Oncofusion

start / graph / ana-interactome5-enumeration

Analysismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q5-interactome5

Interactome round 5: 42 queries, 13 entries written, 9 candidates rejected with counts, and four alias failures of four different kinds

also called interactome5 enumeration · round 5 enumeration ledger

Before writing anything, this round counted how many published papers exist for each candidate protein in KMT2A leukaemia, and checked every dependency number twice - once in myeloid cells and once in B-cells. Most of the proteins the brief asked about had almost no literature; two of them had strong-looking dependency numbers that evaporated when the cells were split by type.

The same thing, for a scientist

**42 PubMed record-count queries** were run through NCBI E-utilities on 2026-09-19, with every record in every decision-bearing query **title-checked individually** rather than counted, plus a scripted read of `analyses/depmap-kmt2a-selective-scan-full` and of this beat's own lineage-split run for all 23 panel genes [E1][E2]. **Thirteen entries were written; nine candidates were rejected or deferred with their counts.** Three findings dominate. **(1) The brief's two 'structural holes' are not the same kind of hole.** FACT has no published physical link to the KMT2A axis at all - 0 records after title-check against AFF1/AFF4/ELL2/SEC/MLLT1/MLLT3, 0 against DOT1L/H3K79 - and both subunits are anti-dependencies. NSL has four directly relevant papers including five NUP98 fusions binding MLL1 and NSL [E3]. **(2) Round 4's SETD1B deferral was wrong on the literature and the reason is an alias class nobody had named**: a primary MLL-rearranged AML paper existed [E4], reachable when the spelling `Set1B` is added to `SETD1B OR KMT2G OR KIAA1076`, which raises the count from 17 to 27. **(3) Round 4's CXXC1 deferral was based on the same 8-record count this round used** - and this round wrote the entry. Two of those 8 tie CXXC1 directly to MLL, including a domain swap showing the CGBP CXXC domain cannot substitute for KMT2A's in a transforming fusion [E5]. **Counting is not reading, and an identical count supported opposite conclusions in two consecutive cycles.**

Go deeper
What was written (13 entries)
EntryBlockWhy it earned an entry
gene-supt16hFACTcounted absence + the CBL0137 p53 mechanism + the ENL thread
gene-ssrp1FACTthe LEDGF/p75 PWWP-to-HMG contact; two mapped steps from the fusion
protein-fact-complexFACTthe within-heterodimer disagreement on the selectivity statistic
gene-kansl1NSLNSL scaffold; the WDR5 mutual-exclusivity result
gene-kansl2NSLcompletes the KANSL1/WDR5/KANSL2 structure the partition claim rests on
protein-nsl-complexNSLfive NUP98 fusions binding MLL1 and NSL; complex-dependent mark specificity
gene-msl1MSLbest-mapped protein in the beat; supplies the missing scaffold for hyp-msl-route-to-h2bub-at-fusion-targets
gene-setd1bCOMPASSthe MLL-rearranged AML tiling result round 4 missed
gene-cxxc1COMPASSbinds MLL1 and recruits it to HoxA7; its CXXC domain CANNOT substitute in a fusion
gene-med12Mediatora named menin-associated protein, with the sign against a target hypothesis
gene-med23Mediatorrank 16 of 17,728, survives the lineage split, zero literature
hyp-wdr5-partition-nsl-versus-compasscross-cuttingthe one new falsifiable idea: NSL and COMPASS compete for WDR5
exp-wdr5-pool-exchange-kansl1-degroncross-cuttingthe degron experiment that decides it, with a cheap kill criterion

Plus this ledger and ana-interactome5-lineage-split.

