-hooks, CXXC (UniProt Q03164)

• ← related toThe KMT2A::MLLT3 chimeric protein: what survives the join on each side, in residue numbers - and the four AF9 modules that do not

  Q03164 coordinates, fetched and read back 2026-09-20

• ← related toMYC - named in a hundred entries of this graph and, until now, written in none

  DOWNSTREAM, NOT A PARTNER: no KMT2A::MYC fusion is claimed; MYC is carried here as an effector of the fusion programme's output.

• ← regulatesIKK/NF-kB retains MLL on chromatin — BOTH the fusion and wild-type MLL, at two promoters, in mouse myeloid AML, and it is occupancy loss rather than protein loss

  IKK inhibition strips wild-type MLL and H3K4me3 from Hoxa9/Meis1

• ← part ofKMT2A::SEPTIN6 (MLL-SEPT6), t(X;11)(q24;q23) and its cryptic insertion variants

  retained KMT2A N-terminus (MBM1/2, AT-hooks, CXXC)

• fuses with →SEPTIN6 (septin-6) — cytoskeletal GTPase, KMT2A fusion partner in t(X;11)(q24;q23)

  t(X;11)(q24;q23) and cryptic ins(X;11)

• ← part ofKMT2A::SEPTIN9 (MLL-SEPT9 / MLL-MSF), t(11;17)(q23;q25)

  retained KMT2A N-terminus

• fuses with →SEPTIN9 (septin-9 / MSF) — cytoskeletal GTPase, KMT2A fusion partner in t(11;17)(q23;q25)

  t(11;17)(q23;q25)

• ← fuses withSEPTIN6 (septin-6) — cytoskeletal GTPase, KMT2A fusion partner in t(X;11)(q24;q23)

  KMT2A::SEPTIN6, t(X;11)(q24;q23)

• ← fuses withSEPTIN9 (septin-9 / MSF) — cytoskeletal GTPase, KMT2A fusion partner in t(11;17)(q23;q25)

  KMT2A::SEPTIN9, t(11;17)(q23;q25)

• ← part ofKMT2A::SEPTIN5 (MLL-SEPT5 / MLL-hCDCrel-1), t(11;22)(q23;q11.2)

  retained KMT2A N-terminus

• fuses with →SEPTIN5 (septin-5 / CDCrel-1) — KMT2A fusion partner in t(11;22)(q23;q11.2)

  t(11;22)(q23;q11.2)

• ← part ofKMT2A::SEPTIN2 (MLL-SEPT2), t(2;11)(q37;q23) — the therapy-related septin

  retained KMT2A N-terminus

• fuses with →SEPTIN2 (septin-2 / NEDD5) — KMT2A fusion partner in t(2;11)(q37;q23)

  t(2;11)(q37;q23)

• ← part ofKMT2A::SEPTIN11 (MLL-SEPT11) — the fusion that cytogenetically impersonates t(4;11)

  retained KMT2A N-terminus including CXXC

• fuses with →SEPTIN11 (septin-11) — KMT2A fusion partner at 4q21, adjacent to AFF1

  t(4;11)(q21;q23), impersonating KMT2A::AFF1

• ← part ofKMT2A::ABI1 (MLL-ABI1), t(10;11)(p11.2;q23) — the partner that binds an elongation partner

  retained KMT2A N-terminus

• fuses with →ABI1 (Abl interactor 1 / e3B1) — WAVE-complex cytoskeletal adaptor, KMT2A partner, and an ENL-binding protein

  t(10;11)(p11.2;q23)

• ← depends onHow a cytoskeletal partner reaches chromatin at all — the KMT2A N-terminus drags it in, and the traffic runs both ways

  nuclear import is supplied entirely by the retained KMT2A N-terminus

• ← depends onWhy a nucleoporin fusion needs menin-MLL1: the second menin-dependent fusion family, and what it borrows rather than carries

  wild-type MLL1 is the borrowed cofactor

• ← fuses withNUP98::KMT2A, inv(11)(p15q23) — the NUP98 fusion that contains KMT2A and is menin-inhibitor RESISTANT

  NUP98 FG repeats to almost the entire KMT2A ORF, exon 1 deleted

• ← bindsFive KMT2A-free NUP98 fusions bind MLL1 and the NSL complex — and one unbiased mass-spectrometry study failed to reproduce it

  GLFG-dependent binding of wild-type MLL1 by five KMT2A-free NUP98 fusions

• ← testsTwo constructs that would decide whether menin dependency is about the menin-binding motif or about HOXA level

  MBM-deleted endogenous Kmt2a allele as the decisive reagent

• ← related to756 children with 11q23/MLL-rearranged AML: 5-year EFS runs from 11% to 92% depending on the partner

  outcome varies by which partner joins the retained KMT2A N-terminus

• ← testsThe clinical outcome data do not support a single KMT2A class, and they do not refute the menin interface either

  whether KMT2A fusions behave as one clinical class

• ← part ofKMT2A::GPHN (MLL-gephyrin), t(11;14)(q23;q24) — the 15-residue counterexample

  retained KMT2A N-terminus: AT-hooks and CXXC (MBM1/2 retained by position, not stated by the source)

• fuses with →GPHN (gephyrin) — synaptic scaffold and molybdenum-cofactor enzyme, KMT2A fusion partner in t(11;14)(q23;q24)

  t(11;14)(q23;q24), two published cases

• ← part ofKMT2A::MLLT11 (MLL-AF1q), t(1;11)(q21;q23) — 28 patients, and not one functional experiment

  retained KMT2A N-terminus (MBM1/2, AT-hooks, CXXC)

• fuses with →MLLT11 / AF1q — a 90-residue protein with no chromatin domain and no oligomerisation module, KMT2A partner in t(1;11)(q21;q23)

  t(1;11)(q21;q23); 28 patients, rank 10 of all direct KMT2A partners

• ← part ofKMT2A::EPS15 (MLL-AF1p), t(1;11)(p32;q23) — the infant-ALL cytoplasmic partner

  1368 N-terminal residues of HRX/MLL: AT-hook, snRNP and methyltransferase-homology regions

• fuses with →EPS15 / AF1p — domain architecture, and the one coordinate nobody has published

  t(1;11)(p32;q23); 55 patients, rank 8, 22 of them infant ALL

• ← part ofKMT2A::GAS7 (MLL-GAS7), t(11;17)(q23;p13) — three patients, and the best mouse model in the partner class

  retained KMT2A N-terminus (MBM1/2, AT-hooks, CXXC)

• fuses with →GAS7 — growth-arrest-specific F-BAR protein, the second pillar of the forced-oligomerisation model

  cryptic t(11;17)(q23;p13); 3 patients in 3401

• ← part ofKMT2A::MLLT6 (MLL-AF17), t(11;17)(q23;q12-q21) — the bridge case: a leucine zipper on a DOT1L-complex subunit

  retained KMT2A N-terminus (MBM1/2, AT-hooks, CXXC)

• fuses with →MLLT6 (AF17) — the fourth KMT2A fusion partner inside DotCom, and the only one no cell needs

  t(11;17)(q23;q12-q21); 27 patients, rank 11, adult-AML-weighted

• ← part ofKMT2A::MYO1F, t(11;19)(q23;p13.2) — recurrent in infant AML, and never functionally tested

  retained KMT2A N-terminus (MBM1/2, AT-hooks, CXXC)

• fuses with →MYO1F — unconventional myosin IF, a KMT2A partner with neither a transactivation domain nor a dimerisation domain

  t(11;19)(q23;p13.2); 5 patients, all AML, 4 infant; recurrent in infant AML

• ← part ofThe "second breakpoint cluster region" of KMT2A (Meyer 2019): the founding report of BCR2, not a third cluster

  KMT2A intron 19/20 to exon 24, DCAL exon numbering

• fuses with →USP2 (ubiquitin carboxyl-terminal hydrolase 2) - the ninth commonest KMT2A partner and the only one that is a deubiquitinase

  inv(11)(q23.3q23.3), 30 patients, rank 9 of all KMT2A partners; minor BCR only

• ← part ofKMT2A::USP2, inv(11)(q23.3q23.3) - the minor-BCR fusion that keeps the PHD fingers

  5' KMT2A through the minor BCR: AT-hooks, MBM1/2, CXXC AND the complete PHD1-3 cluster, bromodomain and ePHD4

• fuses with →USP8 (UBPY) - the second deubiquitinase KMT2A partner, three patients, same UCH logic

  t(11;15)(q23;q21), 3 patients in 3401; minor BCR

• ← part ofKMT2A::USP8, t(11;15)(q23;q21) - three patients, and the second instance of a deubiquitinase partner

  5' KMT2A through the minor BCR, retaining PHD1-3, bromodomain and ePHD4

• fuses with →MAML2 (mastermind-like 2) - the Notch co-activator that is a KMT2A partner in leukaemia and in thymoma

  inv(11)(q21q23), 4 patients in the recombinome and at least 8 individually reported; also the recurrent fusion of type B2/B3 thymoma

• ← part ofKMT2A::MAML2, inv(11)(q21q23) - a therapy-related fusion that also causes a solid tumour

  KMT2A exon 7-11, i.e. a MAJOR-BCR fusion: AT-hooks, MBM1/2 and CXXC retained, PHD fingers lost

• ← related toMixed phenotype acute leukaemia (MPAL) and acute leukaemia of ambiguous lineage

  KMT2A rearrangement is over-represented in paediatric ambiguous-lineage leukaemia and concentrates in the poor-outcome hybrid-treatment group

• ← related toThree pairs of twins who both got leukaemia: the patients who proved the fusion happens before birth

  clonal non-constitutional KMT2A rearrangement identical within each twin pair; breakpoints fall in introns across about 10 kb and differ between individuals

• ← related toPhase-2 dossier: the CXXC-DNA interface - two structures, a genetically validated requirement, and a selectivity problem that decides everything

  LOEUF 0.070 - the organismal warning against whole-protein KMT2A strategies

• ← bindsCK2 phosphorylation of KMT2A S136/S142 tunes the LEDGF grip - a second, non-menin, phospho-gated handle on the fragment every fusion keeps

