Oncofusion

start / graph / mark-h3k9me2

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q102-h3k9-rearrangement-brake

H3K9me2 - named sixty times in this graph with no entry of its own, the mark that defines the lamina-associated domains the fusion's targets sit beside, and the only H3K9 methylation state ever measured at a KMT2A fusion's own binding sites

also called H3K9me2 · H3K9 dimethylation · histone H3 lysine 9 dimethylation · H3K9Me2 · dimethyl-H3K9 · H3K9me1/2 (the compact form used in most abstracts) · H3K9me2/3 (the compact form used when grouped upward)

H3K9me2 is the two-methyl silencing tag that covers huge blocks of the genome, especially the parts pinned to the edge of the nucleus. It is the tag a DOT1L drug pushes onto the leukaemia genes, the tag that floods the KMT2A gene's fragile stretch when its eraser is removed, and it has never once been measured in a patient's infant leukaemia cells.

The same thing, for a scientist

H3K9me2 is the di-methyl state of histone H3 Lys-9: written in euchromatin by the obligate G9a/EHMT2-EHMT1/GLP heterodimer [E1], erased by the KDM3/JMJD1 family with a *preference* for the di- over the mono-methyl substrate [E2], and read by HP1 proteins and MPP8 [E1]. It is the mark that operationally defines **lamina-associated domains**, which is how it already appears sixty times in this corpus without a node - `localization-nuclear-lamina-lads-kmt2a-r` and `ana-lad-border-overlap-kmt2a-targets` are built on it. **Three KMT2A-relevant measurements exist and they point in three directions.** (1) In murine MLL-AF9 leukaemia, DOT1L inhibition *increases* H3K9me2 specifically around the transcription start sites of MLL-AF9-bound genes, via SIRT1 and SUV39H1 - the only targeted H3K9-methylation result at fusion-bound loci in any system, and it is me2, not me3 [E3]. (2) In near-diploid human RPE cells, KDM3B depletion *increases* H3K9me2 on the flank of the KMT2A breakpoint cluster region, lowers CTCF occupancy inside it, and promotes KMT2A copy gains and rearrangements, all of which G9a co-depletion reverses [E4]. (3) In AML cells, KDM3B knockdown *decreases* H3K9me2 at the HOXA1 retinoic-acid response element while raising me1 there [E5] - the opposite sign from (2) for the same perturbation. **And it has never been measured in the index fusion:** `H3K9me2[tiab] AND KMT2A[tiab]` returns 0 and `H3K9me2[tiab] AND ("MLL-AF4" OR "KMT2A::AFF1" OR "t(4;11)")[tiab]` returns 0, both twice, both genuine [E6].

Go deeper
Where this mark already lives in this graph, unnamed

H3K9me2 is named in eleven entries and 60 times, and the places it does real work are worth naming because a reader of those entries cannot currently click through to the mark:

entrywhat it uses H3K9me2 for
localization-nuclear-lamina-lads-kmt2a-rH3K9me2 as the defining mark of lamina-associated domains, the compartment the fusion's targets border
ana-lad-border-overlap-kmt2a-targetsthe LAD-border overlap analysis rests on the same definition
mark-h3k9me3records, in its own E5 and its counter-evidence, that the targeted H3K9-methylation result in this disease is me2 and not me3
gene-ehmt2G9a as the H3K9me1/2 writer, both upstream at the locus and downstream in established AML
gene-rest, gene-brd3, gene-med19, gene-ebna1bp2H3K9me2 as the repressive output of REST-CDYL-G9a and related silencing routes
localization-transcription-factories-pol2-clustersH3K9me2 as the state the fusion's active compartment is defined against

The reason this matters and is not bookkeeping: mark-h3k9me3's central open question is 'which H3K9 methylation state actually matters at fusion targets - me2 or me3?', and it answers that the targeted murine result is me2 while the untargeted human KMT2A::AFF1 result is me3. That question is about this mark and had no entry to be asked in.

What distinguishes me2 from me1 and from me3, with the cases where they part company

Enzymology. The same heterodimer writes me1 and me2 in euchromatin, and UniProt annotates EHMT1/GLP - but not EHMT2/G9a - as also writing me3 1. The KDM3 family erases me1 and me2 with a stated preference for me2 and has weak or no activity on me3 2. The KDM4/JMJD2 family erases me2 and me3. So me2 is the only H3K9 methylation state reachable by both eraser families, which makes it the state most likely to be buffered when either one is inhibited, and the state whose level is hardest to attribute to a single enzyme.