What was rejected or deferred, with counts and reasons
CandidateRecords (title-checked where given)DepMapDecision
MED15 raw, 0 relevant (MLL OR KMT2A) AND (MED1 OR TRAP220 OR DRIP205 OR PPARBP OR PBP OR CRSP1 OR RB18A); the 5 are two LEDGF/p75 papers, an Mcl-1 inhibitor paper, an ovarian-cancer CIN analysis and a linked-read sequencing methodrank 81, MYELOID-IDENTITY ARTEFACT (myeloid z -0.86, B-lineage -0.13)rejected. Both halves of the case failed at once: no literature and no B-lineage dependency. The brief listed it as a priority-4 candidate; the data does not support it 2
MED140 (MLL OR KMT2A) AND (MED14 OR CRSP2 OR DRIP150 OR RGR1 OR EXLM1)rank 453 (-0.1541) on pan -1.1276rejected - pan-essential with no literature
MED241rank 128 (-0.2698), myeloid z -0.96 / B-lineage -0.42deferred - the count is one paper and the split is myeloid-weighted
KANSL30 (MLL OR KMT2A) AND (KANSL3 OR KIAA1310 OR NSL3)rank 13,630 (+0.0371)recorded inside protein-nsl-complex, no separate entry
MCRS11rank 10,971 (+0.0139)recorded inside protein-nsl-complex
PHF201rank 13,286 (+0.0339) on pan -0.0802recorded inside protein-nsl-complex - barely required anywhere
OGT3 (MLL OR KMT2A) AND (OGT OR "O-GlcNAc transferase")rank 1,634 (-0.0829) on pan -1.4669; the only negative difference in the NSL rosterdeferred with regret - a large independent O-GlcNAc literature exists that this beat could not read, and OGT is the only NSL-associated gene leaning the right way
MSL31 (MLL OR KMT2A) AND (MSL3 OR MSL3L1)rank 408 (-0.1615) on pan +0.0179; myeloid z -1.03 / B-lineage -0.40deferred with regret - the second-best-shaped number in the beat after MED23, on a gene required essentially nowhere. The myeloid-weighted split is exactly the artefact shape, which is why it was not written
KANSL1 as a target(entry written, target claim rejected)rank 15,913 (+0.0685)the entry exists; the target reading does not

Pattern. Round 4's rejections were dominated by anti-dependencies - seven chromatin repressors with positive differences. Round 5's rejections are dominated by a different failure: good genome-wide ranks that do not survive being split by lineage. MED1 at rank 81 and CXXC1 at rank 414 are both in the top 2.5% of the genome and neither has a B-lineage dependency worth the name. The instrument changed and so did the failure mode.

Four alias failures, of four different kinds, in one beat

The standing brief names two alias classes: retired symbols (a recall failure, the septin case) and short or generic aliases (a precision failure, the RD/U19 case). This beat hit both, plus two that the current wording does not cover.

1. Case and punctuation variants of a CURRENT symbol. (MLL OR KMT2A) AND (SETD1B OR KMT2G OR KIAA1076) returns 17. Adding Set1B returns 27 4. Same gene, same current symbol, different spelling, different query. Round 4 recorded the 17 and concluded no MLL-rearranged paper existed; the paper is in the 27.

2. Complex-name phrases are not reachable from subunit symbols. The NUP98 result - five fusions binding MLL1 and NSL, the cleanest cross-family extension in this graph - does not appear in any KANSL subunit query. It is reachable only through "NSL complex" and "non-specific lethal" 3. An enumeration keyed to gene symbols would have reported the NSL complex as having no leukaemia literature and been badly wrong.

3. An alias that collides with a DIFFERENT gene that shares the same localisation. KANSL1's synonym NSL1 is also the symbol of an unrelated kinetochore protein - and KANSL1 is itself annotated to the kinetochore. The collision cannot be resolved by localisation, only by reading titles; one of the four hits in the KANSL1 alias query is exactly that false positive 6.

4. A UniProt gene_exact search returning a different protein, caught live. gene_exact:MED1 AND organism_id:9606 AND reviewed:true returns Q15648 (MED1) and O95243 (MBD4), because MBD4 lists MED1 among its synonyms 7. This is the sixth instance of the wrong-gene-resolution failure in this project and the first caught by the assertion rather than by a length check. The standing rule works; it was applied; it fired.

And a fifth thing that is not an alias failure but belongs beside them: "FACT complex" is two ordinary English words. The query "FACT complex" AND (leukemia OR leukaemia) returns 9 records, two of which are coordination-chemistry papers where 'complex' means a metal complex. Corrected relevant count: 0 8.

Corrections reported to entries this beat does not own

Reported, not edited, per the standing rule, and with the citation that settles each.

