Oncofusion

start / graph / mark-h3k9me1

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

H3K9me1 - the only H3K9 state whose level at the KMT2A breakpoint cluster region has ever been measured, and it is LOW there until KDM3B is removed

also called H3K9me1 · H3K9 monomethylation · histone H3 lysine 9 monomethylation · H3K9Me1 · monomethyl-H3K9 · H3K9me1/2 (the compact form used in most abstracts)

H3K9me1 is the first, single methyl tag on histone H3 at lysine 9 - the faintest version of the silencing mark. In healthy cells the stretch of KMT2A where this leukaemia's break happens carries unusually little of it, and when the enzyme that keeps it clear is removed the tag floods in and the gene starts breaking.

The same thing, for a scientist

H3K9me1 is the mono-methyl state of histone H3 Lys-9: the product of G9a/EHMT2 and EHMT1/GLP in euchromatin [E1], and, in a compartment-separated route the field rarely states, of MECOM/PRDM3 and PRDM16 acting on **free histone H3 in the cytoplasm** before deposition, where SUV39H enzymes then use it as the substrate for H3K9me3 [E2]. It is erased by the KDM3/JMJD1 family - KDM3A, KDM3B and JMJD1C - which act on me1 and me2 and not on me3 [E3, E4]. **Its importance to this project is specific and recent.** In near-diploid human RPE cells, H3K9me1 across the 8.3 kb KMT2A breakpoint cluster region (exons 8-14 in the source's numbering) is *lower* than in the flanking sequence at baseline; KDM3B depletion raises it strongly and uniformly across the whole BCR, reduces CTCF occupancy at the conserved CTCF site in exon 11, and promotes site-specific KMT2A copy gains and break-apart events - and the H3K9me1 rise at that one CTCF site is among the strongest of any CTCF site in the genome [E5]. Co-depleting G9a resets it and prevents the alterations [E5]. **So H3K9me1 is currently the best-evidenced chromatin state gating whether the KMT2A locus rearranges at all**, and it is the one H3K9 state with no prior entry in this graph despite being named in four entries. Nobody has measured it in a KMT2A-rearranged cell, in a B-lineage cell, or in infant material: `H3K9me1[tiab] AND ("MLL-AF4" OR "t(4;11)" OR "MLL-AF9")[tiab]` returns 0 twice over [E6].

Go deeper
Why me1 is not just 'less me2', and why that matters here

Three properties separate H3K9me1 from H3K9me2 and H3K9me3, and each one is load-bearing for this beat.

H3K9me1H3K9me2H3K9me3
written byG9a/EHMT2, EHMT1/GLP 1; MECOM/PRDM3 + PRDM16 on free H3 in the cytoplasm 2G9a/EHMT2, EHMT1/GLP 1SUV39H1/2, SETDB1, EHMT1/GLP 1
erased byKDM3A/KDM3B/JMJD1C (me1 + me2) 34KDM3 family, KDM4 familyKDM4 family; not KDM3 3
distributioneuchromatic gene bodies; a substrate pool feeding me2/me3 2large domains, LADs, facultative repressionconstitutive heterochromatin, pericentromere
measured at the KMT2A BCRYES - and it is depleted there at baseline 5yes, at the BCR flank only 5ChIP-seq run and never reported (see absence-...-h3k9me3-at-the-kmt2a-locus...)

First, me1 is a precursor, not only a mark. MECOM/PRDM3 and PRDM16 monomethylate free H3 in the cytoplasm; the monomethylated histone is imported, incorporated, and then used by SUV39H enzymes as the substrate for trimethylation 2. A perturbation that changes the me1 pool therefore changes the me3 ceiling, which is why 'raise me1/me2 and me3 will follow' is a real possibility and not a rhetorical hedge.

Second, the enzymes are not symmetric. G9a/EHMT2 is annotated as an me1/me2 writer and not an me3 writer [E1, and gene-ehmt2 E1]; its obligate heterodimer partner EHMT1/GLP is annotated as writing me1, me2 and me3 1. Both facts come from the same UniProt curation of the same 2002 paper. That single asymmetry is the reason a 'G9a inhibitor' cannot be assumed me3-sparing (see drug-unc0642).

