MBTD1 - the H4K20me1/me2 reader inside NuA4/TIP60, the quiescence brake on adult haematopoietic stem cells, and the 3' partner of ZMYND11::MBTD1 whose methyl-lysine cage the fusion carries and does not use

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

• MBTD1's normal function is a brake. Conditional knockout does not deplete HSCs - it expands them, makes them cycle hyperactively and breaks their stress response. An agent that disables MBTD1, or the EPC1 surface that recruits it, would be releasing that brake in every normal stem cell it reaches. In an infant, whose marrow has to work for eighty years, releasing HSC quiescence is a worse risk than a cytopenia.
• MBTD1 promotes homologous-recombination repair and restrains 53BP1. Disabling it would be expected to shift double-strand-break repair pathway choice towards end joining - which is the pathway this project's break-machinery layer argues produces the KMT2A translocation in the first place. A drug that biases repair towards NHEJ in a haematopoietic progenitor is a drug with a second-malignancy argument attached.
• The complex MBTD1 sits in is near-essential. KAT5 mean gene effect -0.607 with 70.2% of 1,258 lines below -0.5, and EP400 -0.932 with 94.3% below -0.5. 'Inhibit the hijacked acetyltransferase' is an essential-function toxicity proposal. MBTD1 itself is shallower (-0.353, 17.9%), which is the reason to target the reader rather than the enzyme - and also the reason a reader inhibitor may do very little.
• The fusion does not use MBTD1's reader function, so a methyl-lysine-cage ligand against MBT4 would be a chemically tractable drug against the wrong module. This is the mirror of the LEDGF PWWP problem this graph already records: the arm with chemistry is not always the arm with genetics.
• Every haematopoietic statement in this entry is mouse. There is no human MBTD1 loss-of-function phenotype, no human expression atlas reading quoted here, and no patient with MBTD1 haploinsufficiency described. A toxicity argument built on a mouse conditional knockout is a hypothesis about humans.

1. Is MBTD1 or any NuA4/TIP60 subunit differentially required in KMT2A-rearranged lines? The DepMap numbers in this entry are unsplit by lineage or fusion status, and the split is compute on data this project already queries. It is the cheapest MBTD1 question in the corpus.
2. Does the H4K20me cage matter for LEUKAEMIA MAINTENANCE even though it is dispensable for transformation? The cage mutants were scored in a serial-replating immortalisation assay. A dispensable-for-initiation module can be required for maintenance, and the distinction decides whether MBT4 is a target surface or a bystander.
3. Is MBTD1 a reader that protects its own mark? The mouse-oocyte report has it associating with Pr-Set7/SETD8 and stabilising H4K20me1. If that holds in a haematopoietic cell it inverts the usual reader logic and would mean depleting MBTD1 lowers H4K20me1 as well as un-reading it - which the BAHCC1 replication work would care about.
4. Why does adult Mbtd1 loss EXPAND the HSC pool when fetal MBTD1 sustains it? The same paper reports both directions. A stage-dependent reversal in a chromatin reader is the kind of thing this project's fetal-versus-adult origin arguments turn on, and the mechanism is unexplained.
5. Has anyone put MBTD1 and KMT2A in the same query? MBTD1[tiab] AND KMT2A[tiab] and MBTD1[tiab] AND MLL[tiab] were deliberately NOT run in this beat, so the pairing is unbounded rather than empty. One query and its title check would settle it, and by this project's own registry the KMT2A operand must be run alongside MLL before any zero is believed.

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

   MBTD1 is UniProt Q05BQ5 (MBTD1_HUMAN), 628 aa, the single reviewed human record with MBTD1 as primary gene name and no UniProt gene synonyms, HGNC:19866 at 17q21.33 with primary Ensembl id ENSG00000011258 and RefSeq NM_017643, GRCh38 chr17:51,177,419-51,261,114 on the minus strand; its feature table gives an FCS-type zinc finger at 45-80 and four MBT repeats at 141-245, 253-350, 351-456 and 464-560, and its two subcellular-location comments are Nucleus (ECO:0000305, curator inference) and Chromosome (ECO:0000269, PubMed 27153538).