Distribution. me2 occupies large domains rather than sharp peaks, and its most consistent genomic correlate is the lamina-associated domain. me1 sits in euchromatic gene bodies; me3 sits in constitutive heterochromatin. At the KMT2A locus the three are not interchangeable: me1 filled the whole breakpoint cluster region on KDM3B loss while me2 rose only on the flank 4.

Two documented cases where me1 and me2 move in opposite directions from one perturbation. In RPE cells, KDM3B depletion raised both, but with different geometry 4. In AML cells, KDM3B knockdown raised me1 and lowered me2 at the HOXA1 RARE 5. Neither literature cites the other. Any entry, including this one, that writes 'H3K9me1/2' as a single quantity is compressing a disagreement.

And one case where me2 and me3 part company in this disease. In murine MLL-AF9, DOT1L inhibition raised H3K9me2 specifically at fusion-bound TSSs 3; in human KMT2A::AFF1 SEM cells the same class of inhibitor raised H3K9me3 generally, at enhancers, non-enhancers and controls alike, and the authors concluded it did not explain the locus-specific effects (mark-h3k9me3 E4). Same drug class, two marks, two geometries, and the difference tracks the mark.

The identifier traps, both measured

Trap 1 - the compact slashed form is a one-way street. Measured 2026-09-22 on live esearch with querytranslations read back 6:

queryrecords
H3K9me2[tiab]1,546
"H3K9me1/2"[tiab]67, of which 52 (77.6%) are not in the me2 set
"H3K9me2/3"[tiab]121
H3K9me2[tiab] AND KMT2A[tiab]0, twice, genuine

PubMed tokenises H3K9me1/2 so that H3K9me1 matches inside it and H3K9me2 does not. A paper that writes only the compact form is therefore retrievable as the lower state and not as the upper one. This is what makes the KMT2A zero above simultaneously true and worthless: PMID 37788669's whole subject is H3K9me2 at the KMT2A locus. The upward form H3K9me2/3 has the same shape one rung higher and will hide me3 papers from an me3 query.

Trap 2 - ALL cannot be an operand. H3K9me2[tiab] AND ("acute lymphoblastic"[tiab] OR ALL[tiab]) returns 3, and the response's errorlist reads phrasesnotfound: ["ALL"]: PubMed silently dropped the disease abbreviation that matters most to this project and ran only the spelled-out form 6. This graph has recorded ALL misread as a gene symbol before; here it is dropped as a stopword instead, which is quieter.

So the honest bound is: three records mention H3K9me2 and acute lymphoblastic leukaemia in a title or abstract, one of which is the KDM3B/LMO2 paper 7, and the intersection with any KMT2A fusion is empty.

What is known in a KMT2A-rearranged context, and what closes each gap
questionstateclasswhat closes it
H3K9me2 at fusion-bound loci in a KMT2A fusionmeasured - murine MLL-AF9, targeted increase at fusion-bound TSSs under DOT1L inhibition 3-done
H3K9me2 genome-wide in a KMT2A-fusion celldeposited - GSE61021/GSE61022, murine MLL-AF9 8measured-not-analysedcompute
H3K9me2 at the KMT2A locus in any haematopoietic cellnot measured - the only BCR data are in RPE 4measured-not-analysed for the KDM3-family arm: GSE75577/GSE75581 hold murine Jmjd1c-loss H3K9me2 genome-wide, so the Kmt2a-locus question can be asked of public data 8compute
H3K9me2 in human KMT2A::AFF1, any locusnever - two independent zeros 6never-measuredwet work
H3K9me2 in B-lineage or infant materialnever 6never-measuredwet work

The single cheapest measurement in this table is the third row: ask whether Kmt2a-locus H3K9me2 rises in the deposited murine Jmjd1c-loss data. It is the same enzyme family, in a haematopoietic system, in public data, and it would tell you whether the RPE result transfers to blood before anyone runs a cell.

