FANCD2 — the FA effector menin binds, and the weakest FA gene in this graph's own selectivity scan

No three-dimensional model is available for this protein, so the map above is the whole picture. That is itself informative: AlphaFold does not model chains this long, and the parts that have been crystallised are held under the entries for the individual domains.

• FANCD2 is a tumour-suppressor effector. Anything that reduces its function in a child being treated for leukaemia raises the mutation rate in the normal marrow that has to repopulate afterwards.
• If a menin inhibitor turns out to impair FANCD2 function, the class acquires an unplanned genotoxic liability in exactly the population with the longest remaining lifespan.
• The converse bite: if menin restrains FANCD2, inhibiting menin could make leukaemic cells better at repair, which would antagonise the topoisomerase-II-directed and alkylating backbones these patients also receive.
• The FA pathway's documented human requirement is in haematopoietic stem and progenitor cells of people with no fusion (PMID 22683204). Any FANCD2- or FA-directed intervention in KMT2A-rearranged infant leukaemia is therefore predicted to hit the normal HSPC compartment through exactly the mechanism it is meant to exploit, in patients whose marrow has to last a lifetime.
• The subnuclear rationale is borrowed. Every image of FANCD2 in a focus comes from a genotoxin-treated non-haematopoietic cell line; nobody has looked in a KMT2A-rearranged cell. A localisation-based targeting argument here rests on a compartment that has not been shown to exist in the disease.

1. Is the menin-FANCD2 association independent of FANCD2 K561 monoubiquitination, as UniProt states but neither of its attached PMIDs shows? A FANCD2 K561R knock-in plus a menin co-IP would settle it in one experiment.
2. Why is FANCD2 the weakest FA gene in the KMT2A selectivity scan while being the one FA protein menin binds? If the bridge were real, FANCD2 should be at least as selective as the core subunits.
3. Does FANCD2 form damage-induced foci normally in KMT2A-rearranged blasts, and does a menin inhibitor change that?
4. Does the menin-FANCD2 association exist at all in a human haematopoietic cell, as opposed to mouse embryonic fibroblasts?
5. Does FANCD2 form damage-induced foci in a KMT2A-rearranged infant B-ALL cell, and at what baseline frequency without added genotoxin? No imaging or chromatin fractionation of FANCD2 has ever been done in this disease.
6. Is the rank-465 tilt an ID2-complex property or a FANCD2 property? FANCI ranks 52 and FANCD2 ranks 465 in the same scan, which is a large gap for an obligate heterodimer and is either a real asymmetry worth explaining or a reagent artefact in the library.

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

   FANCD2 is UniProt Q9BXW9, 1451 aa, primary gene name read back as FANCD2. It is monoubiquitinated on Lys-561 by FANCL with UBE2T and deubiquitinated by USP1; the FANCI-FANCD2 heterodimer binds and scans dsDNA; monoubiquitination is required for chromatin binding and for interaction with BRCA1, BRCA2 and FAN1. UniProt lists MEN1 among FANCD2's interactors, citing PMID 12874027.

   UniProtKB entry Q9BXW9 (FANCD2_HUMAN, Fanconi anemia group D2 protein) · UniProtKB/Swiss-Prot 2026
   ▸What this evidence cannot show

   A database record, not an experiment. The accession was resolved with gene_exact, which matches synonyms, so the primary gene name and the 1451 aa length were both read back off the entry rather than assumed. UniProt's PTM sentence bundles several facts under two PMIDs and the menin clause is not supported by either of those two abstracts; see E2 and E5.

   ▸Evidence that argues against this (1)

   • UniProt records an interaction as curated from the literature; it does not weigh the evidence. The MEN1 row rests on one 2003 co-immunoprecipitation paper that has never been independently replicated at the protein level. Marek LR et al. 2008
2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   ATM phosphorylates FANCD2 on Ser-222 and that phosphorylation, not the K561 monoubiquitination, is what is required to activate the S-phase checkpoint; the two modifications are independent and regulate discrete cellular signalling pathways.

   Convergence of the fanconi anemia and ataxia telangiectasia signaling pathways. · Taniguchi T, Garcia-Higuera I, Xu B, Andreassen PR, Gregory RC, Kim ST, Lane WS, Kastan MB, D'Andrea AD · Cell 2002
   PMID 12086603DOI
   ▸What this evidence cannot show

   Established for the cell systems used in that paper (human lymphoblast and fibroblast lines with defined FA and ATM genotypes). It says nothing about leukaemia and nothing about menin; it is cited here only to separate the two FANCD2 modifications, because that separation is what makes the checkpoint reading of the menin association distinct from the repair reading.