1. ana-interactome4-enumeration, SETD1B rejection row. The row reads 'the strongest primary paper found is B-cell lymphoma (venetoclax resistance via BCL2-family proapoptotic genes), not KMT2A leukaemia'. A primary MLL-rearranged AML paper exists and predates that enumeration: Izumi et al., Leukemia 2025;39(7):1627-1639, PMID 40341256, DOI 10.1038/s41375-025-02638-y 4. Abstract read, full text not read. The dependency figures in the same row are correct and are reproduced unchanged in gene-setd1b.

2. ana-interactome4-enumeration, CXXC1 deferral row. The row reads 'deferred for want of a primary paper tying it to a KMT2A fusion' against a count of 8. The same query returns 8 now, and two of those 8 tie CXXC1 to MLL: PMID 18082152 (CGBP co-purifies with MLL1/MLL2/hSet1, binds the HoxA7 promoter, knockdown reduces MLL1 recruitment and H3K4me3) and PMID 23990460, which performs a CXXC domain swap in the context of the MLL-AF9 fusion 5. Both abstracts read, neither full text. This is not a retrieval error but a reading error, and it is the more instructive of the two.

Neither correction changes a number. Both change a conclusion.

Facts and figures
scriptnone - PubMed E-utilities esearch/esummary queries plus scripted reads of two existing analyses
statusrun
result summary42 queries, all decision-bearing records title-checked. 13 entries written, 9 candidates rejected or deferred with counts. Four alias failures of four distinct kinds, two of them classes the standing brief does not yet name. Two corrections reported against interactome round 4's enumeration ledger, neither changing a number and both changing a conclusion.
rejections
MED140.000MED14: 0KANSL30.000KANSL3: 0MED241.000MED24: 1MCRS11.000MCRS1: 1PHF201.000PHF20: 1MSL31.000MSL3: 1OGT3.000OGT: 3MED15.000MED1: 5
inputs
[
  "NCBI PubMed E-utilities",
  "analyses/depmap-kmt2a-selective-scan-full/results.json",
  "analyses/depmap-interactome5-lineage-split/results.json",
  "UniProt REST"
]
entries written
[
  "gene-supt16h",
  "gene-ssrp1",
  "protein-fact-complex",
  "gene-kansl1",
  "gene-kansl2",
  "protein-nsl-complex",
  "gene-msl1",
  "gene-setd1b",
  "gene-cxxc1",
  "gene-med12",
  "gene-med23",
  "hyp-wdr5-partition-nsl-versus-compass",
  "exp-wdr5-pool-exchange-kansl1-degron"
]
rejections — the numbers behind the chart
[
  {
    "candidate": "MED1",
    "records_raw": 5,
    "records_title_checked": 0,
    "depmap_rank": 81,
    "lineage_verdict": "MYELOID-IDENTITY ARTEFACT (myeloid z -0.86, B-lineage -0.13)",
    "decision": "rejected",
    "reason": "no relevant literature and no B-lineage dependency; a priority-4 candidate in the dispatch brief that the data does not support"
  },
  {
    "candidate": "MED14",
    "records_raw": 0,
    "depmap_rank": 453,
    "depmap_pan": -1.1276,
    "decision": "rejected",
    "reason": "pan-essential with zero literature"
  },
  {
    "candidate": "MED24",
    "records_raw": 1,
    "depmap_rank": 128,
    "lineage_verdict": "myeloid-weighted (z -0.96 vs -0.42)",
    "decision": "deferred",
    "reason": "one paper and a myeloid-weighted split"
  },
  {
    "candidate": "KANSL3",
    "records_raw": 0,
    "depmap_rank": 13630,
    "decision": "recorded in protein-nsl-complex, no separate entry"
  },
  {
    "candidate": "MCRS1",
    "records_raw": 1,
    "depmap_rank": 10971,
    "decision": "recorded in protein-nsl-complex, no separate entry"
  },
  {
    "candidate": "PHF20",
    "records_raw": 1,
    "depmap_rank": 13286,
    "depmap_pan": -0.0802,
    "decision": "recorded in protein-nsl-complex, no separate entry",
    "reason": "barely required anywhere"
  },
  {
    "candidate": "OGT",
    "records_raw": 3,
    "depmap_rank": 1634,
    "depmap_pan": -1.4669,
    "decision": "deferred with regret",
    "reason": "the only NSL-associated gene with a negative difference, and a large independent O-GlcNAc literature this beat could not read"
  },
  {
    "candidate": "MSL3",
    "records_raw": 1,
    "depmap_rank": 408,
    "depmap_pan": 0.0179,
    "lineage_verdict": "myeloid-weighted (z -1.03 vs -0.40)",
    "decision": "deferred with regret",
    "reason": "second-best-shaped number in the beat on a gene required essentially nowhere, but the split has the artefact shape"
  },
  {
    "candidate": "NPM1",
    "decision": "not re-scored",
    "reason": "the lineage-split run could not score it - too few lines (3 myeloid, 1 lymphoid, 7 comparator); the existing gene-npm1 entry was left untouched as instructed"
  }
]
alias failures
[
  {
    "kind": "case/punctuation variant of a CURRENT symbol",
    "example": "SETD1B versus Set1B",
    "counts": "17 versus 27",
    "consequence": "round 4 concluded no MLL-rearranged paper existed; the paper is in the larger set",
    "new_class": true
  },
  {
    "kind": "complex-name phrase unreachable from subunit symbols",
    "example": "'NSL complex' / 'non-specific lethal' versus KANSL1/2/3",
    "counts": "4 relevant versus 0 from any subunit query",