Facts and figures
markH3K9me2
histoneH3
residueK9, N-terminal tail
chemistrydi-methylation; mutually exclusive with H3K9ac on the same residue; the only H3K9 methylation state reachable by both the KDM3 (from me2 down) and KDM4 (from me3 down) eraser families
genomic distributionlarge domains rather than sharp peaks; the operational mark of LAMINA-ASSOCIATED DOMAINS, which is how it already underlies localization-nuclear-lamina-lads-kmt2a-r and ana-lad-border-overlap-kmt2a-targets in this graph. At the KMT2A locus it rises on the FLANK of the breakpoint cluster region on KDM3B loss, not uniformly across it
distinguished from me1same writers, different geometry at the KMT2A locus (flank vs whole BCR) and OPPOSITE SIGN at the HOXA1 RARE in AML cells - the two states are not one quantity
distinguished from me3different writers (SUV39H1/2, SETDB1 for me3) and a different eraser family; in this disease the TARGETED H3K9-methylation result at fusion-bound TSSs is me2 (murine MLL-AF9) while the UNTARGETED result in human KMT2A::AFF1 is me3
measured in kmt2a rtrue
measured in kmt2a r wheremurine MLL-AF9, at MLL-AF9-bound transcription start sites under DOT1L inhibition (PMID 25822366); GSE61021/GSE61022
measured in kmt2a aff1false
measured in b lineagefalse
measured in infant materialfalse
measured at the kmt2a locusin RPE cells only (PMID 37788669) - never in a haematopoietic cell
shown for fusionYES in murine MLL-AF9 myeloid leukaemia, at fusion-bound TSSs, under DOT1L inhibition (PMID 25822366); NOTHING in KMT2A::AFF1, nothing in B-lineage, nothing in infant material, and nothing at the KMT2A locus in any blood cell
direction in diseaseAMBIGUOUS AND STAGE-DEPENDENT BY DESIGN OF THE EVIDENCE - raising it at the KMT2A locus promotes rearrangement in a normal epithelial cell (PMID 37788669); raising it at fusion-bound promoters silences the leukaemic programme in established murine MLL-AF9 (PMID 25822366). Same mark, opposite therapeutic sign, different DNA and different disease stage
epigenetic readoutH3K9me2 ChIP-seq (Abcam ab1220) genome-wide and at the KMT2A BCR flank in RPE; H3K9me2 ChIP at MLL-AF9-bound TSSs in murine myeloid leukaemia; deposited in GSE210480, GSE61021/GSE61022, GSE75577/GSE75581
absence classmeasured-not-analysed
writers
[
  "EHMT2/G9a (Q96KQ7, 1,210 aa) - entry: gene-ehmt2",
  "EHMT1/GLP (Q9H9B1, 1,298 aa) - obligate heterodimer partner, additionally annotated as an me3 writer; NO entry in this graph as of 2026-09-22",
  "SETDB1 (Q15047, 1,291 aa) - no entry in this graph; note its canonical UniProt FUNCTION comment reads 'Lacks all domains required for histone methyltransferase activity', an isoform-scoped statement that must not be read as a fact about the protein"
]
erasers
[
  "KDM3B (Q7LBC6) - entry: gene-kdm3b",
  "KDM3A (Q9Y4C1) - PREFERS the dimethyl substrate; no entry in this graph",
  "JMJD1C/KDM3C (Q15652) - entry: gene-kdm3c",
  "KDM4 family (me2 and me3) - gene-kdm4c is in this graph"
]
readers
[
  "CBX5/HP1alpha (P45973, 191 aa), CBX1, CBX3 - NO entries in this graph, a gap mark-h3k9me3 also records",
  "MPHOSPH8/MPP8 (Q99549, 860 aa), HUSH complex - no entry",
  "EHMT2/G9a ANK repeats bind H3K9me1 and H3K9me2 (gene-ehmt2 E1)"
]
antibodies rule 74
{
  "H3K9me2": "Abcam ab1220 in BOTH load-bearing results - PMID 37788669 (RPE ChIP-seq) and PMID 25822366 (murine MLL-AF9 ChIP). The two independent me2 results in this beat share one reagent and would share any epitope artefact"
}
identifier traps
{
  "compact_slashed_form": {
    "trap": "'H3K9me1/2' matches H3K9me1[tiab] and NOT H3K9me2[tiab]",
    "slash_records": 67,
    "unreachable_by_me2_query": 52,
    "pct": 77.6,
    "worked_consequence": "H3K9me2[tiab] AND KMT2A[tiab] = 0 for a paper whose subject is H3K9me2 at the KMT2A locus"
  },
  "upward_slashed_form": {
    "\"H3K9me2/3\"[tiab]": 121,
    "implication": "the same trap one rung up will hide me3 papers from an me3 query"
  },
