Oncofusion

start / graph / ana-q114-the-methyl-poor-kmt2a-breakpoint-domain-shows-up-again-in-a-third-lineage

Analysismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q114-suv39h1-at-the-locus

In patient-derived H3K27M glioma, baseline H3K9me3 over the 7,466 bp KMT2A breakpoint cluster core sits at the 7.8th and 28.5th percentile of 59,821 width-matched CTCF-anchor windows (0.56x and 0.90x the random-tile median) - the methyl-poor-BCR pattern this graph found in epithelium and in a blood line, now in a third lineage and for the trimethyl state

also called BCR is methyl-poor in glioma too · Q114 H3K9me3 baseline at the BCR

The stretch of DNA where the KMT2A gene usually breaks carries unusually little of the silencing tag in two children's brain-tumour cell lines - the same pattern this project already found in eye cells and in a leukaemia line. The pattern now shows up in a third kind of tissue, which makes it look like a property of the DNA rather than of one cell type.

The same thing, for a scientist

**A by-product of the writer beat, and it points the same way as two earlier measurements in two other lineages.** This graph holds the *KMT2A* breakpoint cluster region as an **H3K9-methyl-poor broad domain**: H3K9me1 over the BCR is 0.54x the track's chr11 mean against 1.28x in its 20 kb flanks in RPE, and that depletion replicates in K562 as a property of the **7.5 kb BCR** and specifically *not* of the 1 kb CTCF element, where the two tracks disagree in sign [E3]. **Measured here for H3K9me3 at baseline, in a third lineage** [E1][E2]. GSE225028's two H3K9me3 CUT&RUN tracks over `bcr_core` (chr11:118,482,422-118,489,887, 7,466 bp), against 59,821 width-matched windows centred on ENCODE CTCF summits and 20,000 width-matched random tiles: | | SU-DIPG13 | SU-DIPG36 | |---|---:|---:| | mean over `bcr_core` | 0.7502 | 0.6706 | | **percentile vs width-matched CTCF anchors** | **7.78** | **28.52** | | percentile vs width-matched random tiles | 15.25 | 41.43 | | fold over the random-tile median | **0.556** | **0.902** | | class-matched anchor median at this width | 1.5760 | 0.8707 | **Same direction in both lines, strong in one and weak in the other.** The headline number quoted here is deliberately the pair, not the striking one: 7.78 **and** 28.52. A single-line 7.78th percentile would be over-quotable; two lines pointing down, one of them barely, is what was actually measured. **And the depletion is scale-specific in the same way the earlier measurements were.** In the same two tracks the 147 bp exon-11 anchor is at the **57.8th and 63.1st** percentile - i.e. *above* the anchor median - while the 7,466 bp BCR around it is at the 7.8th and 28.5th. The element is not depleted; the domain is. That is exactly the split the K562 H3K9me1 replication found, in a different mark, a different assay and a different lineage [E3]. **Lineage first: this is paediatric H3K27M diffuse midline glioma**, and in a cell type whose defining lesion rewrites the H3K27 landscape one cannot assume the H3K9 landscape behaves as it does elsewhere. What the result adds is a **third independent lineage** - epithelial (RPE), myeloid-erythroid (K562), now neural glioma - showing the same domain-scale depletion at the same coordinates.

Go deeper
Why a third lineage matters for the mechanism this graph argues about

mechanism-h3k9me1-2-ctcf-balance-gates-kmt2a-rearrangement proposes that a KDM3B-G9a balance keeps the BCR methyl-poor and CTCF-bound, and that losing either end makes the locus amplify and rearrange. The load-bearing measurements for the methyl-poor part are in RPE and K562.

A methyl-poor BCR in a cell type with no relationship to either - a neurosphere culture from a paediatric brainstem tumour - is evidence that the depletion is a property of the sequence and its chromatin context rather than of a lineage programme. That is the reading the mechanism entry needs and did not have.

It also does not help the mechanism's causal claim at all. A constitutive methyl-poor domain present in cells that never rearrange KMT2A cannot be what makes the locus rearrange in a B-progenitor; if anything, finding it everywhere makes it a weaker candidate for the specificity. Both halves of that are stated here because only one of them is convenient.