    "consequence": "the cross-family NUP98 result would have been missed entirely",
    "new_class": true
  },
  {
    "kind": "alias colliding with a different gene sharing the same localisation",
    "example": "KANSL1 synonym NSL1 versus the kinetochore NSL1; KANSL1 is itself kinetochore-annotated",
    "counts": "1 of 4 hits is the false positive",
    "consequence": "not resolvable by localisation or topic; title-checking is the only remedy",
    "new_class": false
  },
  {
    "kind": "UniProt gene_exact returning a different protein",
    "example": "gene_exact:MED1 returns Q15648 (MED1) and O95243 (MBD4)",
    "consequence": "sixth instance in this project; FIRST caught by the approvedSymbol assertion rather than by a length check",
    "new_class": false
  },
  {
    "kind": "ordinary-English-word query term",
    "example": "\"FACT complex\" AND (leukemia OR leukaemia)",
    "counts": "9 raw, 0 relevant; two are metal-coordination-chemistry papers",
    "consequence": "inflation on a phrase rather than a symbol",
    "new_class": false
  }
]
corrections reported
[
  {
    "target_entry": "ana-interactome4-enumeration",
    "row": "SETD1B",
    "wrong_text": "the strongest primary paper found is B-cell lymphoma (venetoclax resistance via BCL2-family proapoptotic genes), not KMT2A leukaemia",
    "should_say": "a primary MLL-rearranged AML paper exists: Izumi et al., Leukemia 2025;39(7):1627-1639, PMID 40341256, DOI 10.1038/s41375-025-02638-y",
    "read_level": "abstract only",
    "numbers_affected": "none - the dependency figures in that row are correct"
  },
  {
    "target_entry": "ana-interactome4-enumeration",
    "row": "CXXC1",
    "wrong_text": "deferred for want of a primary paper tying it to a KMT2A fusion",
    "should_say": "two of the same 8 records tie CXXC1 to MLL: PMID 18082152 (co-purification with MLL1/MLL2/hSet1, HoxA7 recruitment) and PMID 23990460 (CXXC domain swap in the MLL-AF9 fusion)",
    "read_level": "both abstracts read, neither full text",
    "numbers_affected": "none - this was a reading error, not a retrieval error"
  }
]
not done
[
  "permutation, lineage-conditioning and line-normalisation on any panel gene",
  "per-line inspection of MED23's four B-lineage values",
  "expression normalisation",
  "full-text reading of any cited paper"
]
How this could bite us back
  • Every count in this ledger is a floor on recall and an upper bound on relevance simultaneously, and both bounds were demonstrated loose within this beat. A counted absence here should be treated as a prompt to look harder, not as a settled fact.
  • Rejecting nine candidates on counts and unsplit ranks is exactly the process that rejected SETD1B in round 4 and had to be reversed in round 5. The rejections recorded here carry the same risk and should be re-read by the next round rather than trusted.
  • Two of the nine deferrals - OGT and MSL3 - were deferred for want of reading time rather than for want of evidence. That is an honest reason and a weak one, and a future round should start there.
Still open
  1. How many other entries in this graph rest on an enumeration that recorded an integer rather than a title-checked relevant set? The CXXC1 case shows the same count can support opposite conclusions, and the ledger format currently stores the integer.
  2. Should every enumeration run complex-name phrase queries alongside subunit symbols by default? On this beat that choice was the difference between finding and missing the most important paper.
  3. How many case-spelling variants are there for the genes already enumerated in this project? SETD1B/Set1B was found by accident, not by design.
  4. Would a standing requirement to run the three disciplining DepMap analyses have changed any conclusion in rounds 4 or 5? Both rounds skipped them and both recorded the omission.
Every source behind this entry (8)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Forty-two PubMed record-count queries were run through NCBI E-utilities esearch on 2026-09-19 for this beat. Key counts against '(MLL OR KMT2A)': SUPT16H/SPT16/FACTP140 1; SSRP1/T160/FACT80 1; the combined FACT query 2 of which 1 survives title-check; FACT versus AFF4/AFF1/ELL2/SEC/MLLT1/MLLT3 1 of which 0 survive; FACT versus DOT1L/H3K79 0; KANSL1/KIAA1267/MSL1v1/NSL1 4 of which 2 survive; 'NSL complex' OR 'non-specific lethal' 4 of which 4 survive; KANSL2/C12orf41/NSL2 1; KANSL3/KIAA1310/NSL3 0; KAT8/MOF/MYST1/hMOF 11; MCRS1 1; PHF20 1; OGT 3; MSL1/KIAA1585/hMSL1 0; MSL3/MSL3L1 1; SETD1B/KMT2G/KIAA1076/Set1B 27; CXXC1/CFP1/CGBP/PCCX1 8 of which 2 survive; MED1 and six aliases 5 of which 0 survive; MED12 and four aliases 4 of which 2 survive; MED14 and four aliases 0; MED23 and four aliases 0; MED24 and four aliases 1. Against leukaemia terms: 'FACT complex' 9 of which 0 survive; SUPT16H/SPT16 5 of which 1 survives; NSL complex/KANSL1-3 8 of which 1 survives; MSL1 1 of which 0 survive; CXXC1/CFP1 10; SETD1B/KMT2G 21; MED12 28 (not title-checked).