  "ALL_is_a_stopword": {
    "query": "H3K9me2[tiab] AND (\"acute lymphoblastic\"[tiab] OR ALL[tiab])",
    "count": 3,
    "errorlist": "phrasesnotfound: [\"ALL\"]",
    "reading": "PubMed dropped the abbreviation silently and ran only the spelled-out form; a bounded absence about ALL written with that operand is bounded by 'acute lymphoblastic' alone"
  },
  "registry_rows_existed_before": false
}
corpus position when written
{
  "entries_naming_h3k9me2": 11,
  "mentions": 60,
  "entries_that_do_real_work_with_it": [
    "localization-nuclear-lamina-lads-kmt2a-r",
    "ana-lad-border-overlap-kmt2a-targets",
    "mark-h3k9me3",
    "gene-ehmt2",
    "gene-rest",
    "localization-transcription-factories-pol2-clusters"
  ],
  "count_method": "grep -rlo / grep -ro over data/graph/nodes/**/*.json, run 2026-09-22 before this entry was written"
}
How this could bite us back
  • This mark carries two opposite therapeutic prescriptions in one disease. Raising it at fusion-bound promoters is the proposed mechanism by which DOT1L inhibition silences the leukaemic programme; raising it at the KMT2A locus is what makes that locus rearrange. A combination that lowers H3K9me2 to protect the locus would be antagonistic to the DOT1L rationale, and a combination that raises it would be antagonistic to locus stability.
  • The G9a inhibitor that lowers this mark is a DUAL G9a/GLP inhibitor, and GLP/EHMT1 is annotated as an H3K9me3 writer as well. So 'inhibit the me1/me2 writer' is not a clean intervention on me1/me2, and the one published blot in a haematopoietic cell shows a G9a/GLP inhibitor lowering H3K9me3 too.
  • H3K9me2 is the state with two eraser families, which means a single-agent eraser inhibitor is the intervention most likely to be silently buffered - and the buffering would look like drug failure rather than like redundancy.
  • Nothing in this entry has been measured in a human B-lineage cell, in infant material, or in KMT2A::AFF1. Two independent PubMed zeros bound that, and the archive check does not rescue it: every deposited dataset here is mouse myeloid or human epithelium.
Still open
  1. Does Kmt2a-locus H3K9me2 rise when a KDM3 family member is lost in a haematopoietic cell? GSE75577/GSE75581 hold murine Jmjd1c-loss H3K9me2 genome-wide data; this is the cheapest test of whether the RPE result transfers to blood and it needs no wet work.
  2. Which H3K9 methylation state actually matters at fusion-bound loci - me2 or me3? mark-h3k9me3 asks this and could not answer it; the answer now has two halves, because the targeted murine result is me2 at fusion TSSs and the locus-genesis result is me1/me2 at the KMT2A gene, and no experiment has measured both in one system.
  3. Is the KMT2A locus inside or beside a lamina-associated domain in a fetal B-lineage progenitor? H3K9me2 defines LADs, this graph already has a LAD-border analysis for the fusion's targets, and nobody has asked the same question about the KMT2A gene itself - which would connect the compartment story to the fragility story.
  4. Does the shared use of Abcam ab1220 in both load-bearing H3K9me2 results in this beat matter? Two independent papers, two cell types, one catalogue number - an orthogonal antibody or a mass-spectrometry readout would make the agreement between them evidential rather than reagent-bound.
Every source behind this entry (8)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    H3K9me2 is written in euchromatin by EHMT2/G9a (Q96KQ7, 1,210 aa) and by its heterodimer partner EHMT1/GLP (Q9H9B1, 1,298 aa), which is additionally annotated as writing H3K9me3; H3K9 methylation is described as a specific tag for epigenetic transcriptional repression by recruiting HP1 proteins, CBX5/HP1alpha (P45973, 191 aa) binds H3K9-methylated H3 tails and may tether heterochromatin to the inner nuclear membrane through the lamin-B receptor, and MPHOSPH8/MPP8 (Q99549, 860 aa) recognises H3K9me as the HUSH complex's reader.