  LEDGF IBD binds the KMT2A IBM (111-160); phospho-gated at S136/S142

• ← related toS821 is not phosphorylated in any dataset we could query: the 14-3-3 motif candidate is dead as stated

  phosphosite audit of Q03164

• ← regulatesCSNK2A1 (casein kinase 2 alpha) - the one kinase that turns up three separate times in KMT2A biology

  phosphorylates S136/S142 (retained) and T2724/S2726 (lost)

• ← related to56 of KMT2A's 130 observed phosphosites survive the translocation, and the retained fragment is phospho-denser than the part that is lost

  174 modified residues on Q03164 split at the breakpoint

• ← regulatesATR phosphorylates KMT2A at S516 to break the SCF(Skp2) contact - a phospho-gated degradation switch sitting inside the retained fragment

  ATR phosphorylation of S516 blocks SCF(Skp2)-mediated degradation

• ← related toNobody has ever shown 14-3-3 binding directly to KMT2A: five records across every paralogue alias, and one of them is about AF4

  counted absence over all seven 14-3-3 paralogue aliases

• ← bindsThe phosphatase side: SET binds both wild-type KMT2A and the fusion, and inhibiting it changes four kinases at once

  SET interacts with both wild-type KMT2A and the fusion; interface unmapped

• ← bindsCks1 and Cks2 hold both halves of KMT2A and set its level through the cell cycle

  Cks1/Cks2 bind both MllN and MllC; retained/lost undetermined

• ← related toEnumeration: every phosphorylation-dependent and kinase-regulated contact on KMT2A and its fusions that this beat could find

  the beat's index over KMT2A phospho-contacts

• ← related toH3K4me3 - the mark KMT2A's own SET domain writes, that the fusion cannot write, that an interface in this graph already binds, and that switches PRC2 off

  written by KMT2A's SET domain, which the fusion loses

• ← related toThe 50-RBP BioID2 proximity map in K562 - the first modern proximity dataset this graph can reach, and AFF1 is in it as prey

  KMT2A appears as prey in twenty-five of fifty bait neighbourhoods

• ← co-occurs withZRANB2 - the only bait in any published proximity experiment that reaches AFF1, AFF4, ELL, MLLT1, CCNT1, KMT2A and DOT1L at once

  KMT2A recovered as prey

• ← co-occurs withSBDS - a ribosome-maturation factor whose loss causes a leukaemia-predisposition syndrome, and whose BioID neighbourhood contains AFF1, KMT2A, CDK9, PAF1 and LEDGF

  KMT2A recovered as prey

• ← related toWild-type KMT2A is one of the most promiscuously labelled proteins in the proximity record - 25 of 50 baits - and menin is one of the least, at 3

  wild-type KMT2A recovered by 25 of 50 baits

• ← related toKMT2A sits in the 14-3-3 proximity neighbourhood, and the YWHAG link reproduces across two independent studies

  KMT2A recovered by six 14-3-3 isoforms; YWHAG reproduces across two studies

• ← related toA 109-transcription-factor BioID network puts KMT2A next to MYC, YY1 and the ETS and KLF families - and AF10 next to ERG and KLF5

  KMT2A recovered as prey by eleven transcription-factor BioID baits

• ← co-occurs withGPKOW - a G-patch spliceosomal protein whose BioID neighbourhood contains AFF4, ELL, ENL, CDK9, KMT2A and DOT1L

  recovered as prey in the GPKOW BioID2 neighbourhood (proximity, not binding)

• ← co-occurs withPPIL4 - a spliceosomal cyclophilin whose BioID neighbourhood is the widest overlap with this graph's subject matter of any bait

  recovered as prey in the PPIL4 BioID2 neighbourhood (proximity, not binding)

• ← co-occurs withHNRNPC - the nascent-RNA packaging protein whose BioID neighbourhood contains AFF4, ELL, ENL, CDK9, cyclin T1 and KMT2A

  recovered as prey in the HNRNPC BioID2 neighbourhood (proximity, not binding)

• ← co-occurs withSMNDC1/SPF30 - a Tudor-domain splicing factor recovering ELL, ENL, KMT2A, PAF1 and KAT7

  recovered as prey in the SMNDC1 BioID2 neighbourhood (proximity, not binding)

• ← co-occurs withAGGF1 - an angiogenic G-patch protein whose BioID neighbourhood contains ELL, KMT2A, LEDGF, WDR5, KAT7 and RNA polymerase II

  recovered as prey in the AGGF1 BioID2 neighbourhood (proximity, not binding)

• ← related toNUTM1 (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

  CONTRAST, not analogy: the 3' partner of the KMT2A-WILD-TYPE infant B-ALL. No reported contact with KMT2A, menin, the MBM, CXXC or the WIN site; the two fusion families share an age band, a disease and a posterior-HOXA output and share neither a mark nor a mechanism

• ← related toPRMT5 - the type II arginine methyltransferase this graph has reached for sixteen times, whose mark on H3R2 RECRUITS WDR5 to chromatin

  MLL1 is named in the PAFc/HOXA9/STAT5 unit activating PRMT5; no physical contact reported

• ← inhibitsPRMT6 - the enzyme whose mark is MUTUALLY EXCLUSIVE with the MLL complex on the H3 tail, and which shuts HOX genes down by evicting WDR5

  loss of promoter H3R2me2a ENHANCES KMT2A binding; the mark restrains it

• ← regulatesEHMT2 / G9a - the H3K9me1/2 writer that sets whether the KMT2A LOCUS ITSELF rearranges, and whose loss delays MLL-AF9 AML in a supplementary figure nobody can find

  H3K9me1/2 balance at the KMT2A LOCUS controls its amplification and rearrangement

• ← regulatesH3K9me1 - the only H3K9 state whose level at the KMT2A breakpoint cluster region has ever been measured, and it is LOW there until KDM3B is removed

  H3K9me1 level across the KMT2A breakpoint cluster region gates copy gain and break-apart

• ← regulatesKDM3B - the H3K9me1/2 demethylase that sits on the KMT2A breakpoint cluster region and keeps it from rearranging, and the only leukaemia gene in this graph whose direction of effect flips between the myeloid and the lymphoid literature

  KDM3B occupies the KMT2A breakpoint cluster region and its loss causes site-specific copy gain and break-apart

• ← causesThe H3K9me1/2-CTCF brake: a KDM3B-G9a balance keeps the KMT2A breakpoint cluster region methyl-poor and CTCF-bound, and losing either end makes the locus amplify and rearrange - in epithelium, in primary cord-blood progenitors and in mice

  site-specific KMT2A copy gain and break-apart, transient at 12 h and inherited by 72 h

• ← causesJDI-12 - published as a JMJD1C inhibitor that preferentially kills KMT2A-rearranged leukaemia and spares cord blood, then used as 'KDM3i' to CREATE KMT2A rearrangements in cord-blood progenitors, by two papers that do not cite each other

  produces site-specific KMT2A copy gains and break-apart in RPE, KG1a, a primary AML line, an AML organoid and cord-blood CD34+ HSPCs

• ← regulatesH3R2me2a - the asymmetric arginine mark two residues from H3K4 that LOCKS the WRAD module off the histone tail, and the first arginine mark node in this graph

  loss of PROMOTER H3R2me2a enhances wild-type KMT2A binding and H3K4me3 deposition

• ← testsMeasure H3R2 and H4R3 methylation states at KMT2A-fusion targets by isomer-resolving mass spectrometry, and test the gate by writer perturbation read out on the enzyme's product - the experiment that has never been done in any fusion

  allele-resolved occupancy: the substrate-side gate should restrain the SET/WIN/PHD-competent wild-type allele and leave the fusion untouched

• ← related toH4K20me2 - the abundant repair-linked H4 mark, the 53BP1 docking site, and the first H4K20 state in this graph; never once measured in a KMT2A-rearranged cell

  BOUNDED ABSENCE: H4K20[tiab] AND KMT2A[tiab] = 0 (parsed zero); H4K20[tiab] AND MLL[tiab] = 1, and that one record is a NEGATIVE - ASH1L's distribution does not resemble H4K20

• ← co-occurs withAFG2A / SPATA5 (Q8NB90) - a 55LCC replisome-proteostasis ATPase that co-purifies with KMT2A and menin, and whose CURRENT symbol cannot retrieve its own source paper or any of its function literature

  one HeLa S3 tandem affinity purification (PMID 38554706); AFG2A is the BAIT and KMT2A the prey; IntAct MI-score 0.35

• ← co-occurs withAFG2B / SPATA5L1 (Q9BVQ7) - the second 55LCC ATPase, the only read-queue member with an annotated nuclear pool, pan-essential in 95.5% of lines, and the one that touches PSIP1

  same single HeLa S3 TAP (PMID 38554706); AFG2B bait, KMT2A prey; MI-score 0.35

• ← co-occurs withYWHAG / 14-3-3 gamma (P61981) - the only cross-laboratory reproduced KMT2A proximity partner in the curated record, and UniProt places it nowhere near the nucleus

  two INDEPENDENT BioID studies, two labs (36931259 HEK293 Flp-In; 39251607 K562); YWHAG bait, KMT2A prey in both; MI-score 0.41. The only row pair in the read queue that is not one screening programme seen twice

• ← related toThe 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

  largest absolute gap: 218 PSICQUIC rows against 376 from IntAct REST (58%), 158 rows unseen, including the taspase1 processed-chain identifiers Q03164-PRO_0000390949 and -390950

• ← part ofRun 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

  canonical exon 11 of ENST00000534358, inside the major breakpoint cluster region

• ← related toTRIB3 / TRB3 (Q96RU7) - a pseudokinase that is the ISR's own brake, falls 3.1 log2 under either inhibitor, and has a published physical link to the MLL complex in the wrong tissue

  UNREAD LEAD: PMID 36142452 reports TRIB3 interfering with the MLL complex in breast cancer cells - title-level only, recorded as a lead not as evidence

• binds →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

  HRX/MLL proteins interact directly with GADD34 by yeast two-hybrid and co-IP in human cells (PMID 10490642, 1999, transfection/overexpression, never reproduced)