The KMT2A promoter against the promoter-matched null, in the same tracks

The pre-registered N3 null exists because a previous positive at the KMT2A promoter reached the top 0.07% of 1.51 M genome tiles and fell to 1.24x against four housekeeping promoters in the same track. Run again here, on the SUV39H1 tracks:

trackkmt2a_tss_2kbACTBGAPDHRPLP0EEF1A1KMT2A / mean(HK)
SU-DIPG13 SUV39H10.21923.0202.2702.0680.3580.114x
SU-DIPG36 SUV39H11.21932.1730.4710.8131.1381.06x

So in SU-DIPG13 the KMT2A promoter carries a ninth of the SUV39H1 signal of an average housekeeping promoter, and in SU-DIPG36 about the same as one. Neither is a positive. The promoter-matched null is doing its job in the direction it was built for.

Facts and figures
scriptanalyses/q114-suv39h1-at-the-kmt2a-locus/run_bigwig.py
statusrun
run date2026-09-22
run byQ114-suv39h1-at-the-locus
result summaryBaseline H3K9me3 over the 7,466 bp KMT2A BCR core is at the 7.78th (SU-DIPG13) and 28.52nd (SU-DIPG36) percentile of 59,821 width-matched CTCF-anchor windows, 0.556x and 0.902x the width-matched random-tile median - while the 147 bp exon-11 anchor inside it is at the 57.8th and 63.1st percentile.
bcr core grch38chr11:118482422-118489887
bcr core width bp7466
cell typeSU-DIPG13, SU-DIPG36 - patient-derived H3K27M diffuse midline glioma neurospheres
lineagepaediatric glioma - NOT haematopoietic, NOT KMT2A-rearranged
shown for fusionNO - nobody has measured H3K9 methylation at this locus in any KMT2A-rearranged cell of any lineage
genome buildGRCh38/hg38, verified from each file's own chrom table
statistic statusNOT pre-registered as a primary endpoint - a by-product of a design aimed at the writer, reported with both lines' numbers in the headline for that reason
inputs
[
  "GSM7037529",
  "GSM7037530",
  "ENCFF685KTA"
]
h3k9me3 bcr
{
  "SU-DIPG13": {
    "mean": 0.7502,
    "pct_vs_class_matched": 7.78,
    "pct_vs_random_tiles": 15.25,
    "fold_vs_random_median": 0.556
  },
  "SU-DIPG36": {
    "mean": 0.6706,
    "pct_vs_class_matched": 28.52,
    "pct_vs_random_tiles": 41.43,
    "fold_vs_random_median": 0.902
  }
}
h3k9me3 exon11 same tracks
{
  "SU-DIPG13": {
    "mean": 1.6301,
    "pct_vs_class_matched": 57.82
  },
  "SU-DIPG36": {
    "mean": 0.8356,
    "pct_vs_class_matched": 63.07
  }
}
promoter matched null on the suv39h1 tracks
{
  "SU-DIPG13": {
    "kmt2a_tss_2kb": 0.2192,
    "ACTB": 3.02,
    "GAPDH": 2.27,
    "RPLP0": 2.068,
    "EEF1A1": 0.358,
    "ratio_to_housekeeping_mean": 0.114
  },
  "SU-DIPG36": {
    "kmt2a_tss_2kb": 1.2193,
    "ACTB": 2.173,
    "GAPDH": 0.471,
    "RPLP0": 0.813,
    "EEF1A1": 1.138,
    "ratio_to_housekeeping_mean": 1.061
  }
}
lineages now showing the domain scale depletion
[
  "retinal pigment epithelium (H3K9me1, GSE210480)",
  "K562 (H3K9me1, ENCODE)",
  "SU-DIPG13 and SU-DIPG36 (H3K9me3, GSE225028)"
]
How this could bite us back
  • If the methyl-poor breakpoint domain is constitutive across lineages, then a therapeutic argument that restoring H3K9 methylation over the BCR would protect against rearrangement is arguing for reversing a state that normal cells of several types maintain - and the KDM3B literature says losing that state causes amplification, not that keeping it is safe to alter.
  • A G9a/GLP inhibitor raises no flag from this measurement, but neither is it exonerated by it: the mark is depleted at the domain and present at the element, so an agent acting on the element and an agent acting on the domain would show different toxicities, and no study has separated them.
Still open
  1. Is the domain-scale depletion present in a cell that actually rearranges KMT2A? Every lineage it has now been seen in - RPE, K562, two glioma lines - has an intact 11q23 and does not generate KMT2A fusions, so a constitutive depletion cannot by itself explain lineage-restricted rearrangement. A fetal-liver or CD34+ measurement is the missing arm, and absence-no-h3k9me1-or-h3k9me2-in-a-cd34-progenitor-in-the-ihec-federation records why it does not exist.
  2. Does the depletion track replication timing or gene-body transcription rather than anything specific to the breakpoint? The BCR is inside a transcribed gene body, and a matched null of 7,466 bp windows inside other transcribed gene bodies of similar expression would separate 'methyl-poor because breakpoint' from 'methyl-poor because transcribed'. Not run, and it is the cheapest way this finding could be deflated.
  3. Why is the depletion 7.78th percentile in SU-DIPG13 and 28.52nd in SU-DIPG36? With one sample per line the deposit cannot say whether that is a line difference, an H3K27M-dose difference, or library depth.
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    GSE225028's two H3K9me3 CUT&RUN bigWigs (SU-DIPG13, SU-DIPG36; CST #13969S; Bowtie2 to hg38) were read over the 7,466 bp KMT2A breakpoint cluster core and six other intervals against 59,821 width-matched CTCF-anchor windows and 20,000 width-matched random tiles.

