The KMT2A locus sits inside a 179-318 kb H3K9me2-domain-free block in human CD34+ bone-marrow progenitors - the 7th, 11th and 37th longest such block on chr11 in three ChIPs - and the same block is present in granulocytes, K562 and most primary AML
also called the H3K9me2 gap at 11q23.3 · rank 7 of 2,076 chr11 gaps · the progenitor baseline, measured
In real human blood stem cells, the KMT2A gene sits in one of the largest stretches of chromosome 11 that carries none of the silencing tag at all - a gap of two to three hundred thousand bases. That is the same baseline scarcity the founding paper reported in eye-derived cells, now seen in the right kind of cell. But the same gap is there in mature white cells and in leukaemia too, so it is a feature of the place, not of the cell.
The same thing, for a scientist
The first link of the published *KMT2A* rearrangement brake is a **resting-state** property: H3K9me1/2 are unusually scarce across the breakpoint cluster region. Until now that had been measured in hTERT-RPE and, for H3K9me1 only, in K562 [E2]. **In normal human CD34+ bone-marrow haematopoietic progenitors, at domain resolution, it holds** [E1]. In all three CD34+ H3K9me2 ChIPs of GSE71809 (two donors), the *KMT2A* BCR, the 1,048 bp CTCF/RAD21/SMC3 element, canonical exon 11, both 20 kb flanks and the whole 90,341 bp gene span carry **zero** called H3K9me2 domain coverage, while **40.0%, 49.3% and 44.9% of chr11** is inside a called domain. The gene span plus its 20 kb flanks lies inside a contiguous domain-free block of **317,700 / 274,500 / 179,300 bp**, ranking **7 of 2,076**, **11 of 2,116** and **37 of 1,749** chr11 blocks by length - the **98.1st to 99.7th percentile** [E1]. **The block is bounded and its edges are reproducible.** Two of the three ChIPs place its centromeric edge within 800 bp of each other - hg19 chr11:118,217,700 and 118,218,000 - about **90 kb centromeric of the *KMT2A* transcription start** and ~135 kb from the BCR. Its telomeric edge varies more: 118,535,400, 118,491,700 and 118,397,300, i.e. between the end of the gene and ~140 kb past it [E1]. **Two facts stop this being read as a progenitor-specific or heterochromatin-related feature.** The block is present in **4 of 4 granulocyte, 2 of 2 K562, 2 of 2 UNC0638-treated K562 and 6 of 10 primary AML** H3K9me2 ChIPs, and in the CD34+ H3K9me3, H3K27me3 and H3K4me2 samples. And the CD34+ **H3K27ac** sample covers the BCR and the element **100%**, and the gene span 90% - so this is an **active** region in a progenitor, not silent chromatin, and an H3K9me2-free block is what active chromatin looks like at this resolution [E1]. **The null, stated because it deflates the headline:** 17.8-21.5% of same-width chr11 windows are entirely domain-free in these samples, so *being* free is roughly a 1-in-5 event. Only the **block length percentile** is a strong statement, and block length is a property of a ~250 kb neighbourhood rather than of the 7.5 kb BCR [E1].
What is new, and what this cannot say
New. The resting-state half of the chain has never before been measured in a primary human haematopoietic progenitor in this graph. ana-h3k9me1-depletion-over-the-kmt2a-bcr-replicates-in-a-blood-cell-but-not-at-the-element reached K562 - a CML-derived erythroleukaemia line - and named the substitution as its principal weakness. This reaches CD34+ bone-marrow progenitors from two donors, which is the cell compartment the disease arises in, and for the mark (H3K9me2) that the published balance model puts on the G9a side.
Cannot say: anything about the 1,048 bp element. The called domains have a median width of 12,750-17,000 bp in these samples, so the element, the exon and the BCR are all sub-domain objects and their coverage values are binary membership, not occupancy (concept-an-enriched-domain-call-is-not-a-coverage-track). The element-scale question remains open and is exactly where the K562 bigWig comparison disagreed in sign - 0.877 in one track and 1.155 in the other.
Cannot say: anything about a change. GSE71809 has no KDM3B arm in any cell type. The G9a arm it does have is pharmacological (UNC0638 in K562), and at a locus already free of domains there is no floor left to fall to, so that arm returns no information here.
Cannot say: anything about CTCF. GSE71809 has no CTCF ChIP. The step that mechanism-kdm3b-depletion-lowers-ctcf-only-at-the-kmt2a-anchor treats as carrying the locus address remains unmeasured in any blood cell, which is what absence-no-human-ctcf-or-cohesin-occupancy-measurement-in-a-blood-cell-at-kmt2a records.