   UniProtKB Q05BQ5 (MBTD1_HUMAN); HGNC:19866 (mbt domain containing 1); Ensembl ENSG00000011258 · UniProt / HGNC / Ensembl 2026
   ▸What this evidence cannot show

   All three records fetched live on 2026-09-23 and the primary gene name read back. Only one reviewed row is returned so the several-rows trap does not fire, and there is no paralogue close in length that an accession swap could hide behind - the nearest four-MBT-repeat relatives L3MBTL1 (772 aa), L3MBTL2 (705 aa) and SFMBT1 (866 aa) are all far longer, and none carries MBTD1 as an alias. The Ensembl coordinates are GRCh38; no GRCh37 coordinates were fetched, so none are quoted.

   ▸Evidence that argues against this (1)

   • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] A UniProt record is a curation decision, not evidence about the protein. 'Single reviewed record with no gene synonyms' is a statement about UniProt's nomenclature at one moment - and the absence of synonyms is exactly the condition under which a symbol-coverage check can miss prior work published under an unindexed alias. STRONGER COUNTER: a literature search under historical names for this locus, which was not run.
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   The crystal structure of MBTD1 residues 130-566, covering the four MBT repeats, at 2.5 A resolution shows an asymmetric rhomboid architecture resembling L3MBTL2, and fluorescence polarization confirms that MBTD1 preferentially binds mono- and di-methyl-lysine histone peptides through one of its four repeats using a semi-aromatic cage.

   Structural studies of a four-MBT repeat protein MBTD1 · Eryilmaz J, Pan P, Amaya MF, et al. · PLoS ONE 2009
   PMID 19841675DOI
   ▸What this evidence cannot show

   Read at abstract level plus the RCSB entry for 3FEO. Binding is to histone PEPTIDES by fluorescence polarization, not to nucleosomes; the MBT family is known to show little peptide-sequence selectivity, so this establishes methyl-state preference more firmly than it establishes H4K20 as the physiological ligand - that assignment comes from E3. The paper frames MBTD1 as a Polycomb-group protein, which the later complex purification contradicts.

   ▸Evidence that argues against this (1)

   • The paper's own framing of MBTD1 as a Polycomb-group transcriptional repressor was overturned by its identification as a stable subunit of the NuA4/TIP60 acetyltransferase complex, a coactivator. Domain homology within the MBT family predicted repression; complex purification found the opposite. Jacquet K et al. 2016
3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   MBTD1 is a stable subunit of the NuA4/TIP60 acetyltransferase complex and, through a histone reader domain for H4K20me1/2, allows TIP60 to associate with specific gene promoters and to promote repair of DNA double-strand breaks by homologous recombination; the complex regulates 53BP1 association partly by competing for H4K20me2 and partly by TIP60 acetylating H2AK15, the residue whose RNF168-dependent ubiquitylation 53BP1 requires.

   The TIP60 Complex Regulates Bivalent Chromatin Recognition by 53BP1 through Direct H4K20me Binding and H2AK15 Acetylation · Jacquet K, Fradet-Turcotte A, Avvakumov N, et al. · Molecular Cell 2016
   PMID 27153538DOI
   ▸What this evidence cannot show

   Read at abstract level; no PMCID was pursued. Human cell lines, not haematopoietic primary cells, and the DNA-repair readouts are reporter and focus assays rather than patient material. UniProt cites this paper as the experimental support for both the H4K20me1/2 binding and the Chromosome localisation of Q05BQ5, and it is the paper the function comment is actually about - so this is a case where the database pointer and the source agree, unlike several logged in this project.

   ▸Evidence that argues against this (1)

   • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] The reader specificity and complex membership come from a study whose primary question was 53BP1 and bivalent chromatin, so MBTD1's role is established in a DNA-damage and bivalency context rather than in leukaemia. Stable subunit membership in a purified complex does not establish that the interaction is required, stoichiometric, or present in haematopoietic cells. STRONGER COUNTER: a haematopoietic-cell proteomic study reporting NuA4/TIP60 composition without MBTD1. Not searched for.
4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   The crystal structure of the MBTD1-EPC1 complex shows a hydrophobic C-terminal fragment of EPC1 engaging the MBT repeats of MBTD1 at a site distinct from the H4K20me-binding site, and cellular assays validate the key residues; EPC1 is the NuA4/TIP60 non-catalytic subunit that enlists MBTD1 into the complex.