    What this evidence cannot show

    One sample per antibody per line - there is no replicate, so the difference between the two lines cannot be decomposed into biology and technical variation. The deposit's H3K9me3 tracks FAILED this beat's pre-registered KRAB-ZNF positive control at all four window scales tried (best 2.20x against a required 3x), which is the main limit on how much weight this depletion can bear.

    Evidence that argues against this (1)
    • A CUT&RUN track with no IgG control cannot distinguish low signal from low accessibility to the antibody, and a broad-domain depletion is exactly the shape a coverage or accessibility artefact would take. The internal comparison to 59,821 same-width CTCF-anchor windows is the guard against that, not a proof. source
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    The intervals, the nulls, the thresholds and a NEGATIVE prior were fixed on disk before any signal at this locus was read.

    Q114 pre-registration - SUV39H1 occupancy at the KMT2A locus · Q114-suv39h1-at-the-locus · oncofusion-knowledge-graph 2026
    What this evidence cannot show

    The pre-registration was written for the WRITER question. The H3K9me3 tracks were included as a companion control (C1b) and as the Question-C comparator, so the BCR-depletion result reported here is a by-product of a design aimed elsewhere and carries no pre-registered threshold of its own.

    Evidence that argues against this (1)
    • A finding that falls out of a design aimed at a different question has not been protected from selection: seven intervals times four tracks is 28 numbers, and picking the striking one is exactly what rule 84 and rule 87 warn about. It is reported here with its direction stated in both lines and with the weaker line's number in the headline. source
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    This graph already holds the KMT2A breakpoint cluster region as an H3K9-methyl-poor broad domain, measured for H3K9me1 in retinal pigment epithelium (0.54x the track's chr11 mean over the BCR against 1.28x in its 20 kb flanks) and replicated in K562 as a property of the 7.5 kb BCR rather than of the 1 kb CTCF element.

    PMID 37788669DOI
    What this evidence cannot show

    CORRECTION MADE WITHIN THIS ENTRY'S OWN SESSION: the title, authors, journal and year of PMID 37788669 were first written from recall as a different paper and were wrong; they were then fetched live from PubMed efetch on 2026-09-22 (DOI read only from PubmedData/ArticleIdList) and replaced with the real record. The paper is quoted here only for the deposit and its cell type; the BCR numbers are this graph's own reads of the deposited tracks (ana-q108-h3k9me3-at-the-kmt2a-bcr-measured, ana-h3k9me1-depletion-over-the-kmt2a-bcr-replicates-in-a-blood-cell-but-not-at-the-element), not statements from the paper. Those tracks are 1 kb binned, which the corpus records as a hard resolution limit.

    Evidence that argues against this (1)
    • The K562 replication found the two tracks DISAGREE IN SIGN at the 1 kb CTCF element while agreeing over the 7.5 kb BCR, so 'the BCR is methyl-poor' and 'the element is methyl-poor' are different claims and only the first replicates. This entry is about the BCR-scale claim only. source
Connected to