Cannot say: whether any of the AML samples is KMT2A-rearranged. Four of the ten AML H3K9me2 ChIPs place a domain over the locus and six do not. The GEO characteristics carry sample id, sex and antibody and no karyotype, so that 6:4 split cannot be read against fusion status - and it must not be assumed to be uninformative either.
Coordinates, for anyone joining to this
All hg19, from this project's GRCh38 values minus the offset 129,285 bp, re-derived this cycle across six intervals from 2 bp to 90,341 bp with Ensembl mappings filtered on seq_region_name == "11" and the offset constant in all six.
| interval | hg19 | width | H3K9me2 domain coverage, 3 CD34+ ChIPs |
|---|---|---|---|
| BCR core (DCAL A-B-C) | chr11:118,353,137-118,360,602 | 7,466 | 0, 0, 0 |
| CTCF/RAD21/SMC3 element | chr11:118,358,874-118,359,921 | 1,048 | 0, 0, 0 |
| canonical exon 11 | chr11:118,359,329-118,359,475 | 147 | 0, 0, 0 |
| DCAL cut point +/-250 | chr11:118,355,327-118,355,827 | 501 | 0, 0, 0 |
| 20 kb centromeric flank | chr11:118,333,137-118,353,136 | 20,000 | 0, 0, 0 |
| 20 kb telomeric flank | chr11:118,360,603-118,380,602 | 20,000 | 0, 0, 0 |
| gene span | chr11:118,307,207-118,397,547 | 90,341 | 0, 0, 0 |
| 500 kb centromeric | chr11:117,807,207-118,307,206 | 500,000 | 0.38, 0.36, 0.48 |
| 500 kb telomeric | chr11:118,397,548-118,897,547 | 500,000 | 0.39, 0.37, 0.72 |
The 500 kb flanks are the internal scale control and they behave as they must: at 36-72% they sit near the chromosome-wide 40-49%, so the zero is a local property of a few hundred kilobases and not a property of the whole neighbourhood or of the sample.
| element | the H3K9me2-domain-free block containing the KMT2A locus in human CD34+ bone-marrow haematopoietic progenitors |
|---|---|
| centromeric edge reproducibility bp | 800 |
| kmt2a gene grch37 | chr11:118,307,207-118,397,547 |
| grch38 minus grch37 offset bp | 129285 |
| offset re derived this cycle | true |
| n windows tested | 13496 |
| resolution bound | median called-domain width 12,750-17,000 bp in these samples, so every interval narrower than that is binary membership and not occupancy; see concept-an-enriched-domain-call-is-not-a-coverage-track |
| localization | chromatin, chr11q23.3 and ~90-140 kb of flanking sequence on each side; a domain-occupancy statement that asserts nothing about protein localisation |
| subnuclear localization | not measured. A large H3K9me2-free, H3K27ac-covered block is the configuration usually associated with the nuclear interior rather than the lamina, and this graph holds a separate entry on KMT2A and lamina-associated domains, but no imaging, Hi-C or DamID was read here and no subnuclear position is claimed. |
| cell type | normal human CD34-positive bone-marrow haematopoietic progenitors, two donors (cryopreserved, AllCells), with peripheral-blood granulocytes, K562 and eight primary AML myeloblast samples as comparators. This is the first primary haematopoietic progenitor measurement at this locus in this graph. |
| cell state | resting / unperturbed, thawed and fixed immediately; no knockdown, no drug in the progenitor arm |
| epigenetic readout | H3K9me2 ENRICHED-domain occupancy, with H3K9me3, H3K27me3, H3K27ac and H3K4me2 in the same donor as within-cell mark comparators and the chromosome-wide domain fraction as the null |
| shown for fusion | Not shown for any fusion. The CD34+, granulocyte and K562 cells carry no KMT2A fusion; the AML samples have no karyotype or fusion status in GEO, so whether any carries one is unknown. |
grch37 block per sample — the numbers behind the chart
{
"GSM1846171": {
"start": 118217700,
"end": 118535400,
"len_bp": 317700,
"chr11_gap_rank": 7,
"n_chr11_gaps": 2076,
"percentile": 0.9971
},
"GSM1846172": {
"start": 118217200,
"end": 118491700,
"len_bp": 274500,
"chr11_gap_rank": 11,
"n_chr11_gaps": 2116,
"percentile": 0.9957
},
"GSM2152593": {
"start": 118218000,
"end": 118397300,
"len_bp": 179300,
"chr11_gap_rank": 37,
"n_chr11_gaps": 1749,
"percentile": 0.9806
}
}chr11 domain fraction in the same samples