   Structural Basis for EPC1-Mediated Recruitment of MBTD1 into the NuA4/TIP60 Acetyltransferase Complex · Zhang H, Devoucoux M, Song X, et al. · Cell Reports 2020
   PMID 32209463DOI
   ▸What this evidence cannot show

   Read at abstract level plus the RCSB records. Europe PMC returns no PMCID and efetch's ArticleIdList carries no pmc id, so routes 1-4 of the unreadable test fail; the DOI route was not attempted. IMPORTANT CONSTRUCT CAVEAT: the deposit associated with this work, 6NFX, is a SINGLE polymer entity of 470 residues annotated to both Q05BQ5 and Q9H2F5 - an engineered single-chain chimera, not a two-chain co-crystal. It models the bound state; the association itself rests on this paper's cellular assays and on the co-purification in E3.

   ▸Evidence that argues against this (1)

   • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] A crystal structure of a complex demonstrates that two fragments CAN bind in vitro at the concentrations and constructs used; it does not establish the interaction occurs in cells or matters for function. The 'site distinct from the histone-binding site' conclusion is a structural inference about one crystal form. STRONGER COUNTER: a cell-based separation- of-function mutant showing EPC1 binding is dispensable. Not searched for.
5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   In mice, MBTD1 is highly expressed in fetal HSCs and sustains their number and function; germline-null animals die soon after birth, apparently from skeletal abnormalities; and in conditional knockout adults HSC and progenitor numbers increase, HSCs show a hyperactive cell cycle and a defective response to exogenous stress, with MBTD1 binding the FoxO3a promoter directly, interacting with TIP60-complex components, and restoration of FOXO3a activity rescuing both the cycling and the pool-size abnormalities.

   MBTD1 preserves adult hematopoietic stem cell pool size and function · Takubo K, Htun PW, Ueda T, et al. · Proceedings of the National Academy of Sciences of the USA 2023
   PMID 37523546DOI
   ▸What this evidence cannot show

   MOUSE THROUGHOUT, conditional knockout; no human data and no leukaemia model. Read at abstract level only. 'Numbers increased' and 'defective stress response' are stated in the abstract without effect sizes, so no magnitude is quoted here. The FoxO3a promoter binding is ChIP in mouse haematopoietic cells; whether MBTD1 acts there through NuA4/TIP60 or independently is not resolved by the abstract.

   ▸Evidence that argues against this (1)

   • The direction of the adult phenotype is the opposite of the fetal one: MBTD1 SUSTAINS fetal HSC number, and its loss INCREASES adult HSC and progenitor number. A single sentence describing MBTD1 as 'required for HSC maintenance' would be true of the fetal result and false of the adult one, and the paper itself is the source of both. Takubo K et al. 2023
6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   In human AML, the recurrent t(10;17)(p15;q21) fuses ZMYND11 exons 1-12 to MBTD1 exons 3-17 in frame, producing a 1,054-residue protein that retains the FCS zinc finger and all four MBT repeats of MBTD1, with the reciprocal product predicted to lack a productive start codon.

   Recurrent translocation t(10;17)(p15;q21) in minimally differentiated acute myeloid leukemia results in ZMYND11/MBTD1 fusion · de Rooij JD, van den Heuvel-Eibrink MM, Kollen WJ, et al. · Genes, Chromosomes & Cancer 2016
   PMID 26608508DOI
   ▸What this evidence cannot show

   Abstract level only; no PMCID by either route. The exon-level junction and the retained-domain statement are cross-checked against the 2026 series' explicit statement that the fusion protein retains PHD, bromodomain and PWWP from ZMYND11, and against the construct boundaries in the founding functional study - three sources agreeing, none of them a protein-level measurement of the patient chimera.

   ▸Evidence that argues against this (1)

   • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] The exon structure and the 1,054-residue product come from a single report. Fusion breakpoints vary between cases even in well-characterised fusions, so exons 1-12 to 3-17 should be read as ONE observed junction and not as the architecture of the entity. STRONGER COUNTER: a second case with a different junction, which would be expected and which nobody has looked for across the 9 records.
7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   In low-grade endometrial stromal sarcoma, MBTD1 is the 5' partner of a novel reciprocal t(X;17)(p11.2;q21.33) producing an MBTD1-CXorf67 fusion, identified in two tumours by paired-end RNA sequencing with FISH and cytogenetics and confirmed by RT-PCR and Sanger sequencing, with a third case found by a purpose-built FISH assay among 25 further uterine stromal tumours negative for JAZF1 and YWHAE rearrangements.