   ▸Evidence that argues against this (1)

   • The same group's later work shows the monoubiquitination site is critical for normal levels of homology-directed repair while the ATM phosphorylation site is not, so the two arms are separable but both real; treating S222 as 'the checkpoint site' and K561 as 'the repair site' is a simplification. Nakanishi K et al. 2005
3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Menin specifically interacts with FANCD2; the interaction is enhanced by gamma-irradiation; loss of menin in mouse embryonic fibroblasts increases sensitivity to DNA damage; and menin localises to chromatin and nuclear matrix, with the nuclear-matrix association enhanced by gamma-irradiation.

   Menin associates with FANCD2, a protein involved in repair of DNA damage. · Jin S, Mao H, Schnepp RW, Sykes SM, Silva AC, D'Andrea AD, Hua X · Cancer Res 2003
   PMID 12874027
   ▸What this evidence cannot show

   `established` only for the proposition that an association was detected in that system, which is the grading rule this graph uses. The full text could not be read through any of the five routes (no PMCID from Europe PMC or from efetch; no DOI in PubMed, Crossref or OpenAlex; publisher HTTP 403; OpenAlex oa_status 'closed'), so everything beyond the abstract comes from the IntAct curation of the same paper and from the 2008 paper's description of it. Confidence for any claim about leukaemia from this paper is weak: the systems are mouse embryonic fibroblasts and tagged constructs, and no leukaemia cell appears in it.

   ▸Evidence that argues against this (1)

   • The only study designed to test whether MEN1 and FANCD2 act together found that they do not: mutation spectra differ, and neither overexpression nor loss of MEN1 modified the crosslink sensitivity of FANCD2 mutants. Marek LR et al. 2008
4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

   In this project's DepMap sweep FANCD2 shows a Chronos difference of -0.150 between 11 KMT2A-rearranged and 43 other leukaemia lines, ranking 411 of 8,269 by raw effect with line-label permutation p = 0.071, against -0.361, rank 15 and p = 0.0026 for its obligate heterodimer partner FANCI.

   FA core complex subunit heterogeneity in the KMT2A-rearranged dependency scan · Oncofusion knowledge graph, mech-fa-core-complex-heterogeneity-kmt2a; DepMap Chronos via Open Targets 2026
   ▸What this evidence cannot show

   Eleven immortalised KMT2A-rearranged lines, none infant-derived, lineage assigned from DepMap free-text disease strings. p = 0.071 is not a negative result; it is an underpowered non-result, and the difference is in the expected direction. A rank of 411 of 8,269 is still the top 5% of the genome.

   ▸Evidence that argues against this (1)

   • The same analysis names the boring rival explanation it could not exclude: the spread across obligate FA subunits may be CRISPR reagent quality rather than biology, in which case FANCD2's weakness carries no mechanistic information at all. Oncofusion knowledge graph, mech-fa-core-complex-heterogeneity-kmt2a; DepMap Chronos via Open Targets 2026
5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   The FANCD2 monoubiquitination site is critical for normal levels of homology-directed repair of chromosomal double-strand breaks, whereas the ATM phosphorylation site is not; the defect in FA cells is mild relative to BRCA1 and BRCA2 mutant cells.

   Human Fanconi anemia monoubiquitination pathway promotes homologous DNA repair. · Nakanishi K, Yang YG, Pierce AJ, Taniguchi T, Digweed M, D'Andrea AD, Wang ZQ, Jasin M · Proc Natl Acad Sci U S A 2005
   PMID 15650050DOI
   ▸What this evidence cannot show

   Human FA patient cells (groups A, G, D2) and mouse Fanca-/- cells, with a reporter assay for homology-directed repair. This paper is one of the two PMIDs UniProt attaches to the sentence claiming the menin interaction is monoubiquitination-independent; its abstract does not mention menin and its subject is not menin.

   ▸Evidence that argues against this (1)

   • 'Critical for normal LEVELS of homology-directed repair' is a quantitative claim whose magnitude decides whether it matters here, and the magnitude is not carried into this entry. The monoubiquitination-versus-ATM-site dissociation is also a statement about FA pathway wiring in the system tested, not about any KMT2A-rearranged cell. Marek LR et al. 2008
6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   A designed genetic test in Drosophila found MEN1 and FANCD2 mutants have strikingly different mutation spectra and no detectable genetic interaction for crosslink sensitivity, and the authors concluded MEN1 plays a role in interstrand crosslink repair distinct from the Fanconi anaemia genes.