• ← related toRead 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

  canonical exon 11 of ENST00000534358 is the break-hotspot element

• ← related toThe 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

  the KMT2A BCR, element, exon 11 and 90 kb gene span carry zero H3K9me2 domain coverage in 3 of 3 CD34+ ChIPs

• ← part ofThe KMT2A exon-11 CTCF anchor, placed in its population: baseline strength 4.92x (91.7th percentile of 12,443 RPE anchors), KDM3B-depletion ratio 0.396x, rank 58 of 1,625 strength-matched anchors - a strong, strongly-responding anchor at the edge of a continuum, not an outlier

  inside Ensembl canonical KMT2A exon 11 (ENST00000534358), 3.8 kb from the DCAL functional cut point

• ← related toRead 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.

  seven KMT2A intervals, all NOT OCCUPIED in both lines

• ← part ofTwo 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

  the element lies inside canonical KMT2A exon 11

• ← related toThe MCF-7 arm, and the clause that saved it: SUV39H1 at the exon-11 anchor reaches the 96.3rd percentile of 56,171 same-cell-type CTCF anchors in one replicate - but all six non-anchor KMT2A intervals are elevated too (medians 73-87), so the pre-registered FIRES clause does not fire, the verdict is NEITHER, and four replicates of one antibody in one cell line span 51.8 to 96.3 at the same 147 bp

  the whole 90 kb gene body is elevated, not the anchor

• ← related toTHE 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

  KMT2A's 90 kb body carries 0.77-0.94x the H3K9me3 of 385 gene bodies matched on length and MCF-7 expression, and 0.65-0.89x of 661 matched silent ones

• ← related toThe H3K9 machinery's address at KMT2A is the PROMOTER, not the exon-11 breakpoint anchor: G9a, HP1a and KAP1 all read the top decile of active promoters there in MCF-7 and KAP1's peak replicates a K562 peak to 267 bp, while five proteins across five cell types are ORDINARY or absent at the anchor - so every H3K9-directed intervention this graph has discussed is aimed 52 kb from where the machinery is

  the H3K9 module s address at this gene is its promoter, not its breakpoint cluster region

• ← related toKDM2B, the PRC1.1 CpG-island recruiter, sits at the 99.7th percentile of its own distribution on the fusion's strongest sites in DMSO and does not gain share under either a menin inhibitor or a DOT1L inhibitor - it holds position while comparable bins drift down

  KDM2B is a documented member of the KMT2A-fusion interactome and this is its chromatin relationship to the fusion under drug

• ← related toThe falsifier that decides the beat, run: 65.6% of the fusion's own top sites overlap a CpG island and 1 of the 52 sites it formally 'gains' under a menin inhibitor does - 1.35-fold over a class-matched null, Fisher p = 0.53 - so the gained set does not carry the signature of CXXC-driven binding and the redistribution model is refuted on its own predicted substrate

  the CXXC domain of KMT2A is retained in the fusion and is the binding activity this test scores

• ← related toAt the same bins in the same cells, menin's share ratio is 0.20-0.26 and the MLL1 N-terminus antibody's is 0.80-0.91 - so a shared-epitope MLL1-N ChIP under-reports the menin inhibitor's occupancy effect by 3.5 to 4.0-fold, and the number is now measured rather than asserted

  the MLL1 N-terminus epitope (Bethyl A300-086A) is retained in the fusion, so the track sums the fusion and wild-type KMT2A

• ← related toAn MLL1 N-terminus ChIP sees the fusion and wild-type MLL1 in one track, and a menin ChIP has the same problem from the other side - so the standard reagents of the menin-inhibitor literature cannot attribute occupancy to the fusion, and the masking factor is now measured at 3.5 to 4.0-fold

  the N-terminus is retained in every KMT2A fusion, which is why an N-terminal antibody cannot separate the fusion from wild-type KMT2A

• ← related toAt the 692 RS4;11 bins where an MLL1 N-terminus and menin co-occupy, VTP-50469 takes menin to 0.20 and DOT1L to 0.20, drops H3K79me2 to 0.66, and leaves MLL1-N-reactive material at 0.83 - four factors, one drug, one chromatin, and a clean ordering

  the MLL1 N-terminus track is the one that does not move, and it cannot be attributed to the fusion

• ← regulatesRUNX2 - 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

  RUNX2 reciprocally binds the KMT2A promoter in KMT2A-R T-ALL, a feed-forward loop with the chimera

• ← inhibitsZMYND11 (BS69) - an H3.3K36me3 reader that restrains elongation, is itself the 5' partner of a chromatin-hijacking AML fusion driving HOXA/MEIS1/MYB, and whose promoter is the third-strongest MLL1-N bin in RS4;11

  2025 PREPRINT: ZMYND11 interacts with and inhibits KMT2A/MLL1 in neurons; a ZRSID point mutation abrogates the contact and revumenib attenuates the loss phenotype

• ← related toAt the MLL1-N HELD SET, floor-normalised, the MLL1-N channel is inert to three decimals in both human lines (1.036 and 1.0359) - AND THAT IS A STATEMENT ABOUT THAT POPULATION, NOT ABOUT THE CHANNEL: at MLL1's own top sites in mouse the same channel reads 0.3012 under a DOT1L inhibitor and 0.3298 under a menin inhibitor

  the Bethyl A300-086A N-terminal epitope is retained in every KMT2A fusion, so this ChIP channel is KMT2A wild-type plus fusion and separates neither

• ← related toThe cheap experiment nobody has run: an MLL1 C-TERMINAL ChIP beside the standard N-terminal one in a single human KMT2A-rearranged line under a menin inhibitor - the C-terminus is lost in the fusion, so the difference of the two tracks is the fusion, with no tag and no knock-in

  exploits the fusion's own architecture: the C-terminal SET domain is lost in every KMT2A fusion, so a C-terminal epitope is wild-type-specific by construction

• ← related toIn the only system where the fusion can be measured separately, a menin inhibitor removes MLL1 from wild-type-dominated sites harder than it removes the tagged fusion from its own sites - 0.1069 against 0.1793 - which is an on-target effect on normal KMT2A-menin complexes and an argument about toxicity nobody in this project has made from occupancy data

  the wild-type protein, measured under a menin inhibitor separately from the fusion for the first time

• ← testsThree blots and one drug curve: does the ZMYND11::MBTD1 fusion protein still bind KMT2A, and does a menin inhibitor work in a leukaemia with no KMT2A rearrangement

  arm A blots endogenous KMT2A at both epitopes, so it also asks which taspase1 fragment ZMYND11 binds

• regulates →Quiescence in the blood stem cell is actively held there by MLL1 - and losing MLL1 pushes the stem cell INTO cycle while pulling the progenitor OUT of it

  MLL1 maintains HSC quiescence; loss causes ectopic cell-cycle entry (PMID 18371366)

• regulates →The cell types that need MLL1 are the uncommitted ones - dependence ends exactly where lineage commitment begins

  Mll dependence is restricted to uncommitted haematopoietic compartments (PMID 18371366)

• ← related toThe two founding conditional Mll knockouts were published back to back and they do not agree: fatal marrow failure in three weeks, or grossly normal blood

  the two founding conditional knockouts of this gene, neither previously cited in the corpus

• regulates →KMT2A is called an H3K4 methyltransferase, and it is required for the mark at under 5% of the promoters that carry it

  KMT2A's own H3K4me3 requirement is under 5% of marked promoters (PMID 19703992)

• regulates →H3K4me3 oscillates around the clock and MLL1 is what lets it - a cell state of KMT2A biology this project has never once mentioned

  MLL1 is essential for cyclic H3K4me3 at circadian promoters (PMID 21113167)

• ← depends onMenin's other address list: the same complex that switches on HOXA in blood writes H3K4 methylation onto the CDK inhibitors p27 and p18 in endocrine cells

  MLL is the enzyme menin recruits to these promoters

• ← related toThe tissue selectivity of MEN1 syndrome is not in MEN1: menin is expressed evenly everywhere and MLL, p27 and p18 are the endocrine-predominant ones

  MLL, not menin, is the tissue-variable element of the module

• ← related toMeasured: the cell-type, cell-state and epigenetics literature for KMT2A and MEN1 was NOT missing from this corpus - 7 of 16 load-bearing papers were already cited, one of them 32 times, and 0 of 16 were in gene-kmt2a

  0 of 16 load-bearing sources were in this entry before cycle 89

• ← related toAll nine unwritten MEN1 partners trace to exactly two publications, neither of which mentions menin outside an antibody catalogue row

  RL3_RL3L_HUMAN, the non-symbol in the KMT2A absent list, resolved to IntAct ambiguous protein group EBI-25396188

• ← related toKMT2A-ABT1 is one spoke-expanded BioPlex 3.0 AP-MS row in HEK293T in which ABT1 is the bait and KMT2A the prey, and ABT1's nucleolus is ECO:0000250 only - declined

  adjudicates the IntAct KMT2A-ABT1 rows; DECLINED - spoke-expanded BioPlex AP-MS, ABT1 bait, KMT2A prey, nucleolar annotation ECO:0000250 only

• ← related toKMT2A-KRR1 is one spoke-expanded BioPlex 3.0 AP-MS row in HEK293T with KRR1 as bait and KMT2A as prey, sitting in a seven-member ribosomal block at the floor score - declined

  adjudicates the IntAct KMT2A-KRR1 row; DECLINED - one of seven ribosomal baits at the 0.35 floor in one screen

• ← related toKMT2A-ANOS1 is one spoke-expanded BioPlex 3.0 AP-MS row with ANOS1 - a secreted, cell-membrane axon-guidance protein - as the bait and KMT2A as the prey in HEK293T - declined

  adjudicates the IntAct KMT2A-ANOS1 row; DECLINED - a secreted, cerebellum-expressed protein as bait in embryonic kidney cells

• ← related toKMT2A-FGFBP1 is spoke-expanded BioPlex 3.0 AP-MS recovered in BOTH HEK293T and HCT116 - it passes BioPlex's own replication criterion and the partner is a secreted extracellular protein - declined

  adjudicates the two IntAct KMT2A-FGFBP1 rows; DECLINED - replicated in HEK293T and HCT116 and the partner is still a secreted protein