   Identification of a novel, recurrent MBTD1-CXorf67 fusion in low-grade endometrial stromal sarcoma · Dewaele B, Przybyl J, Quattrone A, et al. · International Journal of Cancer 2014
   PMID 23959973DOI
   ▸What this evidence cannot show

   Abstract level only. Three cases total. CXorf67 is the gene now widely called EZHIP; NOTE that `MBTD1[tiab] AND EZHIP[tiab]` returns 0 records while `MBTD1[tiab] AND CXorf67[tiab]` returns 4, so the current symbol is the WRONG search token for this literature - the retired-symbol failure mode, live. No functional characterisation of MBTD1-CXorf67 was found and none is claimed.

   ▸Evidence that argues against this (1)

   • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] This is a DIFFERENT DISEASE and, critically, the OPPOSITE ORIENTATION - MBTD1 is the 5' partner in the sarcoma fusion and the 3' partner in the leukaemia one. So it argues against, not for, a single mechanistic story about MBTD1's contribution: whatever MBTD1 supplies must differ when it donates the N-terminus versus the MBT repeats. The entry should not use this to generalise MBTD1's role. STRONGER COUNTER: none needed - the claim's own content limits it.
8. E8established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   PDB 4C5I, titled 'Crystal structure of MBTD1 YY1 complex', contains a 440-residue MBTD1 entity cross-referenced to Q05BQ5 together with a 30-residue human YY1 peptide cross-referenced to P25490, at 2.59 A, and its primary citation is a study of the Drosophila chromatin repressor Sfmbt and Polycomb response elements.

   Structural basis for targeting the chromatin repressor Sfmbt to Polycomb response elements · Alfieri C, Gambetta MC, Matos R, et al. · Genes & Development 2013
   PMID 24186981DOI
   ▸What this evidence cannot show

   The PDB entry and its entity records were fetched live on 2026-09-23; the paper itself was NOT read, at any level, and no biological claim about MBTD1-YY1 is made here. The title above is PubMed's, not RCSB's - RCSB stores a differently capitalised string, which is the mismatch shape the citation checker exists to catch. This evidence item exists only to record the deposit's real contents so that a future beat does not read the RCSB title as an MBTD1-YY1 finding.

   ▸Evidence that argues against this (1)

   • [COORDINATOR SALVAGE NOTE, CYCLE 82 - WITHDRAWN IN CYCLE 86] I flagged PDB 4C5I's title as submitter-authored free text whose partner identity was unverified. **CHECKED AND IT STANDS.** Q123 fetched both polymer entities and aligned them against UniProt the same session: entity 1 is a 3-residue GAM cloning remnant followed by an EXACT 437-residue match to Q05BQ5 130-566, one contiguous block with no substitutions; entity 2 is an EXACT match to YY1 (P25490) 199-228. The title is accurate and so is E8. The only refinement available is that the '440-residue entity' includes the 3-residue tag.
9. E9established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   PDB 6NFX, at 1.95 A and titled 'MBTD1 MBT repeats', is a single polymer entity of 470 residues whose description reads 'MBT domain-containing protein 1,Enhancer of polycomb homolog 1' and which carries UniProt cross-references to both Q05BQ5 and Q9H2F5 - an engineered single-chain MBTD1-EPC1 fusion construct rather than a two-chain complex.

   RCSB PDB entry 6NFX and polymer entity 6NFX_1 (rcsb_polymer_entity.pdbx_description; reference_sequence_identifiers Q05BQ5 and Q9H2F5); PDB entries 3FEO and 4C5I · RCSB Protein Data Bank 2026
   ▸What this evidence cannot show

   Fetched live 2026-09-23 from the RCSB data API. RCSB reports polymer_entity_count_protein = 1 for this entry. The residue ranges modelled were NOT read off the entity sequence and are not quoted; per this project's rule, a span should be derived by aligning the fetched entity sequence to a same-day UniProt sequence, which was not done. The chimeric construction is a statement about the deposit, not a criticism of the paper.