   MEN1 and FANCD2 mediate distinct mechanisms of DNA crosslink repair. · Marek LR, Kottemann MC, Glazer PM, Bale AE · DNA Repair (Amst) 2008
   PMID 18258493DOI
   ▸What this evidence cannot show

   Drosophila. Full text read via the PMC HTML route (PMC2277339) after fullTextXML returned HTTP 500. The authors state their own limit plainly: because MEN1 mutation alone causes no eye phenotype after damage, 'additive or epistatic relationships cannot be evaluated based solely on these genetic interaction studies'. It is a modifier screen with a negative result, not a formal epistasis test.

   ▸Evidence that argues against this (1)

   • The same paper preserves an escape route it cannot close: 'it is possible that these genes function together in the S-phase checkpoint, but not in the repair of interstrand crosslinks'. Both proteins are documented S-phase checkpoint components. Marek LR et al. 2008
7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

   FANCD2's dependency profile correlates with this project's FA-module score at r = 0.714 [0.552, 0.824] across 54 leukaemia lines while being substantially less required than the module members (mean -0.292 against module means two to three times deeper).

   Neither rival mechanism leaves its fingerprint, and the best correlate of Fanconi dependency is menin · Oncofusion knowledge graph, ana-fa-aldehyde-vs-replication; DepMap Chronos via Open Targets 2026
   ▸What this evidence cannot show

   Co-dependency correlation across 54 immortalised leukaemia lines. A correlation between two genes' Chronos profiles reflects shared line-to-line variation, which can be driven by growth rate, screen quality or lineage as easily as by shared function. FANCD2 correlating with the module it belongs to is close to the least surprising possible result.

   ▸Evidence that argues against this (1)

   • In the same analysis MEN1 correlates with the FA-module score at r = 0.670, i.e. almost as strongly as FANCD2 itself does, which shows how little discriminating power this statistic has for the question of who is in the module. Oncofusion knowledge graph, ana-fa-aldehyde-vs-replication; DepMap Chronos via Open Targets 2026
8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   There is no published literature on FANCD2 in KMT2A-rearranged leukaemia. A PubMed title/abstract search over FANCD2, FA-D2 and FACD against KMT2A, MLL and 'mixed-lineage leukemia' returns 1 record, which is a chidamide/MI-3 combination study that names DNA repair as an affected pathway but assays CHK1, CHK2, HR and NHEJ rather than FANCD2. Title-checked count: 0.

   PubMed counted-absence search for FANCD2 against KMT2A/MLL, run 2026-09-20, aliases searched and every hit title-checked · PubMed (NCBI) 2026
   ▸What this evidence cannot show

   A counted absence is only as good as its symbol set. FANCD2 has no widely used retired symbol, which makes this one safer than most, but title/abstract search cannot see supplementary tables or screens that name a pathway rather than a gene — and the single hit removed by title-checking is exactly a paper that names the pathway.

   ▸Evidence that argues against this (1)

   • The one record the query returned should not be dismissed entirely: it reports that combining an HDAC inhibitor with the menin-MLL inhibitor MI-3 in MLL-rearranged AML disrupts the DNA damage checkpoint at CHK1/CHK2 and impairs homologous recombination. That is a menin-inhibitor-to-DNA-repair link in the right disease, obtained without anyone looking at FANCD2. Ye J et al. 2019
9. E9weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.analysis

   A dependency triage over 262 curated interactors with no graph entry reached FANCD2 through the MEN1 bait with a difference of -0.153 (index mean -0.563 against -0.410, pan-essentiality -0.281, lymphoid half -0.70 against myeloid -0.48). The permutation null the triage had skipped was then computed: 50,000 two-sided label permutations per gene over the same 11 KMT2A-rearranged against 43 other-leukaemia lines gives FANCD2 p = 0.1241 - not significant even uncorrected, the second weakest of the five genes in the shortlist, and Bonferroni over the 254 genes scored leaves nothing below 0.05 (DAXX 0.0128, QKI 0.0060, HDGFL2 0.0187, TRAF2 0.1618, FANCD2 0.1241). 22 of the 254 cleared the difference threshold and the pan-essentiality filter did most of the work reducing that to five.