    What this evidence cannot show

    Counts are from the esearch `count` field. **Every record in every decision-bearing query was title-checked individually**, which is why most corrected counts are lower than the raw counts; the two exceptions are explicitly marked as raw. PubMed term mapping expands gene symbols to supplementary-concept records, so a query is not a literal string match and two apparently equivalent queries can differ. Every count is a **floor** on recall and an **upper bound** on relevance.

    Evidence that argues against this (1)
    • This beat demonstrated on itself that counts are unreliable in both directions: a case-spelling variant changed one count from 17 to 27, an identical count of 8 supported opposite conclusions in two cycles, a 9-record query contained two coordination-chemistry papers, and the single most important paper in the beat is unreachable from any subunit symbol. An integer count in this ledger should not be used without the title-checked number beside it. Xu H et al. 2016
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Every candidate in this beat was read out of the project's genome-wide KMT2A-rearranged selectivity scan of 17,728 genes and out of this beat's lineage-split run. Selected: MED23 rank 16 (-0.4827, pan -0.2096), MED1 rank 81 (-0.3114, pan -0.7573, lineage artefact), MED24 128 (-0.2698), MED12 131 (-0.268), MSL3 408 (-0.1615, pan +0.0179), CXXC1 414 (-0.1606, lineage artefact), MED14 453 (-0.1541, pan -1.1276), SETD1B 1,311 (-0.0931), OGT 1,634 (-0.0829), MSL1 2,078 (-0.0728), MSL2 7,112 (-0.0154), KAT8 9,683 (+0.0041), MCRS1 10,971 (+0.0139), PHF20 13,286 (+0.0339), KANSL3 13,630 (+0.0371), KANSL1 15,913 (+0.0685), SUPT16H 17,162 (+0.1165), WDR5 17,587 (+0.1984), KANSL2 17,631 (+0.2265), SSRP1 17,672 (+0.2651).