   ▸Evidence that argues against this (1)

   • [COORDINATOR SALVAGE NOTE, CYCLE 82 - WITHDRAWN IN CYCLE 86] I flagged PDB 6NFX's entity description as a chimeric construct whose 470 residues are not MBTD1's alone. **THAT IS WHAT E9 ALREADY SAID** - it describes 'an engineered single-chain MBTD1-EPC1 fusion construct rather than a two-chain complex'. My note restated the entry's own content as though correcting it. Q123 added the measurement it actually lacked: the 470 residues are MBTD1 140-562 + linker + EPC1 (Q9H2F5) 644-671.
10. E10moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In mouse oocyte meiotic maturation, MBTD1 is associated with Pr-Set7 and stabilises H4K20me1.

    MBTD1 is associated with Pr-Set7 to stabilize H4K20me1 in mouse oocyte meiotic maturation · Luo YB, Ma JY, Zhang QH, et al. · Cell Cycle 2013
    PMID 23475131DOI
    ▸What this evidence cannot show

    TITLE ONLY, plus its retrieval as one of two records under MBTD1[tiab] AND H4K20[tiab]. Mouse oocytes - a system with no relation to haematopoiesis, and one in which maternal protein stores make loss-of-function timing hard to interpret. Recorded because it is the only report in which MBTD1 affects the LEVEL of its own ligand rather than merely reading it, which would make it a stabiliser as well as a reader; nothing about the mechanism is asserted.

    ▸Evidence that argues against this (1)

    • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] Mouse OOCYTE MEIOTIC MATURATION is about as far from human haematopoiesis as a chromatin result can be - different species, a transcriptionally quiescent cell type, and a specialised chromatin state. That MBTD1 stabilises H4K20me1 there is weak evidence that it does so in a leukaemic blast. STRONGER COUNTER: an H4K20me1 measurement on MBTD1 loss in a haematopoietic cell. Does not appear to exist.
11. E11strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In murine haematopoietic stem/progenitor cells, deletion of the four MBT repeats of the MBTD1 portion of ZMYND11-MBTD1, or the triple point mutation A661D/L684D/L688D that disrupts the MBTD1-EPC1 contact, abolishes both Tip60 binding and transformation, whereas the H4K20me1/2-binding-defective mutants W954A and F951A/W954A/Y958A (fusion numbering) bind Tip60 normally and transform normally, and deletion of the FCS zinc finger transforms with a reduced phenotype.

    ZMYND11-MBTD1 induces leukemogenesis through hijacking NuA4/TIP60 acetyltransferase complex and a PWWP-mediated chromatin association mechanism · Li J, Galbo PM Jr, Gong W, et al. · Nature Communications 2021
    PMID 33594072DOI
    ▸What this evidence cannot show

    Full text read this session via PMC7886901. Readout is serial replating immortalisation of primary murine HSPCs plus co-immunoprecipitation - a transformation assay, not a leukaemia-in-vivo assay, for the mutant panel. Residue numbers are in FUSION coordinates as the paper states them; the conversion to MBTD1 coordinates (W529, F526, Y533) is this entry's arithmetic from the stated construct boundaries and is labelled as such. The negative result for the H4K20me cage is a negative in ONE assay in ONE system, and a dispensable module in a transformation assay is not necessarily a dispensable module in leukaemia maintenance or in a patient.

    ▸Evidence that argues against this (1)

    • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] Murine HSPCs with a retrovirally expressed fusion: the construct is over-expressed relative to a translocated allele, which this project has already found inflates occupancy-derived estimates roughly three-fold in a comparable system. A domain-deletion or point-mutant requirement in that setting shows the domain is needed FOR THE OVEREXPRESSED CONSTRUCT'S activity, which is the standard caveat on every retroviral leukaemogenesis model. STRONGER COUNTER: the same mutants at endogenous expression, e.g. knock-in. Not searched for.
12. E12strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    DepMap release 26/06 via Open Targets, fetched 2026-09-23: MBTD1 mean gene effect -0.353 across 1,258 screens (17.9% below -0.5, 0.24% below -1.0), against ZMYND11 +0.011, EPC1 -0.336, KAT5 -0.607 (70.2% below -0.5), EP400 -0.932 (94.3% below -0.5) and MEN1 -0.362; the pan-essential control EEF2 returned -2.709 over 1,258 screens and the low-dependency control OR4F5 -0.043 over 5, while the intended second pan-essential control POLR2A returned zero screens through the identical query shape.