    What this evidence cannot show

    Database records, not measurements. The EHMT1 product-specificity statement including me3 rests on PubMed:12004135, which was not retrieved here - rule 91 applies, and the EHMT1-writes-me3 / EHMT2-does-not asymmetry may be a curation artefact rather than a biochemical distinction. CBX5's reader statement is annotated for 'H3K9me' without a methylation state, so using it for me2 specifically is an inference from the family literature, not from the record. Note also that SETDB1's canonical UniProt FUNCTION comment reads 'Lacks all domains required for histone methyltransferase activity' - an isoform-scoped statement that would be badly misread as a fact about SETDB1, and the reason no SETDB1 claim is made here.

    Evidence that argues against this (1)
    • The reader arm of this mark is entirely unwritten in this graph and therefore untested against it: there is no entry for CBX1, CBX3, CBX5 or MPHOSPH8, a gap `mark-h3k9me3` already records for the me3 side. An annotated reader with no entry and no fusion-context experiment is a lead, not a mechanism. UniProt Consortium 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    KDM3A, the closest paralogue of KDM3B, preferentially demethylates mono- and dimethylated H3 Lys-9 with a preference for the DIMETHYLATED residue and has weak or no activity on trimethylated H3 Lys-9 - making H3K9me2 the state acted on by both the KDM3 family (from below) and the KDM4 family (from above), and therefore the only H3K9 methylation state with two independent eraser routes.

    What this evidence cannot show

    A database record about the paralogue, used for a family-level substrate rule; KDM3B's own record is less specific (see `gene-kdm3b` E1). The 'two eraser routes' clause is this entry's inference from combining the KDM3 and KDM4 substrate preferences, not a statement in any single source, and it has no experimental support in a KMT2A-rearranged cell - `gene-kdm4c` is in this graph and no KDM4 experiment there reports H3K9me2 at fusion-bound loci.

    Evidence that argues against this (1)
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In murine MLL-AF9 leukaemia cells, DOT1L inhibition increases H3K9 DIMETHYLATION (and H3K27 trimethylation) around the transcription start sites of MLL-AF9-bound genes, and Suv39h1 is required for the Sirt1-dependent H3K9 methylation and silencing of the MLL-AF9 leukaemic programme - making this the only targeted, fusion-occupancy-anchored H3K9 methylation measurement in any KMT2A-rearranged system, and it is in the di-methyl state.

    DOT1L inhibits SIRT1-mediated epigenetic silencing to maintain leukemic gene expression in MLL-rearranged leukemia. · Chen CW; Koche RP; Sinha AU; Deshpande AJ; Zhu N; Eng R; et al. · Nat Med 2015
    PMID 25822366DOI
    What this evidence cannot show

    Murine, myeloid, MLL-AF9 - not human, not B-lineage, not KMT2A::AFF1. The H3K9me2 antibody is Abcam ab1220, which is also the H3K9me2 ChIP antibody in PMID 37788669, so the two load-bearing me2 results in this beat share a reagent and would share any epitope artefact (rule 74). This entry does not re-read the paper: the detail is taken from `mark-h3k9me3` E5, which records it as read in full text (PMC4390532), and the human lines in that study were assayed for H3K9ac rather than H3K9 methylation.

    Evidence that argues against this (1)
    • In human KMT2A::AFF1 SEM cells the same inhibitor class produced only a modest and explicitly GENERAL increase in H3K9 methylation - at H3K79me2/3-marked enhancers, non-enhancer enhancers and a non-enhancer control alike - and the authors concluded it did not explain the locus-specific changes. The two results differ in species, lineage, methylation state and in whether the effect is targeted, and nobody has reconciled them. Godfrey L; Crump NT; Thorne R; Lau IJ; Repapi E; Dimou D; et al. 2019
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In immortalised near-diploid human RPE cells, KDM3B depletion produces a genome-wide increase of H3K9me2 skewed towards higher levels, with the increase at the KMT2A gene falling on the FLANK of the 8.3 kb breakpoint cluster region rather than uniformly across it as H3K9me1 does; double knockdown of KDM3B and G9a rescued the H3K9me2 increase across 75% of affected regions genome-wide (219 Mb of 292 Mb) and the H3K9me1 increase across 81% (88 Mb of 109 Mb), and G9a overexpression alone was sufficient to promote KMT2A copy gains and rearrangements.