• ← related toKMT2A-OASL is one spoke-expanded BioPlex 3.0 AP-MS row with OASL as bait and KMT2A as prey; OASL is the only partner in the beat natively expressed in blood leukocytes, and its nucleolar annotation is uncoded - declined, with the compartment argument unavailable

  adjudicates the IntAct KMT2A-OASL row; DECLINED on evidence thinness only - OASL is haematopoietically expressed and annotated nucleolar, so no compartment argument applies

• ← related toKMT2A-PRKRA is one spoke-expanded BioPlex 3.0 AP-MS row with PRKRA (PACT) as bait and KMT2A as prey; PRKRA's only UniProt compartments are cytoplasm and perinuclear region, both uncoded - declined

  adjudicates the IntAct KMT2A-PRKRA row; DECLINED - cytoplasmic dsRNA-binding bait, no nuclease control recorded

• ← related toKMT2A-TMCO1 is a single DSSO crosslink in one detergent fraction of isolated U2OS nuclei with no bait on either side; TMCO1 is an ER-membrane calcium channel and the source paper attributes 4% of its links to nucleus-ER contamination - declined

  adjudicates the IntAct KMT2A-TMCO1 row; DECLINED - a bait-free DSSO crosslink to an ER-membrane calcium channel in isolated nuclei

• ← related toThe 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

  bounds but does not exclude a nucleolar pool for KMT2A; no measurement of that compartment exists in this graph

• ← related toThe 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

  KMT2A is an IntAct bait in 96 of 376 rows across 18 publications, the most recent April 2015; every partner curated since is by-catch from someone else's bait

• ← testsThe missing reagent: a KMT2A-fusion or wild-type-KMT2A proximity bait in a KMT2A-rearranged leukaemia line - IntAct has curated no KMT2A-bait interaction experiment since April 2015 and no PSIP1-bait since October 2012

  would produce the first KMT2A-bait proximity dataset in a KMT2A-rearranged cell; pre-specified readouts include whether any nucleolar protein survives a size-matched control

• ← co-occurs withOGT - the O-GlcNAc transferase, a Polycomb-complexome bait that pulls down KMT2A, and already the metabolic dial on DOT1L in this graph

  OGT bait / KMT2A prey in one spoke-expanded TAP-MS row, IntAct IM-21659, HEK293 - co-purification, not a demonstrated contact

• ← co-occurs withYY1 - the enhancer-promoter looping factor that recovers KMT2A in a 109-transcription-factor BioID screen

  YY1 bait / KMT2A prey, one spoke-expanded BioID row (MI:1314), IntAct IM-28767, HEK293 Flp-In - proximity, not binding

• ← co-occurs withLIN28B - a one-row BioID prey of KMT2A that a 2026 Blood paper shows is REQUIRED for KMT2A::AFF1 to transform human fetal cells

  LIN28B bait / KMT2A prey, one spoke-expanded BioID2 row, IntAct IM-30059, K562 - proximity only

• ← co-occurs withQKI - a KMT2A proximity row in K562, and a tumour suppressor whose loss concentrates in the HOXA subtype of T-ALL

  QKI bait / KMT2A prey, one spoke-expanded BioID2 row, IntAct IM-30059, K562 - proximity only

• ← co-occurs withELF2 / NERF - an ETS factor that works with RUNX1 on a B-cell kinase promoter, and one of eleven TF baits recovering KMT2A

  ELF2 bait / KMT2A prey, one spoke-expanded BioID row, IntAct IM-28767, HEK293 Flp-In - proximity only

• ← related toH3K4me3 is the one mark in GSE260742 with no declared second channel at all - six samples, no replicate suffix anywhere - so no same-condition pairing exists and its noise floor remains a transfer from another mark whatever else is measured.

  H3K4me3 is the wild-type KMT2A SET domain's product

• ← related toPrey 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

  KMT2A is the only core protein recovered above its size-and-abundance prediction, and only in the RBP/K562 map

• ← related toZC3H8 / Fliz1 - the last little-elongation-complex subunit without an entry, discovered in fetal liver haematopoietic progenitors, and a selective BioID2 bait that reaches KMT2A in K562

  ZC3H8 bait / KMT2A prey, one MI:1314 BioID2 spoke row, IM-30059, K562; p = 0.185, NOT RESOLVED

• ← related toELF1 - a direct transcriptional target of MLL fusion proteins and HOXA9 that is ALSO a BioID neighbour of KMT2A, and that is focally deleted in paediatric AML without a KMT2A rearrangement

  ELF1 bait / KMT2A prey, one MI:1314 BioID spoke row, IM-28767, HEK293 Flp-In; p = 0.151, NOT RESOLVED

• ← related toNFYC / 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

  NFYC bait / KMT2A prey, one MI:1314 BioID spoke row, IM-28767; p = 0.084, the lowest of the eleven KMT2A-TF pairs, still NOT RESOLVED

• ← related toDDX55 - a large-subunit rRNA helicase that is nucleolar by imaging and nucleoplasmic by its own founding paper, and the narrowest RBP bait to reach KMT2A after DDX52

  DDX55 bait / KMT2A prey, one MI:1314 BioID2 spoke row, IM-30059, K562; p = 0.162, NOT RESOLVED

• ← related toNHLH1 / 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

  NHLH1 bait / KMT2A prey, one MI:1314 BioID spoke row, IM-28767; p = 0.128, NOT RESOLVED, and NHLH1 has no trustworthy localisation

• ← related toKMT2A-EN1 declined: one spoke-expanded BioID row from an ectopic neural homeobox construct in a kidney cell, and the only screen that ever looked found EN1 unexpressed in AML lines

  declined: one spoke-expanded BioID row from an ectopic neural homeobox construct in a kidney cell

• ← related toKMT2A's proximity promiscuity is BAIT-CLASS-SPECIFIC: it is recovered at twice its size-and-abundance prediction by RNA-binding-protein baits in a leukaemia cell and at exactly its prediction by transcription-factor baits in a kidney cell

  a claim about where KMT2A resides rather than about what it binds

• ← related toResolving cycle 83's threat: the MLL1 channel moves under both drug classes and is therefore not diagnostic, but the MEN1 channel separates them completely - so 'menin displacement' remains a specific readout of menin-interface engagement and the corpus must stop treating the two channels as one

  the MLL1-N channel reports wild-type MLL1 and fusion together and does not separate the drug classes

• ← related toAn H3K4me3-reading PHD finger fused to NUP98 makes a potent leukaemia oncoprotein; the H3K4me3-reading PHD finger of KMT2A is deleted from every KMT2A fusion and reinstating it abolishes transformation. Same module class, opposite sign - and the one NUP98 fusion that carries KMT2A's PHD cassette is the weakest member of its family.

  the PHD cassette deleted from every KMT2A translocation product

• ← related toEWS::FLI1 builds its own enhancers at GGAA microsatellites, and a single inherited base change that lengthens one repeat raises the fusion's binding and the patient's risk. KMT2A fusions occupy pre-existing, invariant CpG islands - which is why this disease has no germline modifier of its target set, and why its mouse models work at all.

  the CXXC domain whose non-methylated-CpG affinity sets the shared address

• ← testsThe specificity control that should have clinched this beat inverts instead: the C-terminal MLL1 epitope the fusion does NOT contain loses MORE than the N-terminal epitope it does contain (68.8% vs 78.4% retained), and because the deposit records different cell-line names for these four samples than for every other one, the arm is reported as uninterpretable rather than interpreted

  An attempted wild-type-MLL1 specificity control using a C-terminal epitope the fusion does not contain. It inverted - the wild-type-only arm lost MORE (68.8%) than the fusion-containing arm (78.4%) - and is reported as uninterpretable, not interpreted.

• ← related toGSE159608 — 419 automated CUT&Tag libraries across 7 KMT2A-rearranged primary patient samples, 5 lines and a normal CD34+ donor, including KMT2A fusion occupancy, a menin-inhibitor arm in SEM, and three single-cell CUT&Tag libraries

  180 CUT&Tag libraries mapping KMT2A fusion occupancy across five partner genotypes

• ← related toKMT2A and several of its fusion partners occupy the same spatial cluster in individual nuclei before any break occurs — single-cell diploid Hi-C, RUNX1-dependent, and 3D architecture does NOT predict which fusion type a cell will make

  KMT2A sits in a multipartite spatial cluster with several partner loci in individual nuclei

• ← related toIn an unsupervised 5 kb correlation of 41 SEM libraries, both KMT2A N-terminal AutoCUT&RUN libraries have menin as their single nearest neighbour - the menin-interface hypothesis appearing as the top-level structure of an analysis that was not looking for it

  The N-terminal fragment carries the menin-binding motifs and is the part retained in the fusion

• ← related toThe menin gap survives matching on menin's own signal, but only two deciles of the two strata overlap

  wild-type MLL1 (C-terminal antibody) shows no preferential loss at fusion-occupied promoters, matched gap -0.055

• ← related toH3K4me3 and wild-type MLL1 are the negative controls for the partition, and both come out flat

  wild-type MLL1 C-terminal channel does not redistribute after fusion degradation

• ← related toThe taspase-1 N/C antibody control separates fusion-occupied from wild-type KMT2A promoters in every primary infant sample, and it is the C-terminal arm that does the work while being the less reproducible arm

  taspase-1 N320/C180 split used as a built-in wild-type control

• ← related toThe KMT2A N-terminal antibody reports wild-type plus fusion and the C-terminal antibody reports wild-type alone - which makes an N-versus-C difference a fusion-specific readout with no tag and no engineering

  the two taspase-1 cleavage products underpin the readout

• ← related toGSE252378 fragment BEDs: a complete KMT2A N-terminal / C-terminal antibody set in 27 patient samples across six fusion partners, deposited and until now unopened by this corpus

  N-terminal and C-terminal KMT2A antibodies, two of each, in every sample

• ← related toThe CD34+ oncoprotein-negative sample that calibrates the published N/C threshold is in GSE159608, not GSE252378 - so the 1.75 threshold cannot be re-derived from the deposit that carries the 27 patients it is applied to

  the threshold calibrates the KMT2A N-terminal over C-terminal ratio

• ← related toRecalibrate the N/C oncoprotein threshold on the fusion-negative CD34+ control at the interval scale it will actually be used at, and report the false-positive count at every scale rather than one