    Open Targets Platform GraphQL API, target.depMapEssentiality, meta.dataVersion year 26 month 06; genes resolved through HGNC primary ensembl_gene_id · DepMap / Open Targets 2026
    ▸What this evidence cannot show

    Screen counts are gene-specific and must not be read as a corpus denominator. NO lineage split and NO KMT2A-rearranged-versus-other comparison was run, so none of these numbers supports a selectivity claim; the mean is a raw effect, not a z. MBTD1 at 17.9% below -0.5 with 0.24% below -1.0 is a shallow, broad dependency with little dynamic range, which is the floor-compression situation in which small differences are uninterpretable rather than small. POLR2A's zero is recorded because it shows a single-gene zero from this endpoint carries no information.

    ▸Evidence that argues against this (1)

    • [COORDINATOR SALVAGE NOTE, CYCLE 82 - WITHDRAWN IN CYCLE 86] I wrote that the Open Targets API does not state a release and that E12's 'DepMap release 26/06' string 'needs re-checking or removing'. **WRONG, AND THE PREMISE WAS FALSE.** I queried the API directly in cycle 86: `{meta{dataVersion{year month iteration}}}` returns year "26", month "06". The figures also reproduce on a fresh call (MBTD1 -0.3531 against E12's -0.353, over 1,258 screens). **The only real defect is the LABEL:** `meta.dataVersion` is the OPEN TARGETS PLATFORM release, not a DepMap release name like 24Q2. The honest field is "Open Targets dataVersion 26.06". Nothing needed removing.
13. E13established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    MBTD1 had no entry in this graph before 2026-09-23 and was named in exactly two node files, both written in the preceding cycle; scripts/is-symbol-covered.py returns 'MBTD1 no match under 5 candidate names'.

    scripts/is-symbol-covered.py MBTD1 and scripts/beat-inventory.py --symbols MBTD1 ..., run 2026-09-23 over 2331 node files (HGNC rows 45083, retired/alias index 58536) · oncofusion-knowledge-graph repository 2026
    ▸What this evidence cannot show

    A no-match verdict from that tool is explicitly NOT a gap verdict - the tool's own output says the thing may live inside a complex, mechanism or mark entry with no name of its own. The word-boundary naming count of 2 and the absence of any fusion, analogy or translocation entry were checked separately, and the two naming files are gene-kat5 (which names MBTD1 as a NuA4/TIP60 subunit) and gene-zmynd11 (which names the fusion and asks for this entry). The URL field carries a repository placeholder because the source is a local script run, not a web resource.

    ▸Evidence that argues against this (1)

    • [ADDED BY THE COORDINATOR DURING SALVAGE, 2026-09-23 cycle 82. Q121 was killed by the account session limit before it wrote this field. This is a reasoned methodological limitation derived from the claim and its citation type, NOT a literature-sourced rebuttal - no counter-citation was fetched, and none is asserted.] A self-referential claim about this repository's own state, which cannot be checked by a reader outside it and which was true only at one commit. It also rests on is-symbol-covered.py, whose directional output I personally mis-stated three times in three cycles - so 'no match under 5 candidate names' bounds the tool's alias index, not the literature. STRONGER COUNTER: none applicable; the item is provenance, not biology, and should be read as such.

• part of →ZMYND11::MBTD1, t(10;17)(p15;q21) - a chromatin reader welded to the NuA4/TIP60 acetyltransferase, maintaining Hoxa/Meis1/Myb/Myc/Sox4 in AML: the KMT2A-fusion architecture built from different parts

  contributes the 3' half: exons 3-17, the FCS zinc finger and all four MBT repeats - the EPC1-binding scaffold the fusion needs, and a methyl-lysine cage it does not

• part of →KAT5 / Tip60 - the acetyltransferase in the only direct KMT2A-fusion histone-variant result, and both fusions tested were myeloid

  stable subunit of the same NuA4/TIP60 complex; MBTD1's H4K20me reading is what lets TIP60 reach specific promoters