    Epigenetic balance ensures mechanistic control of MLL amplification and rearrangement. · Gray ZH; Chakraborty D; Duttweiler RR; Alekbaeva GD; Murphy SE; Chetal K; Ji F; Ferman BI; Honer MA; Wang Z; et al. · Cell 2023
    PMID 37788669DOI
    What this evidence cannot show

    **Full text read 2026-09-22 (efetch db=pmc, PMC10591855, 80,638 stripped characters, <body> confirmed; Europe PMC fullTextXML returned HTTP 500).** The system is RPE, an epithelial line chosen for genomic stability - not haematopoietic, not lymphoid, not leukaemic. The megabase rescue percentages are genome-wide aggregates over 10 kb bins and say nothing about the KMT2A locus specifically; the locus-level statements are ChIP-seq and ChIP-qPCR at named regions. No spike-in normalisation is described, which is a sharper limitation for a mark reported to change genome-wide than for a locus-specific one. H3K9me2 was NOT measured in any of the paper's haematopoietic arms.

    Evidence that argues against this (1)
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In AML cells, KDM3B knockdown resulted in increased H3K9 mono-methylation but DECREASED H3K9 di-methylation at the retinoic-acid response elements of HOXA1 while eschewing the HOXA1 promoter region; KDM3B is underexpressed in AML patients and in AML cell lines carrying MLL-AF6, MLL-AF9 or PML-RARA translocations, and KDM3B overexpression repressed colony formation of an AML line with 5q deletion.

    KDM3B shows tumor-suppressive activity and transcriptionally regulates HOXA1 through retinoic acid response elements in acute myeloid leukemia. · Xu X; Nagel S; Quentmeier H; Wang Z; Rosenwald A; Meyer C; Drexler HG; MacLeod RAF · Leuk Lymphoma 2018
    PMID 28540746DOI
    What this evidence cannot show

    **Abstract only; full text not retrieved**, so the ChIP antibodies (rule 74), element coordinates, replicate number and effect sizes behind the opposite-sign me1/me2 result are uncharacterised - and that is the load-bearing part of this item. Myeloid lines only; no B-lineage arm and no KMT2A::AFF1. The KMT2A-fusion relevance is expression-level (MLL-AF6 and MLL-AF9 lines underexpress KDM3B), not a chromatin measurement in a fusion cell. The causal direction is tested by overexpression in a 5q-deleted line, so the physiological contribution of endogenous KDM3B at HOXA1 is inferred.

    Evidence that argues against this (1)
    • The same first author's later work uses the opposite framing: small-molecule KDM3-family modulators are proposed as a treatment that preferentially kills MLL-rearranged acute leukaemia cells, which requires KDM3-family activity to be pro-leukaemic in exactly the disease where this paper reports KDM3B to be underexpressed and tumour-suppressive. Xu X; Wang L; Hu L; Dirks WG; Zhao Y; Wei Z; Chen D; Li Z; Wang Z; Han Y; et al. 2020
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    As of 2026-09-22 live PubMed esearch returns 1,546 records for H3K9me2[tiab], 181 for "H3K9 dimethylation"[tiab], 121 for "H3K9me2/3"[tiab] and 67 for "H3K9me1/2"[tiab] of which 52 are unreachable by H3K9me2[tiab]; H3K9me2[tiab] AND KMT2A[tiab] returns 0 and H3K9me2[tiab] AND ("MLL-AF4" OR "KMT2A::AFF1" OR "t(4;11)")[tiab] returns 0, both with a non-empty querytranslation and warninglist ["No items found."] and both run twice; H3K9me2[tiab] AND MLL[tiab] returns 2 records, neither about a KMT2A fusion; and H3K9me2[tiab] AND ("acute lymphoblastic"[tiab] OR ALL[tiab]) returns 3 with errorlist phrasesnotfound ["ALL"], PubMed having dropped the abbreviation entirely.