    Genome-wide KMT2A-rearranged selectivity scan, 17,728 genes (DepMap Chronos via Open Targets) · Broad Institute DepMap / Open Targets Platform; analyses/depmap-kmt2a-selective-scan-full 2026
    What this evidence cannot show

    All rows read out of analyses/depmap-kmt2a-selective-scan-full/results.json and analyses/depmap-interactome5-lineage-split/results.json on 2026-09-19 in scripted passes, so no number in this table was typed from recall. Eleven KMT2A-rearranged against 43 other leukaemia lines, AML-heavy, none infant-derived. **No gene was permutation-tested, lineage-conditioned or line-normalised in this cycle**, although scripts for all three exist - the same shortcut round 4 recorded against itself.

    Evidence that argues against this (1)
    • Within every complex written in this beat, subunits that cannot have independent requirements differ widely on this statistic: FACT's two obligate subunits by 0.1486 and 510 ranks, the three NSL scaffolds across 4,000 ranks, the three MSL subunits across ranks 408 to 7,112. That spread is an internal measurement of the statistic's noise and argues that differences below roughly 0.15 should not be interpreted anywhere in this graph. Broad Institute DepMap / Open Targets Platform; analyses/depmap-interactome5-lineage-split 2026
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Five NUP98 fusion proteins containing no KMT2A - NUP98-HOXA9, NUP98-HOXD13, NUP98-NSD1, NUP98-PHF23 and NUP98-TOP1 - physically interact with MLL1 and with the NSL complex; NUP98-HOXA9 and MLL1 co-localise at HOX gene promoters; and MLL1 is required for the proliferation of NUP98-HOXA9 cells in vitro and in vivo.

    NUP98 Fusion Proteins Interact with the NSL and MLL1 Complexes to Drive Leukemogenesis · Xu H, Valerio DG, Eisold ME, Sinha A, Koche RP, Hu W, Chen CW, Chu SH, Brien GL, Park CY, Hsieh JJ, Ernst P, Armstrong SA · Cancer Cell 30(6):863-878 2016
    PMID 27889185DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read, every identifier copied from that fetched record. **This paper is not returned by any KANSL subunit query run in this beat** and is reachable only through the complex-name phrases - the methodological point the ledger is making. Murine models and human expression signatures; no KMT2A fusion tested anywhere in it.

    Evidence that argues against this (1)
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A CRISPR-tiling screen against known H3K4 methylation modifiers in an MLL-rearranged AML model showed that disrupting the SETD1B catalytic SET domain depletes FLT3-ITD- or NrasG12D-expressing AML cells, downregulates the MYC pathway and decreases H3K4me3 breadth, with exogenous MYC or KDM5C disruption restoring growth. As of 2026-09-19 the query '(MLL OR KMT2A) AND (SETD1B OR KMT2G OR KIAA1076 OR Set1B)' returns 27 records; interactome round 4 recorded 17 for the same query without the spelling 'Set1B'.

    Regulation of H3K4me3 breadth and MYC expression by the SETD1B catalytic domain in MLL-rearranged leukemia · Izumi S, Ohtani K, Matsumoto M, Shibata S, Rahmutulla B, Fukuyo M, Nishimoto M, Miyagawa H, Sakaida E, Yokote K, Kitabayashi I, Araki K, Kaneda A, Hoshii T · Leukemia 39(7):1627-1639 2025
    PMID 40341256DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read, every identifier copied from that fetched record. The 17-record figure is quoted from interactome round 4's own recorded query string and was **not re-run in its original form in this session**, so the 17-versus-27 comparison is between this session's run and that entry's record. The mechanism by which the counts differ may partly be PubMed term mapping rather than a clean spelling effect; what is robust is that the paper is reachable from one query and was not recorded from the other.

    Evidence that argues against this (1)
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Replacing the MLL CXXC domain with the CGBP/CFP1 CXXC domain in the context of the leukemogenic MLL-AF9 fusion abrogated colony-forming ability and leukaemogenicity, while the DNMT1 CXXC domain supported leukaemogenesis comparable to MLL-AF9; all swapped fusions still targeted the Hoxa9 locus and the isolated CGBP CXXC domain still bound unmethylated DNA, with the discriminating property being protection of specific CpG residues from methylation, which prevented H3K9 trimethylation and enabled Hoxa9 expression.