[ 0.449, 0.4, 0.493 ]
genome domain fraction in the same samples
[ 0.435, 0.388, 0.466 ]
the null same width windows entirely free
[ 0.1889, 0.2148, 0.1781 ]
internal scale control 500kb flanks covered fraction
{
"centromeric": [
0.38,
0.36,
0.48
],
"telomeric": [
0.39,
0.37,
0.72
]
}present in other cells
{
"granulocyte_H3K9me2": "4 of 4",
"K562_H3K9me2": "2 of 2",
"K562_plus_UNC0638_H3K9me2": "2 of 2",
"primary_AML_myeloblast_H3K9me2": "6 of 10 free, 4 of 10 covered",
"CD34_H3K9me3": "free",
"CD34_H3K27me3": "free",
"CD34_H3K4me2": "free",
"CD34_H3K27ac": "NOT free - BCR and element 100% covered, gene span 90%"
}what it does not measure
[ "the 1,048 bp CTCF element, which is sub-domain", "any change - GSE71809 has no KDM3B arm in any cell type", "CTCF or cohesin, which this series does not assay", "whether any of the eight AML samples is KMT2A-rearranged" ]
- This is the measurement most likely to be quoted as 'the mechanism holds in blood', and it does not support that sentence. What holds in blood is the resting-state scarcity of one mark over a large neighbourhood - which is also true in granulocytes, in a leukaemia line, in drug-treated cells and in most AML, and which the H3K27ac result suggests is just what an active locus looks like. None of the perturbation, none of the CTCF step and nothing at element scale is measured here.
- For an intervention the implication is deflationary and should be said plainly: if the locus already carries no H3K9me2 domain in a progenitor, a G9a inhibitor given prophylactically has no H3K9me2 at this locus to remove, so its protective action - if any - must be somewhere other than where the model places it. And four of ten AML samples DO carry a domain there, which points the drug towards established disease rather than prevention, the opposite of how this graph has framed it.
- Is the KMT2A locus H3K9me2-depleted at coverage resolution in a CD34+ progenitor, or merely outside a called domain? The 31 SRA libraries of GSE71809 would give a BCR/flank signal ratio directly comparable with K562's 0.563 and 0.560 and RPE's 0.422, and that is the number this graph actually needs. It is an alignment job.
- Where exactly is the block's telomeric edge? Its centromeric edge reproduces to 800 bp across two ChIPs while its telomeric edge varies over 138 kb (118,397,300 to 118,535,400). If that asymmetry is real it is interesting, because the paper reports de novo breakpoints skewed centromeric and therapy-related breakpoints skewed telomeric within the BCR, and an H3K9me2 boundary with a sharp centromeric edge and a fuzzy telomeric one is a shape worth explaining.
- Does this block coincide with a lamina-associated domain boundary or a replication-timing transition in a progenitor? Both are measurable from public CD34+ data - Repli-seq and DamID exist for haematopoietic cells - and either would give the block a structural identity rather than leaving it as an absence of a mark.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
In all three CD34+ H3K9me2 ChIP-seq samples of GSE71809 (GSM1846171, GSM1846172 from donor CD34+ 1 and GSM2152593 from donor CD34+ 2) the KMT2A BCR, the 1,048 bp CTCF element, canonical exon 11, the DCAL cut point, both 20 kb flanks and the 90,341 bp gene span carry zero called ENRICHED-domain coverage while 44.9%, 40.0% and 49.3% of chr11 is inside a called domain; the gene span plus 20 kb flanks lies inside a contiguous domain-free block of 317,700, 274,500 and 179,300 bp, ranking 7 of 2,076, 11 of 2,116 and 37 of 1,749 chr11 blocks (98.1st-99.7th percentile), with reproducible centromeric edges at hg19 chr11:118,217,700 and 118,218,000; the 500 kb flanks are 36-72% covered, near the chromosome-wide value; 17.8-21.5% of same-width chr11 windows are entirely domain-free, which is the null; the block is also present in 4 of 4 granulocyte, 2 of 2 K562, 2 of 2 UNC0638-treated K562 and 6 of 10 primary AML H3K9me2 ChIPs and in CD34+ H3K9me3, H3K27me3 and H3K4me2; and CD34+ H3K27ac covers the BCR and the element 100% and the gene span 90%.