• part of →EP400 - the ATPase that swaps H2A.Z in; its complex is required by MLL leukaemia stem cells, and in DepMap it is one of the three genes LEAST required in KMT2A-rearranged lines

  co-subunits of NuA4/TIP60; EP400 is the H2A.Z-exchange ATPase and is near-pan-essential (DepMap -0.932, 94.3% of 1,258 lines below -0.5, release 26/06)

• binds →H4K20me2 - 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

  second H4K20me2 reader, via MBT repeat 4 (464-560); COMPETES WITH 53BP1 for the same mark - an addition to that entry's reader list, which holds only 53BP1 and Crb2

• related to →TP53BP1 (53BP1) - the end-protection factor that decides between the two repair routes, reachable only under the wrong symbol

  NuA4/TIP60 restrains 53BP1 chromatin binding through MBTD1 competing for H4K20me2 and through TIP60 acetylating H2AK15 to block the ubiquitylation 53BP1 needs

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

  MOUSE: high in fetal HSCs and required to sustain them; conditional knockout in adults EXPANDS HSCs and progenitors and makes them cycle hyperactively

• analogous to →In leukaemia KMT2A gives the front of the protein; in sarcoma it gives the back. Every consequence follows from that.

  same phenomenon in a second gene: MBTD1 is the 3' partner in AML (ZMYND11::MBTD1) and the 5' partner in low-grade endometrial stromal sarcoma (MBTD1::CXorf67)

• fuses with →ZMYND11 (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

  t(10;17)(p15;q21); ZMYND11 exons 1-12 to MBTD1 exons 3-17, in frame, 1,054 aa

• related to →The differentiation block — what it actually is, operationally, in KMT2A::AFF1 leukaemia

  only through the fusion - wild-type MBTD1's own state axis is quiescence versus cycling, through direct binding at the FoxO3a promoter (MOUSE)

• binds →H4K20me1 - the transcription- and replication-linked H4 state, whose gene-body level tracks Pol II; it HAS been profiled genome-wide in human haematopoietic stem cells twice, and a query scoped to 'leukemia' cannot see either dataset

  MBT repeat 4 (Q05BQ5 464-560) reads me1 as well as me2; ZMYND11::MBTD1 retains this cage and its cage mutants transform normally, so the fusion carries the reader and does not use it

• ← related toZMYND11::MBTD1 is the KMT2A-fusion architecture assembled from different parts - and the part it throws away is a direct physical brake on KMT2A itself

  the effector-recruiting module

• ← related toMBTD1 is the effector-recruiting half of a leukaemia fusion and the addressing half of a sarcoma fusion while donating nearly all of itself to both - so the junction, not the parts, assigns the role

  the gene that occupies both structural roles while donating nearly all of itself to both fusions

• ← part ofMBTD1::EZHIP (MBTD1-CXorf67), t(X;17)(p11.2;q21.33) - MBTD1's other fusion, in which a NuA4 subunit delivers an H3K27M-mimicking PRC2 poison to low-grade endometrial stromal sarcoma

  MBTD1 aa 1-589 is the 5' module; only the final 39 residues are absent

• fuses with →MBTD1::EZHIP (MBTD1-CXorf67), t(X;17)(p11.2;q21.33) - MBTD1's other fusion, in which a NuA4 subunit delivers an H3K27M-mimicking PRC2 poison to low-grade endometrial stromal sarcoma

  MBTD1 is the 5' partner here and the 3' partner in ZMYND11::MBTD1 - the polarity reversal

• ← fuses withEZHIP (CXorf67, CATACOMB) - a protein that jams PRC2 by pretending to be an H3K27M oncohistone, and the 3' payload of MBTD1's other fusion

  MBTD1::EZHIP in low-grade endometrial stromal sarcoma, with MBTD1 as the 5' partner

• ← bindsEPC1 - the NuA4/TIP60 subunit whose C-terminal fragment grips MBTD1, making it the contact a leukaemic fusion cannot do without and the mildest dependency in that whole complex

  a hydrophobic C-terminal fragment of EPC1 (644-671 in PDB 6NFX) grips MBTD1's MBT repeats at a site distinct from the H4K20me cage