    What this evidence cannot show

    **Titles and abstracts only, and in this beat that bound is demonstrably loose:** the founding paper itself ran H3K9me3 ChIP-seq that appears in no figure, and this project has previously recovered a KMT2A::AFF1 H3K9me3 measurement from a supplementary figure invisible to its own bounded query. Every cytogenetic term is quoted per rule 95 and no querytranslation contained a [UID] term. The grouped `(A OR B)[tiab]` form appears only in the fusion-name OR list; that form is broader than labelled, so the zeros are conservative. Name list tried: H3K9me2, "H3K9 dimethylation", "H3K9me1/2", "H3K9me2/3", "H3 lysine 9", H3K9. `ALL` cannot be used as an operand at all.

    Evidence that argues against this (1)
    • The literature zero is not a data zero. GEO returns 18 series for H3K9me2 AND (KMT2A OR MLL) AND leukemia, including deposited H3K9me2 ChIP-seq in murine MLL-AF9 cells and in murine Jmjd1c-loss marrow - so the honest statement is not 'unmeasured' but 'measured in mouse myeloid systems, never in human B-lineage, and never analysed at the KMT2A locus'. NCBI Gene Expression Omnibus 2026
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    KDM3B is an H3K9me1/2 demethylase that induces expression of the leukaemogenic oncogene LMO2 through synergy with CBP, represses leukaemia-cell differentiation, and is UP-regulated in blood cells from ALL-type leukaemia patients; the study mapped genome-wide H3K9me2 occupancy by ChIP-chip during ATRA-induced differentiation of HL-60 promyelocytic leukaemia cells and found KDM3B down-regulated during differentiation through corepressor recruitment.

    KDM3B is the H3K9 demethylase involved in transcriptional activation of lmo2 in leukemia. · Kim JY; Kim KB; Eom GH; Choe N; Kee HJ; Son HJ; Oh ST; Kim DW; Pak JH; Baek HJ; et al. · Mol Cell Biol 2012
    PMID 22615488DOI
    What this evidence cannot show

    **Abstract only; full text not retrieved.** ChIP-chip rather than ChIP-seq, in HL-60 - which is one of the two KDM3B-LOH lines used in PMID 37788669, so the baseline KDM3B level in this system is already abnormal. 'Up-regulated in blood cells from ALL-type leukaemia patients' has no cohort size, assay or KMT2A status attached in the abstract, so it cannot be read as a statement about KMT2A-rearranged ALL specifically; it is recorded because it is the only lymphoid-lineage KDM3B direction in the literature and it is opposite to the myeloid one.

    Evidence that argues against this (1)
  8. E8established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    Deposited H3K9me2 data exist in KMT2A-fusion and KDM3-family systems: GEO returns 18 series for H3K9me2 AND (KMT2A OR MLL) AND leukemia, among them GSE61021/GSE61022 (murine MLL-AF9, the DOT1L/SIRT1 study), GSE75577 and GSE75581 (murine Jmjd1c loss, one titled 'Effect of loss of Jmjd1c on H3K9me2 level across the genome'), and GSE210480 (60 human RPE samples, PMID 37788669's own deposit, including H3K9me1, H3K9me2, H3K9me3, CTCF and Input across control, siKDM3B, siG9A and double-knockdown arms).

    What this evidence cannot show

    A GEO hit list is not a usability assessment: two of the 18 series are C. elegans, several match on 'MLL' in free text rather than on a KMT2A fusion, and this project has previously found a correctly named, well-formed, empty supplementary archive and a SuperSeries that denied having raw data in three different ways. The Jmjd1c series is murine and is about a different KDM3 family member than the one PMID 37788669 depletes - it is offered as the cheapest available proxy for 'does a KDM3-family loss raise H3K9me2 at the Kmt2a locus in blood', not as the same experiment.

    Evidence that argues against this (1)
    • A deposited genome-wide dataset can be the wrong instrument for a locus question: the KMT2A breakpoint cluster region is 8.3 kb, the analyses in the founding paper were binned at 10 kb, and a single 10 kb bin cannot separate the BCR from its flank - which is precisely the distinction that separates the me1 result from the me2 result. Gray ZH; Chakraborty D; Duttweiler RR; Alekbaeva GD; Murphy SE; et al. 2023
Connected to