Third, at the KMT2A BCR the two states move differently. In RPE cells, KDM3B depletion raised H3K9me1 uniformly across the whole BCR while H3K9me2 rose on the flank of the BCR 5. And in AML cells the same perturbation moved them in opposite directions at one element: KDM3B knockdown raised H3K9 mono-methylation and lowered di-methylation at the HOXA1 retinoic-acid response element 7. Any model that treats 'H3K9me1/2' as a single quantity is unsupported by both of the locus-resolved datasets that exist.

The identifier trap, measured: 'H3K9me1/2' is the compact form and it is invisible to an H3K9me2 query

This mark's literature is mostly written as the compact slashed form. Measured on live esearch, 2026-09-22, with every querytranslation read back 6:

queryrecordsrelation
H3K9me1[tiab]236
H3K9me2[tiab]1,546
"H3K9me1/2"[tiab]67all 67 are a subset of the H3K9me1 set - PubMed tokenises the slash so H3K9me1 matches inside H3K9me1/2
of those 67, reachable by H3K9me2[tiab]1552 of 67 (77.6%) are NOT

So a paper that writes only H3K9me1/2 is retrievable as me1 and not as me2, although it is about both. PMID 37788669 is one of the 52. That is why H3K9me2[tiab] AND KMT2A[tiab] returns a clean, well-formed 0 - run twice, querytranslation non-empty, warninglist ["No items found."] - for a paper whose whole subject is H3K9me2 at the KMT2A locus 6. The graph's own exposure is the mirror image: H3K9me1 is named in 4 entries and H3K9me2 in 11, and neither had a node.

The rule this generalises to: for any mark written in compact slashed form (H3K9me1/2, H3K9me2/3, H3K27me2/3, H3K4me2/3), run the slashed string as its own operand and compare ID SETS. A count comparison would have shown 67 versus 1,546 and told you nothing; the set comparison shows that the lower state's query is the only one that reaches them.

What is known in a KMT2A-rearranged context: almost nothing, and the bound is tight

Bounded absence, all queries run 2026-09-22 with querytranslation recorded, load-bearing zeros run twice 6:

querycountverdict
H3K9me1[tiab] AND KMT2A[tiab]1PMID 37788669 only - and that is RPE cells, not a leukaemia
H3K9me1[tiab] AND MLL[tiab]237788669 plus a doxorubicin-resistance breast-cancer paper matching MLL spuriously
H3K9me1[tiab] AND ("MLL-AF4" OR "t(4;11)" OR "MLL-AF9")[tiab]0genuine zero, twice
H3K9me1[tiab] AND infant[tiab]137788669 again, matching on the word 'infant' in its first sentence
"H3K9 monomethylation"[tiab]12none in a fusion context

The absence is about abstracts, and the archive check changes its class. GEO holds GSE210480, 60 human samples, which is PMID 37788669's own RPE deposit and includes H3K9me1 ChIP-seq in control, siKDM3B, siG9A and double-knockdown arms, two replicates each 8. It also holds GSE75577/GSE75581, murine Jmjd1c-loss H3K9me2 data 8. So the question 'does H3K9me1 change at the Kmt2a locus when a KDM3 family member is lost in a haematopoietic cell' is measured-not-analysed - public data, no wet work - while 'what is H3K9me1 at the KMT2A BCR in an infant KMT2A::AFF1 blast' remains never-measured.

The tension with mark-h3k9me3, stated and not flattened

mark-h3k9me3 argues that H3K9 methylation restrains the KMT2A-fusion programme: SETDB1-deposited promoter H3K9me3 represses Hoxa9 and Meis1, leukaemia stem cells from MLL-r models carry less H3K9me3 than their normal counterparts, and its bite_back therefore warns that a G9a or SETDB1 inhibitor 'is a candidate for making the leukaemia worse'. This entry's source says the opposite about direction: raising H3K9me1/2 promotes KMT2A rearrangement, so G9a inhibition is protective 5.

Both can be true, and the reason is that they are different questions about different objects.

mark-h3k9me3this entry
the DNA in questionthe fusion's target genes (HOXA9, MEIS1 promoters)the KMT2A gene itself, at its breakpoint cluster region
the time pointan established leukaemia, maintenancea normal cell, before any fusion exists
the markme3me1, and me2 on the flank
the enzymeSETDB1, SUV39H1G9a/EHMT2, opposed by KDM3B
direction of benefitraise itlower it

So mark identity plus locus identity plus disease stage resolve the contradiction, and nothing in mark-h3k9me3 needs amending.