   Dependency triage of curated interactors with no graph entry, ranked by DepMap selectivity · Oncofusion knowledge graph, analyses/unwritten-interactors-dependency-triage; DepMap Chronos via Open Targets 2026
   ▸What this evidence cannot show

   The number does not support anything. It is recorded so that it is not rediscovered and mistaken for new evidence, and so that the three-leg convergence originally described for the menin-FANCD2 bridge is visible as having lost its gene-level leg. `weak` overstates it; the vocabulary has no lower grade. The lymphoid mean rests on four cell lines and the effect is small even before the null.

   ▸Evidence that argues against this (1)

   • This is now the second independent computation putting FANCD2 outside significance: the graph's own FA analysis gave rank 411 of 8,269 with permutation p = 0.071, and this one gives p = 0.1241 on a different arm construction. Two non-significant results in the same direction are still two non-significant results. Oncofusion knowledge graph, mech-fa-core-complex-heterogeneity-kmt2a; DepMap Chronos via Open Targets 2026
10. E_q71_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In HUMAN non-FA (normal) and FA patient cell lines, FANCD2 is monoubiquitinated in response to DNA damage and the activated isoform is TARGETED TO NUCLEAR FOCI (dots), colocalising with BRCA1 in ionizing-radiation-induced foci and in the synaptonemal complexes of meiotic chromosomes; the monoubiquitination that produces this relocalisation requires a NUCLEAR complex containing FANCA, FANCC, FANCF and FANCG.

    Interaction of the Fanconi anemia proteins and BRCA1 in a common pathway. · Garcia-Higuera I, Taniguchi T, Ganesan S, Meyn MS, Timmers C, Hejna J, Grompe M, D'Andrea AD · Mol Cell 2001
    PMID 11239454DOI
    ▸What this evidence cannot show

    Abstract read in full; full text not read. Human fibroblast/lymphoblast FA complementation-group lines and meiotic chromosome spreads - NOT leukaemia, not haematopoietic, not KMT2A-rearranged (brief rule 73). The compartment is DAMAGE-INDUCED: the paper describes where FANCD2 goes after a genotoxin, not where it sits in an unperturbed leukaemic nucleus, and no such measurement exists in any KMT2A-rearranged cell. 'Foci' is an immunofluorescence readout, not a biochemical fractionation, so it carries no size or stoichiometry information.

    ▸Evidence that argues against this (1)

    • The foci are a property of the ID2 heterodimer, not of FANCD2 alone, and reading them as a FANCD2 address over-attributes. FANCI was subsequently identified as a monoubiquitinated FANCD2 paralog; FANCI and FANCD2 associate as the ID complex and localise to chromatin TOGETHER in response to DNA damage, with ubiquitination of each required to maintain ubiquitin on the other. A subnuclear claim written against either gene alone therefore describes a shared compartment. Smogorzewska A et al. 2007
11. E_q71_ct1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In FANCONI ANAEMIA PATIENTS (all complementation groups pooled) and in human FA or FA-like cells, haematopoietic stem and progenitor cells show a profound defect that is present BEFORE the onset of clinical bone marrow failure; in response to replicative stress and unresolved DNA damage p53 is hyperactivated and triggers a late p21(Cdkn1a)-dependent G0/G1 cell-cycle arrest, and knockdown of p53 rescues the HSPC defect in several in vitro and in vivo models.

    Bone marrow failure in Fanconi anemia is triggered by an exacerbated p53/p21 DNA damage response that impairs hematopoietic stem and progenitor cells. · Ceccaldi R, Parmar K, Mouly E, Delord M, Kim JM, Regairaz M, Pla M, Vasquez N, Zhang QS, Pondarre C, Peffault de Latour R, Gluckman E, Cavazzana-Calvo M, Leblanc T, Larghero J, Grompe M, Socié G, D'Andrea AD, Soulier J · Cell Stem Cell 2012
    PMID 22683204DOI
    ▸What this evidence cannot show

    Abstract read in full; full text not read. HUMAN FA patients and FA/FA-like cells in NORMAL haematopoiesis - not leukaemia, not KMT2A-rearranged, not an infant B-lineage measurement (brief rule 73). Critically, this is a PATHWAY-LEVEL result: the abstract pools complementation groups and does not attribute the HSPC defect to any single FA gene, so it is cited here as the cell-type context of the FA pathway and NOT as a measurement on this gene. The rescue is p53 knockdown, which is an epistasis result about the response, not about the lesion.