GSE71809 RSEG ENRICHED-domain BED files for 31 samples, parsed over hg19 intervals with per-sample chromosome-wide domain-coverage nulls, a same-width window null and a chr11 gap-length ranking, 2026-09-22; code and full per-sample output at analyses/q111-cd34-h3k9me2-over-the-kmt2a-bcr/run.py, nulls.py, results.json and results_nulls.json · Salzberg AC, Harris-Becker A, Popova EY, Keasey N, Loughran TP, Claxton DF, Grigoryev SA · NCBI GEO (data accompanying PLoS One 12:e0173723) 2017What this evidence cannot show
Domain calls, not coverage: 'zero' means RSEG called no enriched domain, and a sub-threshold enrichment is invisible. Median called-domain width is 12,750-17,000 bp in these samples, so the element, the exon and the BCR are sub-domain objects whose coverage values are binary membership rather than occupancy, and the mark-identity control cannot discriminate at this scale - in the CD34+ H3K9me3 sample the ZNF274 body, the GAPDH promoter, the ACTB promoter and the MLLT1 and TCF3 gene bodies are all 100% covered. Two donors, three ChIPs, cryopreserved commercial bone-marrow CD34+ cells sorted only on CD34, chromatin digested with micrococcal nuclease rather than sonicated, two sequencing platforms mixed across the series, and chrX/chrY reads duplicated before domain calling in male samples (irrelevant here, since every interval is autosomal). The chr11 domain fraction varies 40.0-49.3% between three samples of the same cell type, a 23% relative spread that bounds how precise any gap percentile can be. The 100 bp mask used for the gap analysis rounds every boundary outward, which can only shorten a gap, so the block lengths are conservative. hg19 coordinates are this project's GRCh38 values minus 129,285 bp, re-derived in analyses/q111-ctcf-anchor-matched-ranking/run.py and not re-lifted here.
Evidence that argues against this (1)
- The H3K27ac result is the strongest argument that this block is uninteresting. If the KMT2A locus is fully covered by an H3K27ac domain in the same progenitor, then it is active chromatin, and active chromatin is H3K9me2-depleted by definition - so the 'baseline depletion' that the published chain treats as a licence for instability may be nothing more than the ordinary chromatin state of a transcribed gene. On that reading the first link of the chain is not a special property of KMT2A at all, and the fact that the same block appears in granulocytes, K562, UNC0638-treated K562 and most AML is exactly what one would expect of a housekeeping-active locus. NCBI Gene Expression Omnibus 2017
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
Before this measurement the resting-state H3K9me1/2 depletion over the KMT2A breakpoint cluster region had been measured only in hTERT-RPE cells and, for H3K9me1 alone, in K562 - where it replicated at BCR scale (BCR/flank 0.563 and 0.560 against RPE's 0.422) but where the two available tracks disagreed in SIGN at the 1,048 bp element (0.877 and 1.155), leaving the element-scale result unavailable; and no CD34+ progenitor H3K9me1 exists in ENCODE, which holds four H3K9me1 experiments in total.
ana-h3k9me1-depletion-over-the-kmt2a-bcr-replicates-in-a-blood-cell-but-not-at-the-element, this graph's own analysis of ENCODE ENCSR000AKW K562 H3K9me1 fold-change-over-control bigWigs (ENCFF271KWX, ENCFF654SLZ), read 2026-09-22 · Iqbal Lab oncofusion knowledge graph 2026What this evidence cannot show
That analysis measured H3K9me1 and this one measures H3K9me2 - different marks with different writers, different domain sizes and, in the published balance model, opposite roles - so this is a complementary measurement and not a replication of it. Its readout was a signal ratio from base-resolution bigWigs and this one is a domain call, so the two are not numerically commensurable. Its cell was K562, a transformed erythroleukaemia line; these are primary CD34+ progenitors.
Evidence that argues against this (1)
- The K562 element-scale sign disagreement is the warning that should be carried into any reading of this entry: at the scale that matters for a CTCF anchor, the one blood-cell dataset with enough resolution to answer the question gave two opposite answers from two replicates of one experiment. A domain-level agreement in CD34+ cells does not resolve that, and quoting the progenitor result as though it settled the element would be exactly the substitution this graph has already made once. Iqbal Lab oncofusion knowledge graph 2026
the locus-level record of that measurement, with block coordinates and chr11 gap ranks
a 179-318 kb H3K9me2-domain-free block containing KMT2A in a primary human haematopoietic progenitor
the resting-state H3K9me2 baseline that mechanism starts from, now measured in a CD34+ progenitor at domain resolution
- expressed in →Bone marrow haematopoietic stem cell — the postnatal comparator, and why the same fusion behaves differently here
CD34-positive bone-marrow haematopoietic progenitors, two donors, resting
the same block is 100% H3K27ac-covered at the BCR and the element in the same progenitor donor - active chromatin, not silent
the substrate argument: no H3K9me2 domain at the locus in a healthy progenitor, one in 4 of 10 primary AML samples