  N-terminal versus C-terminal KMT2A antibody ratio

• ← related toThe N-over-C oncoprotein score is not fusion-specific by construction, because taspase1 cleaves wild-type KMT2A into an N320 and a C180 fragment that need not occupy the same chromatin - so a fusion-negative cell has a real N/C signal

  the two taspase1 cleavage fragments N320 and C180

• ← related toThe unbounded oncoprotein score spans 1.73-12.32 log2 units within a sample - it is not the compressed statistic q142 was stuck with - and the pre-registered magnitude verdict is STILL void, because the score's amplitude is no larger than the disagreement between two antibodies against the same end of KMT2A

  N-terminal over C-terminal KMT2A signal

• ← related toScored in the published criterion's own currency, the number of oncoprotein-target promoters runs 0 to 2282 across 27 KMT2A-rearranged patient samples - 20.01x between the 10th and 90th percentiles - and the spread widens to 61.10x after depth matching

  KMT2A N-terminal over C-terminal call count

• ← related toIn 18 of 27 patient samples, comparing two antibodies raised against the SAME end of KMT2A calls at least as many oncoprotein-target promoters as the real N-over-C comparison does, and swapping the antibody pair moves a sample's own call count by a median of 2.57x

  four KMT2A antibodies, two per terminus

• ← related toThe fusion-specific score's per-promoter reliability inside one patient is r = 0.183, and the median correlation between two DIFFERENT patients is 0.196 - so the ordering question cannot be resolved at promoter resolution, and the corpus's 0.82-0.97 figures cannot be carried to this readout

  per-promoter N-over-C ordering

• ← related toGSE252378 holds eight KMT2A::AFF1 oncoprotein panels from two centres, not the three this corpus had used

  KMT2A N- and C-terminal antibody panels across seven patients

• ← related toThe GMP-like promoter class carries no N-over-C excess in 25 of GSE252378's 27 samples: whatever occupies it is seen equally by the fusion-side and the wild-type-side antibody

  N/C near parity is consistent with wild-type KMT2A rather than the fusion holding the class

• ← related toGSE68720 - 97 infant ALL patients from COG P9407 with the KMT2A partner named per sample (mll_par) and a GPL570 expression matrix, 74 partner-named KMT2A-r cases and 17 germline controls

  per-sample KMT2A partner in the mll_par field, 74 partner-named cases

• ← related toaml_stjude_2024 - 887 paediatric AML cases with per-sample KMT2A fusion-partner structural-variant calls for all 180 KMT2A-rearranged cases, plus mRNA-seq and FAB morphology

  237 KMT2A structural-variant records over 180 samples

• ← related toEPHA7 - ephrin type-A receptor 7, a published direct promoter target of the ALL1/AF4 (KMT2A::AFF1) fusion protein

  target of KMT2A fusion proteins in the 2007 founding experiment

• ← related toThe 2.03-fold over-representation of shared windows among fusion-occupied promoters is generic promoter activity, not a fusion preference - and if it is not, the rule is finding something different from what it claims

  H-B would require the fusion's CXXC unmethylated-CpG address to prefer bidirectional CpG-island promoters beyond their activity; prior 0.2

• ← related toRun on the reciprocal KMT2A-C antibody instead of the fusion-side KMT2A-N, the same rule gives zero AFF1-exclusive recurrent gene bodies against a null of 0.52 - the gene-scale class is fusion-arm specific

  the C-terminal fragment's gene-body occupancy is more uniform across patients than a random relabelling predicts

• ← related toThe 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

  315 of the 879 IntAct REST records are KMT2A's; 91 prey-only partners

• ← bindsKMT2A-AKAP8 (AKAP95): graded 'prey-only' by this project, and it is a direct positive modulator of MLL H3K4 methyltransferase activity from a dedicated MLL paper - ADMITTED

  AKAP95 co-elutes with the MLL1 complex on Superose 6 (molecular sieving, PMID 23995757)

• ← bindsKMT2A-CLOCK: one paper, six IntAct records, and the bait/prey grade splits it down the middle - CLOCK falls in the discard bin and SIRT1 does not, from the same experiments

  single two-hybrid, CLOCK bait / KMT2A prey, miscore 0.37

• ← bindsKMT2A-HMGN5: recovered by endogenous CRISPR tagging in OpenCell, not by over-expression - a chromatin-decompacting nucleosome binder with a real reason to be near KMT2A

  KMT2A prey in an endogenously tagged HMGN5 pull-down, OpenCell, HEK293T

• ← bindsKMT2A-KLHL22 - DECLINED. The single source is a cryo-EM structure of KLHL22-CUL3 bound to glutamate dehydrogenase, and KMT2A is a HEK293F pull-down passenger with no nuclear rationale

  single HEK293F preparative pull-down from a cryo-EM paper; DECLINED

• ← bindsKMT2A-DRC5 - DECLINED. A ciliary dynein-regulatory-complex protein, recovered by BioID in a centrosome-cilium map, and the only one of 135 uncovered partners never mentioned anywhere in this corpus

  two identical BioID proximity rows from a centrosome-cilium map, miscore 0.27; DECLINED

• ← part ofThe K1130/K1133 numbering is HUMAN, stated in the paper's own words, and the immunising peptide proves it a second time without relying on that sentence

  human Q03164 residues 1130 and 1133

• ← regulatesCREBBP (CBP) — binds the KMT2A transactivation domain, which no fusion keeps

  CBP acetylates MLL1 (pan-acetyl readout, C-terminally biased); not shown at K1130/K1133 specifically

• ← regulatesEP300 — the acetyltransferase whose paralogue split is the finding: p300 is the 108th most KMT2A-selective gene, CBP is the 16,950th

  p300 acetylates MLL1 (pan-acetyl readout, C-terminally biased); not shown at K1130/K1133 specifically

• ← part ofThe 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

  174 MS-evidenced PTM site calls on Q03164

• ← part ofEvery MS-evidenced KMT2A modification site, flagged retained or lost against the breakpoint - and the retained portion is NOT enriched for acetylation once you use the right null

  71 retained, 4 variable, 99 lost PTM sites on Q03164

• ← part ofKMT2A K1130 acetylation - the clock-controlled, SIRT1-erased site that every KMT2A fusion retains, 17 residues from the CXXC and 208 from the breakpoint

  Q03164 K1130, exon 5, retained in every classical KMT2A fusion

• ← part ofKMT2A K1133 acetylation - the better-evidenced half of the pair, and the one that appears in the CBP/p300 acetylome itself

  Q03164 K1133, exon 5, retained in every classical KMT2A fusion

• ← part ofKMT2A K1130 ubiquitination - the same retained lysine as the clock acetylation site, carrying a chemically exclusive second mark that no KMT2A paper has ever mentioned

  Q03164 K1130 ubiquitination, retained

• ← part ofKMT2A K636 acetylation - the forgotten member of the Choudhary quartet, retained, and outside the SIRT1-interaction region

  Q03164 K636, exon 3, retained with a 702-residue margin

• ← part ofKMT2A K1235 acetylation - inside the SIRT1-interaction region, retained, and the one member of the Choudhary quartet the source paper named and then never assayed

  Q03164 K1235, exon 7, retained with 99 residues of margin

• ← part ofKMT2A K2528 SUMO2 - the single best-supported modification site on the whole protein, in the N320 fragment, and LOST in every fusion

  Q03164 K2528 SUMO2, in N320, LOST in every fusion

• ← part ofKMT2A K1364 SUMO - the one modification site inside the breakpoint cluster, retained in some patients' fusion proteins and lost in others

  Q03164 K1364 SUMO, inside the breakpoint cluster, patient-variable

• ← bindsAKAP8 (AKAP95) - the only nuclear A-kinase anchoring protein, and the one that turned out to be required for MLL-rearranged leukaemia

  AKAP95 IDR1 (101-390) co-condenses with the MLL1 translocated fragment - the region every KMT2A fusion retains

• ← bindsAKAP95 IDR1 co-condenses with the KMT2A N-terminal fragment - the one AKAP95 contact that lands inside the region every KMT2A fusion retains

  co-condensation with the MLL1 translocated fragment, charge-complementarity driven

• ← related toZNF384 fusions make lineage-ambiguous leukaemia by acting in a haematopoietic stem cell - and the founding lesion is in the HSC itself, unlike Ph-like B-ALL

  KMT2A appears in the ZNF384-fusion RIME interactome of 74 proteins - an UNVALIDATED mass-spectrometry hit, flagged as a lead not a contact

• ← related toKMT2A::AFF1–LEF1 is one AP-MS row (p 0.0261, 10|8 spectra) in a Jurkat T-ALL host, where LEF1 is a resident lineage factor; it is the only one of the six AF4-specific candidates with non-zero negative-control signal, and no Wnt co-factor was co-recovered — declined

  declined KMT2A::AFF1-LEF1 pair: one prey row, one method, one publication, no interface

• ← related toACADM is a mitochondrial matrix beta-oxidation enzyme with an experimentally coded matrix-only localisation, recovered on 8 spectra from a nuclear-extract pulldown whose specific tail is 14.5% mitochondrial — declined

  declined KMT2A::AFF1-ACADM pair: one prey row, a mitochondrial matrix enzyme, from a nuclear extract

• ← related toGLUD1 is a hexameric mitochondrial matrix dehydrogenase recovered on 11 spectra with non-zero control signal from a nuclear pulldown — and this corpus has already declined one KMT2A partner whose sole source was a cryo-EM structure of GLUD1 — declined

  declined KMT2A::AFF1-GLUD1 pair: 11 spectra, a matrix hexamer, non-zero negative control

• ← related toRGPD8 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

  declined KMT2A::AFF1-RGPD8 pair: the identification cannot be resolved to the gene

• ← related toRPA2 is the only compartment-compatible candidate of the six and the corpus already has a KMT2A resection thread it would fit — but it is an abundant ssDNA-binding subunit on 6 spectra in a benzonase-treated lysate with no RPA1 or RPA3 co-recovered — declined, thread flagged

  declined KMT2A::AFF1-RPA2 pair: 6 spectra and neither obligate partner subunit co-recovered