    ▸Evidence that argues against this (1)

    • This is the strongest single argument that an FA-gene requirement in a haematopoietic cell is a LINEAGE property and not a fusion property. The HSPC defect is measured in Fanconi anaemia patients who carry no fusion oncoprotein at all, and it is driven by a generic p53/p21 response to replication stress rather than by anything specific to KMT2A. Any claim that this gene is required 'because of the fusion' has to explain why a fusion-free human shows the same compartment failing. Ceccaldi R et al. 2012

• binds →FANCI — ID2 module; rank 15 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  obligate ID2 heterodimer; binds and scans dsDNA

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

  FANCL with UBE2T monoubiquitinates FANCD2 on K561

• ← regulatesUBE2T — E2; rank 33 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  E2 for the K561 monoubiquitination

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

  one 2003 co-IP in MEFs; menin 214-390 sufficient; gamma-IR enhanced

• ← part ofMenin-FANCD2: one 2003 co-immunoprecipitation, a curated sufficient region of menin 214-390, and nothing since

  FANCD2 side of the interface

• refutes →KMT2A rearrangement creates a requirement for the FA ubiquitin-transfer module — the claim, its prior, and the four results that would kill it

  the ID2 effector is the weakest FA gene in the scan: d = -0.150, rank 411, p = 0.071

• supports →The FA core complex does not move together, and that is the strongest argument against reading this as a Fanconi-pathway dependency at all

  the substrate-versus-module discrepancy this entry documents

• related to →The Fanconi signal is in both lineage halves, survives the correlation-proof null, and is still there when lineage is held constant

  the ID2 effector's numbers within the FA lineage-split analysis

• related to →BRIP1 — downstream effector; rank 22 of 8,269 for KMT2A-selective dependency, and no KMT2A literature at all

  FA pathway member in the same DepMap selectivity scan

• ← related toUniProt's menin 214-390 FANCD2 region is not a surface: it spans three spatial lobes and 29.9% of menin's solvent-accessible area, so its twenty-residue overlap with the gamma-H2A.X motif is what chance gives you

  the whole protein-level menin-FANCD2 record is one 2003 co-immunoprecipitation with no mapped interface

• related to →The soluble-nucleoplasm versus chromatin-bound axis - the one localisation readout that keeps a size dimension, and nobody has run it on a KMT2A fusion

  FANCD2 enters damage-induced nuclear foci and the ID complex loads onto chromatin, so the chromatin-bound versus soluble axis is the readout this gene needs - and it has never been run on it in a KMT2A-rearranged cell (human FA lines; PMID 11239454)

• regulates →Bone marrow haematopoietic stem cell — the postnatal comparator, and why the same fusion behaves differently here

  REQUIRED IN: in Fanconi anaemia patients the FA pathway is required in haematopoietic stem and progenitor cells, which are depleted before clinical marrow failure - pathway-level, no fusion present (PMID 22683204)

• regulates →Quiescence (G0) — the state 15 entries invoke to explain measurable residual disease, defined here for the first time

  FA-pathway loss IMPOSES a p53/p21-dependent G0/G1 arrest on human HSPCs; the state appears when the gene is lost rather than being maintained by it (PMID 22683204)

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

  the nuclear FANCA/FANCC/FANCF/FANCG complex is required to monoubiquitinate FANCD2, which is what targets FANCD2 to nuclear foci - so FANCF sets FANCD2's subnuclear address rather than having one of its own (PMID 11239454)

• ← related toBOD1L (BOD1L1) — the top SETD1A co-dependency, and the only candidate in this cycle tested by CRISPR tiling in a KMT2A::AFF1 line

  same fork-protection pathway by phenotype: BOD1L loss and a chaperone-defective FANCD2 both cause DNA2-dependent resection of stressed forks and loss of RAD51 nucleofilament stability, with SETD1A H3K4 methylation upstream of FANCD2 (human cells; PMIDs 26166705, 29937342)

• regulates →Replication stress and interstrand-crosslink repair as a cell state - an ATR dependency measured in mouse MLL-rearranged AML, and the same stress implicated in breaking the MLL locus in the first place

  ID2-module effector and fork histone chaperone; the state gene-fancd2 named as missing vocabulary rather than approximating to quiescence