Between-patient variation in oncoprotein occupancy is not a fusion-partner effect: across six partners the within-partner range of the amplitude statistic is as wide as the between-partner range of its medians
Which partner gene the fusion has does not explain how much chromatin it occupies. Two children with the same partner differ about as much as two children with different partners.
The same thing, for a scientist
**The question.** The magnitude of oncoprotein occupancy varies enormously between patients in this cohort. The first structural explanation to test is the fusion partner, because the published claim is that the partner influences binding-site selection [E2]. | partner | n | median amplitude | amplitude range | called-promoter range | |---|---:|---:|---:|---:| | KMT2A-AF9 | 3 | 1.237 | 1.001-1.730 | 147-2282 | | KMT2A-ENL | 7 | 1.005 | 0.347-1.890 | 0-2060 | | KMT2A-AFF1 | 8 | 0.951 | 0.789-2.219 | 429-2095 | | KMT2A-AFDN | 4 | 0.935 | 0.683-1.548 | 37-1069 | | KMT2A-AF10 | 3 | 0.920 | 0.652-2.259 | 22-1207 | | KMT2A-EPS15 | 2 | 0.828 | 0.636-1.020 | 1233-2063 | **The between-partner spread of medians is 1.49x** (0.828 to 1.237), while the **within-partner range reaches 5.44x** in the widest partner group and the whole-cohort amplitude spread is **6.51x** [E1]. **Partner medians occupy a small part of the total range.** **This is a null and it is reported with its limits, not as a finding.** n is 8, 7, 4, 3, 3 and 2 per partner. Those are not group sizes that can exclude a partner effect; the widest partner group spans nearly the whole cohort range on its own, which is what makes the null informative rather than merely underpowered. The honest statement is that **partner identity does not organise the occupancy range**, not that partner has no effect. **It matters because it pushes the explanation onto the host cell.** If the amount were a partner property it would be a property of the fusion and, in principle, of the molecule. Since it is not, the candidates are oncoprotein expression level, cofactor availability, chromatin accessibility, tumour purity - all properties of the patient's cells rather than of the rearrangement.
Why this is not simply the published claim restated
The paper's claim is about WHERE the oncoprotein binds - partner-specific binding patterns over gene sets, correlated with the prevalence of each rearrangement in ALL versus AML. That is an address claim and this entry does not test it. This entry tests the AMOUNT, which is a different quantity and is the one this beat was dispatched to measure. Partner can organise the address without organising the amount, and the data here are consistent with exactly that.
| script | analyses/q144-unbounded-magnitude/run.py |
|---|---|
| status | run |
| result summary | Amplitude medians by partner span a small part of the cohort range; within-partner ranges are as wide. |
by partner — the numbers behind the chart
{
"KMT2A-AF10": {
"n": 3,
"median_A": 0.9195295439012148,
"min": 0.6516077987372128,
"max": 2.2593809605938557
},
"KMT2A-AF9": {
"n": 3,
"median_A": 1.2365497110353125,
"min": 1.0011176709745866,
"max": 1.7303467278165785
},
"KMT2A-AFDN": {
"n": 4,
"median_A": 0.9346093471810206,
"min": 0.6833229026775943,
"max": 1.5477822284705594
},
"KMT2A-AFF1": {
"n": 8,
"median_A": 0.9509379811057602,
"min": 0.7888130939849658,
"max": 2.2188001740707923
},
"KMT2A-ENL": {
"n": 7,
"median_A": 1.004577858746427,
"min": 0.3473269790646345,
"max": 1.8895908405144881
},
"KMT2A-EPS15": {
"n": 2,
"median_A": 0.8282315542476336,
"min": 0.6363706909774549,
"max": 1.0200924175178123
}
}calls by partner
{
"KMT2A-ENL": [
0,
345,
767,
890,
893,
947,
2060
],
"KMT2A-AFDN": [
37,
332,
400,
1069
],
"KMT2A-AF10": [
22,
1087,
1207
],
"KMT2A-AF9": [
147,
237,
2282
],
"KMT2A-EPS15": [
1233,
2063
],
"KMT2A-AFF1": [
429,
712,
1048,
1151,
1894,
2000,
2045,
2095
]
}- Does oncoprotein occupancy track blast percentage or oncoprotein mRNA level? Both are patient properties and neither is in the deposit.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
Across 27 KMT2A-rearranged patient samples spanning six fusion partners, the median score amplitude per partner spans 1.49x while the whole-cohort amplitude spans 6.51x and the widest single-partner group spans 5.44x.
GSE252378 - KMT2A oncoproteins induce epigenetic resistance to targeted therapies (171 GSMs, AutoCUT&RUN / AutoCUT&Tag, hg38 fragment BEDs and bigWigs) · Janssens DH, Henikoff S (Fred Hutchinson Cancer Center / Van Andel Institute) · NCBI Gene Expression Omnibus 2024What this evidence cannot show
Group sizes are 8, 7, 4, 3, 3, 2 - no formal test was run because none would be meaningful at these n, and none is claimed. The amplitude statistic itself is voided as a between-sample magnitude measure by this beat's own pre-registered noise-floor clause, so this comparison is reported on a statistic whose absolute level is not trusted; the called-promoter ranges are given alongside for that reason.
Evidence that argues against this (1)
- With n as low as 2 and 3 in three of the six groups, a real partner effect of moderate size would be invisible here. This is a statement that partner identity does not ORGANISE the range, which is a weaker claim than no effect, and the entry should not be cited for the stronger one. Janssens DH et al. 2024
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The source paper reports that the fusion partner and its expression level influence genome-wide binding-site selection, with partner-specific binding patterns over pro-B and HSC gene sets correlating with each rearrangement's prevalence in ALL versus AML.
MLL oncoprotein levels influence leukemia lineage identities · Janssens DH, Duran M, Otto DJ, Wu W, Xu Y, Kirkey D, Mullighan CG, Yi JS, Meshinchi S, Sarthy JF, Ahmad K, Henikoff S · Nature Communications 15:9341 2024What this evidence cannot show
That is an address claim measured on the authors' called intervals with dimensionality reduction, not an amount claim measured as a count.
Evidence that argues against this (1)
- The paper's title is 'MLL oncoprotein LEVELS influence leukemia lineage identities', so the authors explicitly do treat level as consequential; what they attribute to the partner is the binding pattern, and what they attribute to level is the lineage program. The reading here - partner organises address, something else organises amount - is compatible with the paper and is not offered as a disagreement. Janssens DH et al. 2024
the same coverage confound bounds any absolute read of that occupancy spread
the occupancy range this partitions
pushes the explanation for the amount onto the host cell rather than the fusion
- related to →KMT2A::MLLT1 (MLL-ENL), t(11;19)(q23;p13.3)
seven KMT2A::ENL samples, the widest within-partner range
two KMT2A::EPS15 samples, the smallest group
the six-partner cohort
q144 already found oncoprotein occupancy does not sort by fusion partner across 27 samples