What does NOT resolve, and is the honest residue. The dispatch brief for this beat said the tool compound UNC0638 lowers me2 and me3 together so no experiment using it can separate them. That is right about the published cellular observation (PMID 33054052, mouse lineage-negative bone marrow, cited in mark-h3k9me3 E2) and wrong about which compound PMID 37788669 used: it used UNC0642, not UNC0638 [E5, and see drug-unc0642]. The correction does not dissolve the problem, because (a) UNC0642 is a dual G9a/GLP inhibitor and UniProt annotates GLP/EHMT1 as an me3 writer 1, and (b) PMID 37788669 ran H3K9me3 ChIP-seq in every arm and reported no H3K9me3 result at all - the eight deposited samples are in GSE210480 8. The experiment that would separate the arms exists, is public, and has never been shown.

Facts and figures
markH3K9me1
histoneH3
residueK9, N-terminal tail
chemistrymono-methylation; mutually exclusive with H3K9ac on the same residue; the substrate pool from which H3K9me2 and H3K9me3 are built
genomic distributioneuchromatic gene bodies and large intergenic domains; at the KMT2A locus specifically it is DEPLETED across the 8.3 kb breakpoint cluster region relative to flanking sequence at baseline in RPE cells, which is the opposite of what a passive chromatin-state model would predict for an intragenic region
distinguished from me2at the KMT2A BCR the two states do not co-vary: KDM3B depletion raises me1 UNIFORMLY across the whole BCR while me2 rises only on the FLANK (PMID 37788669); at the HOXA1 RARE in AML cells the same perturbation raises me1 and LOWERS me2 (PMID 28540746)
distinguished from me3written by a different enzyme set (SUV39H1/2, SETDB1) and erased by a different family (KDM4, not KDM3); H3K9me1 is the substrate for me3 rather than a competitor with it
measured in kmt2a rfalse
measured in kmt2a aff1false
measured in b lineagefalse
measured in infant materialfalse
measured at the kmt2a locusYES - in RPE cells only (PMID 37788669, GSE210480)
direction at the kmt2a locusRAISING it promotes KMT2A copy gain and break-apart; the protective state is LOW H3K9me1 across the BCR with CTCF bound at exon 11
note on direction versus mark h3k9me3OPPOSITE to mark-h3k9me3's direction, and not in contradiction with it: that entry is about me3 at the fusion's TARGET GENES in an established leukaemia, this one is about me1 at the KMT2A GENE ITSELF in a normal cell. Neither entry needs amending.
epigenetic readoutH3K9me1 ChIP-seq and ChIP-qPCR across the KMT2A breakpoint cluster region and at the exon-11 CTCF site; the only published readout is in RPE cells (GSE210480)
absence classmeasured-not-analysed
writers
[
  "EHMT2/G9a (Q96KQ7, 1,210 aa) - euchromatic me1/me2; entry: gene-ehmt2",
  "EHMT1/GLP (Q9H9B1, 1,298 aa) - annotated me1/me2/me3; NO entry in this graph as of 2026-09-22",
  "MECOM/PRDM3 (Q03112, 1,230 aa) - monomethylates FREE H3 in the CYTOPLASM; no entry in this graph",
  "PRDM16 (Q9HAZ2, 1,276 aa) - named as co-primary in MECOM's record but not in its own; no entry in this graph"
]
erasers
[
  "KDM3B (Q7LBC6, 1,761 aa) - entry: gene-kdm3b",
  "KDM3A (Q9Y4C1, 1,321 aa) - no entry in this graph",
  "JMJD1C/KDM3C (Q15652, 2,540 aa) - entry: gene-kdm3c"
]
readers
[
  "EHMT2/G9a ANK repeats bind H3K9me1 and H3K9me2 - the writer reads its own product (gene-ehmt2 E1)",
  "CBX5/HP1alpha (P45973, 191 aa) and the HP1 family bind H3K9me generically; NO entry in this graph",
  "MPHOSPH8/MPP8 (Q99549, 860 aa), the HUSH-complex H3K9me reader; no entry in this graph"
]
antibodies rule 74
{
  "H3K9me1": "Abcam ab8896-100 (PMID 37788669 ChIP-seq)",
  "H3K9me2": "Abcam ab1220 - the SAME catalogue number used for the H3K9me2 ChIP in PMID 25822366, which mark-h3k9me3 E5 records",
  "H3K9me3": "Abcam ab8898",
  "CTCF": "clone D31H2, Cell Signaling #3418"
}
identifier trap
{
  "trap": "the compact slashed form 'H3K9me1/2' is retrievable as H3K9me1 and NOT as H3K9me2",
  "measured_2026_09_22": {
    "H3K9me1[tiab]": 236,