• ← related toCEP112 is a 955-residue centrosomal coiled-coil protein on 7 spectra with zero leukaemia literature — its only nuclear annotation is a single 2024 citation, and a long coiled-coil is the classic non-specific AP-MS passenger — declined

  declined KMT2A::AFF1-CEP112 pair: weakest p of the six, no literature on any side

• ← related toThe 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

  RPA ChIP enrichment at the KMT2A 11q23 breakpoint cluster region after etoposide; occupancy at the locus, NOT a contact with the protein or the fusion

• ← related toIf RPA loading at the KMT2A breakpoint cluster region is the licensing step for the etoposide-induced translocation, then the nuclear RPA reserve is a modifier of therapy-related KMT2A-rearranged leukaemia risk - and an ATR inhibitor, which kills by exhausting RPA, would raise that risk rather than lower it

  the locus whose breakpoint cluster region loads RPA after etoposide

• ← bindsThe menin acidic shelf, Glu366-Asp370-Val371: the counter-charge that MBM2 binds, resolved in 3U88 after all - and the one patch where the drug pocket, the SvID and the gamma-H2A.X motif all touch

  menin E366/D370 meet KMT2A Arg24; menin 371-378 pack against KMT2A Leu27

• ← related toRUN, 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

  the locus interrogated: gene span, BCR core, exon-11 CTCF element, DCAL cut point and TSS

• ← related toCBX1 / HP1-beta (P83916, 185 aa) - the HP1 paralogue with the fitness cost and no role in this graph's cascade: DepMap mean -0.349 with 16.6% of 1,257 lines below -0.5, four times deeper than HP1-gamma or HP1-alpha, and zero papers place it anywhere near KMT2A

  measured absent at the locus in ENCODE K562 and HepG2 by this beat

• ← related toThe 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

  the exon-11 CTCF element at chr11:118,488,614-118,488,760 inside the breakpoint cluster region

• ← related toA genome-wide DAMAGE-STATE occupancy dataset for CtIP and MRE11 does exist and is public - GSE218590, 58 samples, +/- 4-OHT and a light-activated Cas9 time course - and it cannot answer the KMT2A question, because there is not one AsiSI site in the 90,377 bp of the gene

  zero AsiSI sites (GCGATCGC) in the 90,377 bp gene body; nearest is 84,343 bp past the 3' end

• ← related toHypothesis: SET::NUP214 activates HOXA partly through recruited wild-type MLL1, which would make a KMT2A-unrearranged leukaemia menin-inhibitor-sensitive

  SET/TAF-Ibeta's acidic region binds the AMINO-TERMINAL region of MLL - the same end that carries MBM1 and MBM2, so the fusion may compete with menin rather than deliver a menin-loaded MLL1

• depends on →The N320-C180 clamp is the only KMT2A module whose disruption has ever been shown to change subnuclear localisation - and the fusion is the disruption

  the KMT2A subnuclear address is a property of the N320-C180 clamp, which no fusion can form

• related to →The MLL-side twin of the 1998 anti-AF4 result: in 1997 the same punctate nuclear pattern was seen with and without an 11q23 translocation, and the authors said so

  the punctate nuclear image of KMT2A in translocated cells is unattributable by its own authors

• related to →Nuclear matrix and mitotic chromosome scaffold — KMT2A sits with topoisomerase II, in the same place the gene breaks

  MLL amino terminus targets nucleolus, interphase nuclear matrix and mitotic chromosomal scaffolds

• depends on →Nuclear import of every KMT2A fusion is a property of retained KMT2A sequence, and the determinant was mapped in 1997 to a region containing the AT-hook motifs

  nuclear import maps to the AT-hook region; UniProt annotates no NLS motif on KMT2A

• related to →The nucleolus — the one compartment with direct published images of a KMT2A fusion inside it

  nucleolar localisation attributed to the KMT2A amino terminus under conditional overexpression

• depends on →Taspase1 and its two KMT2A cleavage sites: an enzyme-substrate interface with four apo structures, one co-crystal with a small molecule, and no structure with MLL at all

  the two-fragment state that carries the subnuclear address is produced by taspase1 cleavage

• ← related toUTP18 is not a KMT2A partner - it is WDR5, mis-curated into IntAct from the founding paper, and it manufactured the only KMT2A-bait gap in the interactome recount

  the one unwritten bait-supported KMT2A partner in the 2026-09-21 recount is a curation error, so KMT2A's directed-interactome gap is zero

• ← part ofWDR5 — the partner that is not a partner: it binds the half of KMT2A the fusion throws away

  WDR5 is one of the eight polypeptides resolved in the founding MLL complex purification (PMID 15199122)

• ← related toZero of eight unwritten AFF4/MEN1 partners has an experimentally coded nuclear localisation, and the two that are called nuclear are called it by inference

  positive control: KMT2A returns ECO:0000269 Nucleus through the identical code path, plus its two uncoded cleavage-fragment location comments

• ← related toTwo thirds of the curated interactome of this project's six core proteins is spoke-expanded - and the pairs this graph chose to write are significantly better than the ones it did not

  71.9% of KMT2A's 146 IntAct partner pairs rest only on spoke-expanded rows

• ← related toEvery IntAct number in this project is drawn through a query that misses half the record - and what it misses is the taspase1 fragments

  the records the query cannot reach are filed against the taspase1 fragments Q03164-PRO_0000390949 and -PRO_0000390950

• ← bindsSKP2 - 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

  SCF(Skp2) grips the retained N-terminal fragment; 2 IntAct records, neither spoke-expanded

• ← bindsTOPBP1 - a real KMT2A contact that no abstract-level query can see, because it is a control panel in a paper about a different protein

  yeast two-hybrid, KMT2A ePHD vs TopBP1 BRCT region; one record, Figure 7D control panel

• ← bindsAR / androgen receptor - the best-supported partner in the absence set by independent-publication count, and the reason a menin inhibitor is in prostate cancer trials

  endogenous reciprocal co-IP in VCaP plus BioID and co-IP in LAPC4/HEK293T; 2 independent publications

• ← part ofThe human CXXC/CpG-reading family is exactly twelve proteins, and TET2 is not one of them

  KMT2A is one of the twelve, and the reason the census was run.

• ← fuses withTET1 (LCX / CXXC6) - a CXXC reader that is also a KMT2A fusion partner, and whose leukaemic fusion throws its own CXXC domain away

  t(10;11)(q22;q23) MLL-LCX in AML-M2 with trilineage dysplasia; the fusion protein lacks TET1's own CXXC domain and retains its coiled coil.

• ← related toDNMT1 - the antagonist: its CXXC domain reads unmethylated CpG in order to STOP methylation, and it is the only heterologous CXXC that runs an MLL-AF9 leukaemia

  Decitabine-induced DNMT1 degradation coincides with loss of KMT2A protein in ALL lines, attributed by the authors to structural similarity between the two DNA-binding domains, and is synergistic with menin inhibition.

• ← inhibitsDecitabine - degrades DNMT1, and in ALL cells it degrades KMT2A too: the hypomethylating agent with two mechanisms and a menin-inhibitor combination already in the clinic

  KMT2A protein fell in every decitabine-responsive ALL line that degraded DNMT1, and only those lines showed biological effects; mechanism proposed on DNA-binding-domain similarity and not directly demonstrated.

• ← inhibitsDecitabine degrades DNMT1 - and in ALL cells the same treatment lowers KMT2A protein, with the shared CXXC-family DNA-binding module offered as the reason

  KMT2A protein fell in every decitabine-responsive ALL line that degraded DNMT1; species (full-length, N320, C180 or fusion) not distinguished.

• ← related toThe CXXC-DNA affinity is a ten-point ladder from 44 uM to 33 nM, not two numbers 1,300-fold apart, and every step but the last is under 5.5-fold

  All nine measurements are of Q03164 fragments spanning 1136-1244.

• ← part ofThe MT/BCB domain, Q03164 1147-1244: the CXXC fold plus a 41-residue basic tail nobody has measured on its own

  Q03164 1147-1244, the MT/BCB domain, derived from Birke's primers against a live UniProt fetch.

• ← bindsThe MLL-protected Hoxa9 CpG sites bind the CXXC domain 4- to 9-fold tighter than the unprotected ones, in one ITC series at fixed everything else

  KMT2A CXXC 1147-1203 binds Hoxa9-derived CpG duplexes at 3.5 to 32 uM by ITC.

• ← related toOne domain, two names - and `MLL[tiab]` does not match `MLL1`. Two token traps, two missing affinity papers each

  The KMT2A region 1147-1244 carries two names with disjoint literatures.

• ← testsRerun the 2002 SPR with the two controls it did not have: cleave the GST, and put the protein on the chip instead of the DNA

  Re-measures the KMT2A MT domain-DNA contact with the controls the 2002 experiment lacked.

• ← related toTen affinity measurements for this contact and not one uses a fusion protein, full-length KMT2A or a nucleosome

  The largest KMT2A species ever titrated against DNA is 98 of 3,969 residues.

• ← inhibitsTwo entries record zero chemical matter against the CXXC-DNA interface. There is one: dimethyl fumarate, IC50 about 1 uM, and it is already an approved drug

  Dimethyl fumarate inhibits the MLL1 CXXC-DNA complex at a relative IC50 of 0.91 uM in a single-molecule assay; isolated domain, naked DNA, no cell.