    Functional specificity of CpG DNA-binding CXXC domains in mixed lineage leukemia · Risner LE, Kuntimaddi A, Lokken AA, Achille NJ, Birch NW, Schoenfelt K, Bushweller JH, Zeleznik-Le NJ · Journal of Biological Chemistry 288(41):29901-10 2013
    PMID 23990460DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read, identifiers copied from that fetched record. MLL-AF9 in murine primary bone marrow progenitors. **This paper was inside the same 8-record query that round 4 used to conclude no primary paper existed** - the reason this ledger argues that enumeration should record the title-checked relevant set rather than an integer.

    Evidence that argues against this (1)
    • A domain swap can fail for folding, spacing or expression reasons unrelated to the mechanism proposed, and the abstract does not describe expression-matching of the chimeras, so the CpG-protection interpretation is the authors' and not independently established here. NCBI PubMed, queried 2026-09-19 2026
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt lists NSL1 and MSL1V1 among KANSL1's synonyms, and annotates KANSL1 to the kinetochore (alias CENP-36) as well as the nucleus, mitochondrion and spindle pole; NSL1 is separately the symbol of an unrelated kinetochore protein, and one of the four records returned by '(MLL OR KMT2A) AND (KANSL1 OR KIAA1267 OR MSL1v1 OR NSL1)' is a kinetochore Blinkin interaction-mapping paper matched on that unrelated symbol.

    UniProtKB Q7Z3B3 (KANL1_HUMAN), KAT8 regulatory NSL complex subunit 1 · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
    What this evidence cannot show

    Record fetched 2026-09-19 via UniProt REST with gene_exact:KANSL1 AND organism_id:9606 AND reviewed:true; primary gene name read back as KANSL1 and length 1,105 confirmed from the same fetched record. The false-positive record was identified by title from an esummary fetch the same day. **Note the symmetric hazard: KANSL1's other synonym MSL1V1 means an unchecked MSL1 sweep can pull NSL papers into an MSL count.**

    Evidence that argues against this (1)
    • This alias collision is not resolvable by the usual heuristics - both genes are kinetochore-associated, so neither localisation nor topic filtering separates them. Only title-checking works, which makes it a worse case than the generic-alias class the standing rule already covers. NCBI PubMed, queried 2026-09-19 2026
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    A UniProt REST search for gene_exact:MED1 AND organism_id:9606 AND reviewed:true returns two reviewed entries: Q15648, whose primary gene name is MED1 (1,581 aa), and O95243, whose primary gene name is MBD4 (580 aa) and which lists MED1 among its synonyms.

    What this evidence cannot show

    Both records fetched 2026-09-19 in the same scripted pass, and the primary gene name of each was read back from the fetched record. This is the assertion step the standing accession rule prescribes, and it fired: a `gene_exact` search returned a different protein first-class in the result set. MED1 was subsequently rejected on other grounds, so the collision did not propagate into an entry.

    Evidence that argues against this (1)
    • Five previous wrong-gene resolutions in this project were caught by a length check or not at all. This one was caught by the assertion, which is evidence the rule works - but it also shows the rule is only as good as its application, since the same search pattern is used throughout this repository and a synonym match still looks like an exact match. NCBI PubMed, queried 2026-09-19 2026
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The curaxin CBL0137 exerts its effect by trapping FACT into chromatin, activating the p53 pathway and inducing an interferon response, and is active across 12 MLL-rearranged leukaemia lines, 3 MLL-r ALL patient-derived xenograft cell populations, an MV4;11 model and 5 MLL-r ALL xenograft models. The query '"FACT complex" AND (leukemia OR leukaemia)' returns 9 records of which 0 concern KMT2A on title-check, two being metal-coordination-chemistry papers where 'complex' has a different meaning.

    Potent antileukemic activity of curaxin CBL0137 against MLL-rearranged leukemia · Somers K, Kosciolek A, Bongers A, El-Ayoubi A, Karsa M, Mayoh C, Wadham C, Middlemiss S, Neznanov N, Kees UR, Lock RB, Gudkov A, Sutton R, Gurova K, Haber M, Norris MD, Henderson MJ · International Journal of Cancer 146(7):1902-1916 2020
    PMID 31325323DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read, identifiers copied from that fetched record. The mechanism is p53- and interferon-dependent rather than KMT2A-directed, and the abstract describes no non-MLL-r leukaemia comparator arm. Trapping is gain of occupancy, not inhibition.

    Evidence that argues against this (1)
Connected to