    "H3K9me2[tiab]": 1546,
    "\"H3K9me1/2\"[tiab]": 67,
    "slash_set_subset_of_me1_set": true,
    "slash_records_unreachable_by_me2_query": 52,
    "slash_records_unreachable_pct": 77.6,
    "H3K9me1[tiab] AND KMT2A[tiab]": 1,
    "H3K9me2[tiab] AND KMT2A[tiab]": 0
  },
  "worked_consequence": "PMID 37788669, whose subject is H3K9me1/2 at the KMT2A locus, is invisible to H3K9me2[tiab] AND KMT2A[tiab], which returns a clean zero twice",
  "generalises_to": "H3K9me2/3, H3K27me2/3, H3K4me2/3 and any slashed compact mark form - run the slashed string as its own operand and compare ID SETS, not counts",
  "registry_row_existed_before": false
}
corpus position when written
{
  "mark_nodes_before": 14,
  "h3k9_mark_nodes_before": [
    "mark-h3k9ac",
    "mark-h3k9me3"
  ],
  "entries_naming_h3k9me1": 4,
  "mentions_of_h3k9me1": 26,
  "entries_naming_h3k9me2": 11,
  "mentions_of_h3k9me2": 60,
  "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's direction is an argument for GIVING a G9a inhibitor alongside topoisomerase II chemotherapy, and `mark-h3k9me3` carries a published argument for NOT giving one in KMT2A-rearranged disease. Both arguments are about G9a and they point opposite ways, so any protocol built on either must state which locus and which disease stage it means.
  • The intervention with the most direct evidence here is not a drug that raises or lowers this mark: it is KDM3B or CTCF over-expression, which is not a deliverable modality in a patient. What is deliverable - a G9a inhibitor - acts on the enzyme whose obligate partner GLP is annotated to write H3K9me3, so the pharmacological handle is less mark-selective than the mechanism.
  • Every chromatin measurement behind this entry is in retinal pigment epithelium. A mark set point measured in an epithelial cell line is a weak basis for predicting anything about a fetal haematopoietic progenitor, which is the cell this project's index disease arises in.
  • The KDM3-family inhibitor that phenocopies KDM3B loss is published as an anti-leukaemic agent that preferentially kills KMT2A-rearranged cells. If both results hold, the same compound class treats established KMT2A-rearranged leukaemia and induces KMT2A rearrangements in normal cord-blood progenitors - a therapeutic window that depends entirely on which cells are exposed.
Still open
  1. Does H3K9me1 at the KMT2A breakpoint cluster region differ between fetal/cord-blood CD34+ progenitors and adult HSPCs? The whole in-utero-origin argument for infant t(4;11) needs a baseline in the cell of origin, and the only published BCR measurement is in retinal pigment epithelium.
  2. Is the BCR H3K9me1 valley a property of that sequence or of the locus's transcriptional state? The rival fragility model makes break frequency a function of transcriptional output, and KMT2A is highly transcribed in progenitors - so the valley could be a consequence of elongation rather than a set point maintained by KDM3B.
  3. What does the deposited H3K9me3 arm of GSE210480 show at the KMT2A BCR? Eight samples exist across all four knockdown arms and no H3K9me3 result was reported anywhere in the paper; this is the cheapest open question in this beat and it decides whether the me1/me2 mechanism is mark-specific.
  4. Do MECOM/EVI1 and PRDM16, the cytoplasmic H3K9me1 writers, affect KMT2A locus stability? EVI1 is itself a leukaemia driver, the enzymology is annotated, and nobody has connected the two.
Every source behind this entry (8)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    G9a/EHMT2 mono- and dimethylates histone H3 Lys-9 in euchromatin and its ANK repeats bind H3K9me1 and H3K9me2, while its heterodimer partner EHMT1/GLP is annotated as specifically mono-, di- AND trimethylating H3 Lys-9 in euchromatin (H3K9me1, H3K9me2 and H3K9me3) and also weakly methylating H3 Lys-27 - so the two halves of the obligate G9a/GLP heterodimer carry different annotated product specificities in the same database, curated from the same 2002 paper.