• ← related toEviction is a claim about menin. The field, and this graph, transfer it to the fusion without a measurement that supports the transfer

  the shared N-terminal epitope is why MLL1 and the fusion cannot be separated in these assays

• ← bindsSETD2 (Q9BYW2) - the H3K36 trimethyltransferase that binds all seven MLL fusions, and the only one of them with a published KMT2A::AFF1 dependency test

  co-IP maps the contact to the MLL N-terminus and the SETD2 C-terminus - in transiently transfected HEK293

• ← bindsJADE3 / PHF16 (Q92613) - the HBO1-complex scaffold with the strongest pan-fusion evidence in this beat, and an independent published contact with MLL

  independent evidence: MLL proteins contact the HBO1 complex via THD2, mediated mainly by the ING4/5 and PHF16 (=JADE3) subunits; transformation arm shown for MLL-ELL

• ← related toNUDT21 / CFIm25 (O43809) - the CFIm subunit that picks poly(A) sites, in the pan-MLL-fusion core, and a paraspeckle protein

  pan-MLL-fusion core interactor; the bait is the shared KMT2A N-terminus

• ← related toWDR33 (Q9C0J8) - the subunit that reads AAUAAA, in the pan-MLL-fusion core, and the plant homolog of which physically links 3' processing to H3K4 demethylation

  the plant WDR33 homolog FY sits in a physical chain with an H3K4 demethylase module

• ← inhibitsSON (P18583) - a nuclear speckle scaffold that BINDS MENIN and inhibits MLL complex assembly, and the one protein in this tranche the Skucha authors name in their own discussion

  inhibits MLL complex assembly, lowering H3K4me3 - wild-type MLL, not tested on a fusion

• ← related toZC3H18 (Q86VM9) - the CBC-to-exosome bridge in the pan-MLL-fusion core, and the ONLY member of this tranche whose own complex came with it

  a phosphorylation-dependent ZC3H18 isoform associates with core histones, which is why a decay adaptor is in a chromatin purification

• ← related toMAP7D1 (Q3KQU3) - the strongest index-fusion p-value of the nine deferred pan-MLL-fusion preys, on a protein UniProt places entirely outside the nucleus

  pan-MLL-fusion core prey; the shared bait is the KMT2A N-terminus

• ← fuses withVIM / vimentin (P08670) - a pan-MLL-fusion prey with a clean negative control, and separately a genuine KMT2A rearrangement partner in sarcoma, which are two different facts

  VIM::KMT2A recurrent sarcoma fusion, VIM exon 4 :: KMT2A exon 2, 4 published cases - KMT2A is the 3' partner and nearly all of KMT2A is retained, so this is NOT a leukaemia-type MLL fusion

• ← fuses withKMT2A::MAPRE1 (MLL-EB1) - a coiled-coil dimerisation fusion from pro-B ALL with a mouse model, low leukaemogenic potential, and Ets1 as its effector

  MLL exon 6 fused in frame to EB1 exon 5 in adult pro-B ALL; reciprocal EB1-MLL also verified

• ← related toEIF4G1 (Q04637) - a translation-initiation scaffold in the pan-MLL-fusion core, alone, with no other eIF4F subunit anywhere in the 960

  pan-MLL-fusion core prey; the shared bait is the KMT2A N-terminus

• ← related toZNF512 (Q96ME7) - the one protein of the nine deferred pan-MLL-fusion preys that is a sequence-specific chromatin factor, and it initiates pericentric heterochromatin

  pan-MLL-fusion core prey; the shared bait is the KMT2A N-terminus

• ← related toPRKAR1A (P10644) - a PKA regulatory subunit in the pan-MLL-fusion core with no catalytic subunit anywhere in the network, and a fusion oncogene in variant APL

  pan-MLL-fusion core prey; the shared bait is the KMT2A N-terminus

• ← related toLUZP1 / fimbacin (Q86V48) - the weakest index-fusion call of the nine, and the strongest non-index call: p = 0.127 against MLL-AF4 and 2.6e-10 against MLL-ENL

  pan-MLL-fusion core prey; the shared bait is the KMT2A N-terminus

• ← related toNUP214 / CAN - the cytoplasmic-face nucleoporin that CRM1 docks against, and the 3' partner of four leukaemia fusions

  A nucleoporin fusion relocalises wild-type MLL1: NUP98::HOXA9 and SET::NUP214 phase-separated nuclear bodies PROMOTE MLL1 CONDENSATION and are co-localized with MLL1/CRM1 ON CHROMATIN (PMID 37516964) - MLL1 as client, not driver (Q91 2026-09-22)

• ← part ofWhich KMT2A exon carries which interface: exon 1 is the menin-binding motif, exons 5-6 are the CXXC zinc finger (computed, not recalled)

  exon-to-residue map of KMT2A's interfaces (exon 1 = MBM, exons 5-6 = CXXC)

• ← part ofYAP1::KMT2A::YAP1 (YKY) - an insertional KMT2A rearrangement that keeps the CXXC zinc finger and throws the menin-binding exon away

  contributes only KMT2A exons 4/5-6 (the CXXC zinc finger) as an internal insert

• ← part ofVIM::KMT2A - the one KMT2A rearrangement that keeps the SET domain and still throws away the menin-binding exon

  KMT2A exon 2 onward = aa 145-3969, SET domain retained, exon 1 discarded

• ← fuses withYAP1 (P46937) - the Hippo pathway's transcriptional co-activator, and a KMT2A fusion partner from a completely different functional class

  predominant KMT2A rearrangement partner in soft tissue sarcoma (16 of 34)

• ← depends onKMT2A-rearranged soft tissue sarcoma - the same gene as infant B-ALL, in a young adult's leg, as a solid tumour

  defined by rearrangement of KMT2A, with KMT2A as the 3-prime or internal contributor

• ← part ofLDB1::KMT2A - the clearest statement in this literature that a sarcoma fusion keeps KMT2A exons 4 to 36 and discards exons 1 to 3

  KMT2A exons 4-36 = aa 1053-3969; exons 1-3 discarded

• ← fuses withLDB1 (Q86U70) - and the sarcoma fusion keeps the part that self-associates while losing the LIM interaction domain entirely

  LDB1 exons 1-10 to KMT2A exons 4-36, one published sarcoma case

• ← fuses withDOCK10 (Zizimin-3) - the Cdc42/Rac1 exchange factor fused to KMT2A exon 8 in an undifferentiated pleomorphic sarcoma, and the only one of the seven this graph already held as an H3K79me2-marked gene

  KMT2A exon 8 :: DOCK10 exon 2 (case S13); KMT2A 5-prime, aa 1-1362, MBM+CXXC retained, break inside the leukaemic BCR

• ← fuses withAHDC1 (Gibbin) - an AT-hook chromatin-architecture factor that is already a cofactor of a DIFFERENT sarcoma fusion oncoprotein, and whose KMT2A breakpoint lands in an untranslated exon

  KMT2A exon 15 :: AHDC1 exon 2 (case S16); KMT2A 5-prime, aa 1-1668, MBM+CXXC retained - but AHDC1 exon 2 is 5'-untranslated, so no partner protein

• ← fuses withELP4 (PAXNEB) - a required subunit of Elongator, the one partner of the seven with a genuine histone-acetyltransferase connection, and the one whose fusion supplies only its last 43 residues

  KMT2A exon 9 :: ELP4 exon 10 (case S15, angiosarcoma); KMT2A 5-prime aa 1-1406, break inside the leukaemic BCR; ELP4 contributes only its terminal 43 residues

• ← fuses withCPT1A - the outer-mitochondrial-membrane fatty-acid gatekeeper that is also a lysine succinyltransferase, fused to KMT2A exon 3 in the one case that is tether WITHOUT address

  KMT2A exon 3 :: CPT1A exon 15 (case S12, MPNST); KMT2A 5-prime aa 1-1052 - MBM and AT-hooks retained, CXXC LOST

• ← fuses withC2CD3 - the centriole-elongation regulator fused to KMT2A exon 8 in an MPNST, predicted OUT OF FRAME, and the partner whose RefSeq transcript is 390 residues shorter than UniProt's canonical

  KMT2A exon 8 :: C2CD3 exon 24 (case S14, MPNST); KMT2A 5-prime aa 1-1362, break inside the leukaemic BCR; junction predicted OUT OF FRAME

• ← fuses withTMPRSS13 (MSPL) - a cell-surface serine protease fused in frame to KMT2A exon 20, the clearest NOT-APPLICABLE epigenetics call of the seven, and a symbol whose approved form has a third of its alias's recall

  KMT2A exon 20 :: TMPRSS13 exon 2 (case S17, high-grade endometrial stromal sarcoma); KMT2A 5-prime aa 1-1888, predicted in frame but topologically incoherent

• ← fuses withRALGAPA2 (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

  KMT2A exon 22 :: RALGAPA2 exon 2 (case S18, conventional osteosarcoma); KMT2A 5-prime aa 1-1987, the largest KMT2A contribution of the seven; junction predicted OUT OF FRAME

• ← recruitsHCFC1 (HCF-1) — a founding member of the wild-type KMT2A complex that the fusion cannot bind

  HCF-1 recruits the MLL-family H3K4 methyltransferases to E2F-responsive S-phase promoters; the motif it binds (KMT2A aa 1799-1802) is deleted from every fusion

• ← related toELP1 (IKBKAP/IKAP) - the Elongator scaffold that was first named for the IKK complex, is a paediatric-cancer predisposition gene, and has an RNA Pol II contact in IntAct

  BOUNDED ABSENCE, not an interaction: ELP1 x KMT2A = 0 records in titles/abstracts, 0 IntAct records among 130, 0 GEO leukaemia datasets, and absent from all five KMT2A-fusion interactome analyses in this repo

• ← related toELP2 (STATIP1/StIP1) - the double beta-propeller of the Elongator core, whose informal name collides with a different human gene and whose STAT3 function is a by-similarity annotation

  BOUNDED ABSENCE: 0 literature records, 68 IntAct records with no KMT2A-axis partner AND no POLR2A, absent from all five fusion-interactome analyses in this repo

• ← related toELP3 - the catalytic subunit, the only Elongator protein with a real acetyltransferase domain, and the one whose loss in mouse blood causes marrow failure and then p53-mutant leukaemia

  BOUNDED ABSENCE, bounded four ways and nearly fabricated by an instrument failure: 0 literature records (ELP3[tiab] first returned 0 with an EMPTY querytranslation; true count 202), 0 of 107 IntAct records, 0 GEO leukaemia datasets, absent from all five fusion-interactome analyses here

• ← related toELP5 (DERP6) - the bridge that joins ELP3 to ELP4, the most essential Elongator subunit in DepMap, and a gene whose cytoplasmic call rests on an overexpression experiment

  BOUNDED ABSENCE, and small-literature: 0 records against controls of 104/30/218, but ELP5 total literature is 22 records and its IntAct bait has only 13 curated partners