    What this evidence cannot show

    A database record, not a measurement, and both product-specificity statements trace to the same citation, PubMed:12004135, which was NOT retrieved for this entry - rule 91 applies and the asymmetry between EHMT1 and EHMT2 should be checked against that paper before it is built on. The asymmetry may be a curation artefact rather than a biochemical fact: G9a and GLP are the same SET-domain family and several primary reports treat them as interchangeable me1/me2 enzymes. Recorded here because the asymmetry is the difference between 'a G9a inhibitor cannot touch me3' and 'a dual G9a/GLP inhibitor might', and because reasoning in either direction without noticing it is the error to avoid.

    Evidence that argues against this (1)
    • Whatever the annotated enzymology, the measured cellular consequence of a G9a/GLP inhibitor is not me3-sparing: UNC0638 reduced both H3K9me2 and H3K9me3 in mouse lineage-negative bone marrow. So an annotation-level argument that an EHMTi is me1/me2-selective is contradicted by the one published blot in a haematopoietic cell. Ropa J; Saha N; Hu H; Peterson LF; Talpaz M; Muntean AG 2020
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    MECOM/PRDM3 displays histone methyltransferase activity and monomethylates H3 Lys-9, probably catalysing monomethylation of FREE histone H3 in the CYTOPLASM which is then transported to the nucleus and incorporated into nucleosomes, where SUV39H methyltransferases use it as the substrate for H3K9 trimethylation - MECOM being one of the primary enzymes, with PRDM16, directing cytoplasmic H3K9me1.

    What this evidence cannot show

    A database record, and the record itself hedges twice - 'probably catalyzes' and 'likely to be one of the primary histone methyltransferases'. The underlying primary paper was not retrieved for this entry, so this is recorded at database strength with the hedges preserved; rule 91 applies. Note the strong incidental relevance this beat did not pursue: MECOM/EVI1 is itself a leukaemia oncogene and a 3q26 rearrangement partner, so 'the cytoplasmic H3K9me1 writer' and 'a leukaemia driver' are the same protein. No experiment links EVI1's methyltransferase activity to KMT2A locus stability, and this entry does not claim one.

    Evidence that argues against this (1)
    • PRDM16's own UniProt FUNCTION comment (Q9HAZ2, 1,276 aa) does not mention H3K9 monomethylation at all - it records DNA binding, transcriptional regulation, TGF-beta repression and granulocyte differentiation. So the claim that PRDM16 is a co-primary cytoplasmic H3K9me1 writer is asserted in MECOM's record and absent from PRDM16's, in the same database on the same day. UniProt Consortium 2026
  3. E3strong 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 - so the KDM3/JMJD1 family is an me1/me2 eraser family and cannot erase me3.

    What this evidence cannot show

    A database record about KDM3A, used here for the family-level substrate rule; KDM3B's own record states only 'specifically demethylates Lys-9 of histone H3' with a catalytic-activity reaction written as dimethyl-to-unmethylated in two steps, and does not explicitly exclude me3 (see `gene-kdm3b` E1). The family rule is therefore imported from the paralogue and should be read as such.

    Evidence that argues against this (1)
    • The compound used to phenocopy KDM3B loss in this beat's founding experiment (JDI-12) is reported by its own discovery paper to induce a GLOBAL increase in H3K9 methylation without the methylation state being separated, and JDI-4 was shown to reverse 'the H3K9 demethylation conferred by KDM3B' in an assay whose readout state is not stated in the abstract. So the cleanly me1/me2-restricted picture is not what the pharmacology reports. Xu X; Wang L; Hu L; Dirks WG; Zhao Y; Wei Z; Chen D; Li Z; Wang Z; Han Y; et al. 2020
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    KDM3B (Q7LBC6, 1,761 aa) is a histone demethylase that specifically demethylates Lys-9 of histone H3, with an annotated catalytic reaction converting N(6),N(6)-dimethyl-L-lysyl(9)-[histone H3] to L-lysyl(9)-[histone H3] in two 2-oxoglutarate-dependent steps (EC 1.14.11.65), and it is described in its own record as possibly having tumour suppressor activity.