• ← related toELP6 (TMEM103/C3orf75) - an Elongator subunit named after a membrane protein it is not, whose entire localisation record comes from a Drosophila paper its own symbol cannot retrieve

  BOUNDED ABSENCE, small-literature, and the second protein in this beat on which the empty-querytranslation instrument failure fired (ELP6[tiab] returned 0 before returning 43)

• ← related toSMYD3 - 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

  IDENTIFIER TRAP ON THIS PROJECT OWN INDEX GENE: SMYD3[tiab] AND KMT2A[tiab] = 2 and BOTH are off topic, while AND MLL[tiab] = 5 and AND HOXA9[tiab] = 1 carry the mechanism. Treat X AND KMT2A = 0 as uninformative until X AND MLL has been run

• ← related toILF3 / NF90 - a nucleolar dsRNA-binding coactivator whose chromatin residency was mapped in the same leukaemia line that proximity-labelled KMT2A

  ILF3-bait BioID2 in K562 recovered KMT2A as prey (MI:1314 proximity, MI-score 0.27) - a proximity row, not a contact

• ← related toBUD13 - a RES-complex splicing factor 1.66 Mb from KMT2A on 11q23.3, whose one chromatin paper puts it in the soluble fraction

  BUD13-bait BioID2 in K562 recovered KMT2A as prey (MI:1314, MI-score 0.27); one n-ary record EBI-56895223

• ← part ofAFF1::KMT2A (AF4-MLL), the der(4) product: what it is made of, residue by residue

  Carries KMT2A residues from the breakpoint cluster region to 3969, including PHD1-4, bromodomain, FYRN, TAD, FYRC, WIN and SET.

• ← regulatesTaspase1 is what turns the reciprocal product into an oncoprotein - and it is a conditional oncogene because of it

  The same protease cleaves wild-type KMT2A at the same two sites, which is both the toxicity risk and the reason the released MLL-C fragments are indistinguishable.

• ← part ofFYRN/FYRC: the clip that holds the cut reciprocal protein together, and the only fusion-adjacent interface anyone has tried to break with a decoy

  FYRN 2018-2074 and FYRC 3666-3747 are wild-type KMT2A domains, which is why a decoy cannot be fusion-selective.

• ← related toAt the DNA level almost every KMT2A-rearranged patient has a reciprocal allele - 94% across partners, and 5% have nothing else

  Counts reciprocal alleles across the whole KMT2A recombinome, not only t(4;11).

• ← part oft(6;11) makes four fusion genes, not two, and the PHD/bromodomain cassette is what opens chromatin in all of them

  Nominates the KMT2A PHD/bromodomain cassette as the causal element in reciprocal products generally.

• ← part ofKMT2A*: a menin-blind, LEDGF-blind, SET-competent KMT2A protein made from out-of-frame reciprocal alleles

  A 5'-truncated KMT2A protein initiating from a gene-internal promoter upstream of exon 12.

• ← bindsSET / TAF-Ibeta domain map (Q01105, 290 aa) - dimerisation, earmuff, and the acidic tail that binds MLL

  the SET acidic region binds the MLL N-terminus in both the wild-type and the SET-NUP214 direction

• ← bindsSET::NUP214 is KMT2A-free in its sequence but not in its mechanism: it binds and condenses wild-type MLL1, which reopens the menin question this beat was sent to close

  the recruited protein is wild-type MLL1/KMT2A, which carries MBM1 and MBM2

• ← related toThe MLL-side twin of the 1998 anti-AF4 result: in 1997 the same punctate nuclear pattern was seen with and without an 11q23 translocation, and the authors said so

  wild-type HRX/KMT2A is punctate and widely expressed, which is what makes the fusion invisible in an image

• ← part ofNuclear import of every KMT2A fusion is a property of retained KMT2A sequence, and the determinant was mapped in 1997 to a region containing the AT-hook motifs

  the import determinant is retained KMT2A sequence, shared with the wild-type protein

• ← part ofThe clustered PHD fingers (KMT2A aa 1416-1640) are a chromatin-targeting module and the canonical t(4;11) junction removes them by ten residues

  clustered PHD fingers of KMT2A, aa 1416-1640

• ← part ofThe N320-C180 clamp is the only KMT2A module whose disruption has ever been shown to change subnuclear localisation - and the fusion is the disruption

  the FYRN-FYRC clamp between the taspase1-generated halves of KMT2A

• ← related toHypothesis: what the KMT2A half loses is control of its own abundance, not control of its activity

  [describes] three degradation determinants C-terminal to the breakpoint are absent from every der(11) fusion

• ← regulatesTASP1 (taspase1) - the protease that cuts KMT2A, and the timer the fusion escapes

  [cleaves] D2666 and D2718; cleavage primes MLL1 for proteasomal degradation (contested: Takeda reads it as activation)

• ← regulatesSilmitasertib (CX-4945) - a CKII inhibitor that works on KMT2A-rearranged leukemia by strengthening the allele the fusion does not have

  [stabilises] raises full-length wild-type KMT2A protein and its chromatin occupancy

• ← testsExperiment: titrate wild-type KMT2A up and down in one leukemia cell and watch what happens to the fusion's grip on chromatin

  [perturbs] degron-titrated wild-type allele, plus non-cleavable knock-in for the upward arm

• ← part ofLost from the KMT2A half: the four-residue HCF-binding motif (DHNY, aa 1799-1802)

  HCF-binding motif DHNY at aa 1799-1802, lost in every fusion

• ← part ofLost from the KMT2A half: PHD3 and the bromodomain are one cassette, not two modules (aa 1566-1765)

  aa 1566-1765; also flags a ~65-residue error in the stored bromodomain range

• ← part ofLost from the KMT2A half: the WIN motif and the whole WRAD interface (aa 3762-3767, Arg3765)

  WIN motif aa 3762-3767, Arg3765; ~2,400 residues past the breakpoint cluster

• ← part ofLost from the KMT2A half: the SET domain (aa 3829-3945) - the lost module whose loss demonstrably does not matter

  SET aa 3829-3945, post-SET 3953-3969; lost in every fusion

• ← part ofLost from the KMT2A half: the transactivation domain (aa 2829-2883) and its CBP contact

  TAD aa 2829-2883 (9aaTAD 2847-2855); lost in every fusion

• ← part ofLost from the KMT2A half: both taspase1 cleavage sites (D2666 and D2718) - the fusion is never processed

  taspase1 sites at 2666/2667 and 2718/2719; neither is present in any fusion

• ← part ofLost from the KMT2A half: the FYRN-FYRC clamp (2018-2074 / 3666-3747) that holds wild-type KMT2A together

  FYRN 2018-2074 and FYRC 3666-3747; both lost

• ← part ofLost from the KMT2A half: PHD2 (aa 1479-1533), an E3 ubiquitin ligase that limits KMT2A's own activity

  PHD2 aa 1479-1533, an E3 ligase that limits KMT2A abundance and occupancy; lost in every fusion

• ← part ofLost from the KMT2A half: the MOF/KAT8 contact - required in the cell, but not shown to matter to the fusion

  C-terminal MOF-recruitment contact, coordinates unresolved

• ← related toThe third category: modules lost from the fusion protein but supplied in trans by the wild-type KMT2A allele

  the untranslocated allele still encodes every module catalogued as lost

• ← part ofLedger: every module the KMT2A half loses, with coordinates, how each was confirmed, and whether its loss has been shown to matter

  complete coordinate ledger of the modules C-terminal to the breakpoint cluster region

• ← related toKMT2A partial tandem duplication: an addition that only works once the normal copy has been switched off — 'recessive gain-of-function'

  recessive gain-of-function challenged: Mll(PTD/WT) knock-in mice show the Hox and progenitor phenotype with the wild-type allele intact (Dorrance 2006)

• ← related toAuditing the seven strikes before applying them: nine of ten decisions stand, four for different reasons than were given, and one clinical argument fails

  two strikes rest on MLL1 SET-domain dispensability; Chen 2017 finds endogenous MLL1 dispensable altogether and MLL2 the major dependency

• ← related toEPS15 (AF1p): the EH domains may or may not be discarded, and either way nobody has shown it matters

  KMT2A::EPS15, t(1;11)(p32;q23); no fusion node exists for this pair

• ← related toHLA-DRA - the non-polymorphic alpha chain, and the only HLA-DR chain any source on this beat names explicitly

  common MLL/COMPASS subunits bind MHC-II and CIITA promoters and their overexpression enhances HLA-DRA expression - family-level, not KMT2A-specific

• ← related toThe named-but-unwritten queue's top two entries were a web API and a survival endpoint, and MYC was the real answer all along

  PZP IS NOT A GENE HERE: 21 entries use PZP for KMT2A's PHD-Zinc-finger-PHD cassette, while HGNC assigns it to pregnancy zone protein.

• ← bindsKDM4C (JMJD2C/GASC1) — recruited by the retained KMT2A N-terminus, drugged in a KMT2A::AFF1 line, and invisible to CRISPR

  interaction domain mapped by deletion to the KMT2A 5' region

• ← related toIn human KMT2A::AFF1 MV4;11 (monocytic, not B-lineage), the vehicle-versus-vehicle noise floor is 9.4% at the top 0.1% of menin/MLL1 sites and 0.1-2.6% in every other stratum - so cycle 102's borrowed 4-10% band is about right AT THE PEAKS and four to five times too generous everywhere else, and it was applied everywhere
• ← bindsWDR5: an established wild-type KMT2A contact, and no fusion contact ever measured

  WDR5 binds the KMT2A WIN motif at approximately aa 3762-3773 with Arg3765 in the WDR5 arginine pocket - a wild-type contact roughly 2,400 residues past the breakpoint cluster region.

• ← bindsRBBP5: the thinnest fusion literature of the four subunits, and one chromatin datum

  RBBP5 engages a non-active-site surface of the MLL1 SET domain as a heterodimer with ASH2L, and separately contacts nucleosomal H2B-ubiquitin.

• ← depends onWRAD reaches KMT2A fusions only as a passenger on a recruited wild-type complex

  The proposed route is the wild-type KMT2A allele, which KMT2A fusions require for stable chromatin association and for leukaemia maintenance.