    What this evidence cannot show

    The FUNCTION and CATALYTIC ACTIVITY statements both rest on a single experimental citation, PubMed:16603237, which was not retrieved here; rule 91 applies. The annotated reaction starts from the DImethyl substrate, so the record does not by itself establish that KDM3B erases the MONOmethyl state, even though 'H3K9me1/2 demethylase' is how the leukaemia literature uses it and how PMID 37788669's experiments treat it. That gap between the database reaction and the literature usage is recorded rather than smoothed over.

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

    In immortalised near-diploid human RPE cells, H3K9me1 across the 8.3 kb KMT2A breakpoint cluster region (exons 8-14 in this source's numbering) is LOWER at baseline than in adjacent regions; KDM3B siRNA depletion causes a strong, uniform increase of H3K9me1 across the whole BCR with H3K9me2 rising on the BCR flank, reduces CTCF occupancy specifically at the conserved CTCF site in KMT2A exon 11, and produces site-specific KMT2A copy gains and break-apart events - and the H3K9me1 increase at that one CTCF site is among the strongest of all CTCF sites genome-wide; co-depletion of G9a/EHMT2 completely resets H3K9me1 at the BCR and prevents the alterations, while EHMT1 co-depletion does not.

    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 via efetch db=pmc on PMC10591855 after Europe PMC fullTextXML returned HTTP 500; the stripped body is 80,638 characters and a <body> element was confirmed present.** The system for every chromatin measurement in the paper is **RPE - immortalised retinal pigment epithelium, chosen precisely because it is near-diploid, genomically stable and carries no cancer mutations**. It is not haematopoietic, not lymphoid and not a leukaemia. The leukaemic and primary arms (KG1a, HL60, U937, a primary AML-derived line, a primary AML organoid, and cord-blood CD34+ HSPCs) are scored by **DNA FISH for copy gain and break-apart only** - no H3K9me1 ChIP was performed in any haematopoietic cell. The ChIP antibodies are H3K9me1 Abcam ab8896-100, H3K9me2 ab1220, H3K9me3 ab8898, CTCF clone D31H2 (CST #3418), with at least two chromatin preparations from two independent siRNAs; no spike-in normalisation is described, which matters for a mark reported to change genome-wide. 'Break apart' is a two-colour FISH probe separation, a cytogenetic proxy for rearrangement, not a sequenced junction - the paper says explicitly that integration sites and sequence maps remain to be determined.

    Evidence that argues against this (1)
    • The rival model for the same locus makes CTCF a marker of fragility rather than a protector: TOP2-mediated double-strand breaks are ENRICHED at CTCF- and cohesin-bound chromatin loop anchors, break frequency correlates positively with transcriptional output, and the genes that recurrently translocate to drive leukaemias - KMT2A among them - are highly transcribed and enriched at loop anchors. On that account CTCF presence marks where KMT2A breaks; on this entry's account CTCF presence is what stops it. PMID 37788669 cites this paper and does not address the conflict. Gothe HJ; Bouwman BAM; Gusmao EG; Piccinno R; Petrosino G; Sayols S; Drechsel O; Minneker V; Josipovic N; Mizi A; et al. 2019
  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 236 records for H3K9me1[tiab], 1,546 for H3K9me2[tiab] and 67 for "H3K9me1/2"[tiab]; all 67 of the compact-form records are a subset of the H3K9me1 set and 52 of the 67 (77.6%) are NOT reachable by H3K9me2[tiab]; H3K9me1[tiab] AND KMT2A[tiab] returns exactly 1 record (37788669), H3K9me2[tiab] AND KMT2A[tiab] returns 0, and H3K9me1[tiab] AND ("MLL-AF4" OR "t(4;11)" OR "MLL-AF9")[tiab] returns 0 - both zeros with a non-empty querytranslation and warninglist ["No items found."], each run twice.

    What this evidence cannot show

    **Titles and abstracts only.** A mark measured in a supporting figure is invisible to every one of these queries (rule 80), and this beat has a worked example: PMID 37788669 itself ran H3K9me3 ChIP-seq that appears in no figure and no result sentence. Each cytogenetic term is quoted, per rule 95, and no querytranslation contained a [UID] term. One tag per term throughout, per rule 94 - the grouped form was used only inside an OR list of quoted fusion names where the tag applies to the group, and that form is broader than labelled, which makes the zeros conservative rather than optimistic. The name list tried was H3K9me1, H3K9me2, "H3K9me1/2", "H3K9 monomethylation", "H3K9 dimethylation" and "H3 lysine 9"; it did not include every spelling a paper might use, and "ALL" cannot be used as an operand at all - it returns errorlist phrasesnotfound because PubMed treats it as a stopword.

    Evidence that argues against this (1)
    • The zero is a property of the index, not of the field: the same archive query that bounds this absence also shows the data exist. GEO returns 18 series for H3K9me2 AND (KMT2A OR MLL) AND leukemia, including the deposited H3K9me2 ChIP-seq from the murine MLL-AF9 DOT1L/SIRT1 study - so 'no H3K9me2 measurement in a KMT2A-fusion cell' would be false, and only the narrower claim about the KMT2A locus survives. Chen CW; Koche RP; Sinha AU; Deshpande AJ; Zhu N; Eng R; et al. 2015
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In AML, KDM3B is underexpressed in patient samples and in AML cell lines carrying MLL-AF6, MLL-AF9 or PML-RARA translocations; KDM3B overexpression repressed colony formation of an AML line with 5q deletion; and KDM3B knockdown increased H3K9 MONO-methylation while DECREASING H3K9 DI-methylation at the retinoic-acid response elements of HOXA1, while sparing the HOXA1 promoter region.

    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 method, the antibodies (rule 74), the element coordinates and the effect sizes for the opposite-sign me1/me2 result are not characterised here, and that result is the load-bearing part - it should be read in full before being built on. Myeloid cell lines throughout, no B-lineage arm, no KMT2A::AFF1, and the causal arm is overexpression in a 5q-deleted line rather than loss in a normal cell. The same abstract states in its own first sentence that KDM3B 'reportedly shows both tumor-suppressive and tumor-promoting activities in leukemia' and that the function is likely cell-type dependent, which is the honest reading of the whole KDM3B literature.

    Evidence that argues against this (1)
    • In the lymphoid lineage the direction inverts: KDM3B is an H3K9me1/2 demethylase that activates the leukaemogenic oncogene LMO2, represses leukaemia-cell differentiation, and is UP-regulated in blood cells from ALL-type leukaemia patients. Since this graph's index disease is infant B-ALL, the lineage in which KDM3B looks pro-leukaemic is the lineage that matters most here. Kim JY; Kim KB; Eom GH; Choe N; Kee HJ; Son HJ; Oh ST; Kim DW; Pak JH; Baek HJ; et al. 2012
  8. E8established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    GEO series GSE210480 (Homo sapiens, 60 samples) is PMID 37788669's own deposit and contains H3K9me1, H3K9me2, H3K9me3, CTCF and Input ChIP-seq in RPE cells across control, siKDM3B, siG9A and siKDM3B+siG9A arms with two biological replicates each - including eight H3K9me3 samples; separately, GSE75577 and GSE75581 hold murine Jmjd1c-loss H3K9me2 genome-wide data, and GSE61021/GSE61022 hold the murine MLL-AF9 H3K9me2 ChIP-seq from the DOT1L/SIRT1 study.

    What this evidence cannot show

    A listing of deposited sample titles establishes that the assay was run and the files exist; it does not establish that the data are of usable quality, that the H3K9me3 arm has adequate depth, or that a re-analysis would be informative. Sample titles are also inconsistently cased in the deposit (H3K9ME2, H3K9me2, CTRLsiKDm3B), which will break a naive string-matched pipeline. The GEO series title also differs from the published title, so a title-matching join between the paper and its data would fail.

    Evidence that argues against this (1)
    • A GEO series can announce data it does not usably contain - this project has recorded a SuperSeries asserting no raw data in three different ways, all three wrong, and a well-formed, correctly named, empty supplementary archive. The presence of eight H3K9me3 sample records is a strong prior that the analysis is cheap, not a guarantee. Gray ZH; Chakraborty D; Duttweiler RR; Alekbaeva GD; Murphy SE; et al. 2023
Connected to