An expression subgroup is not an occupancy subgroup — and this project keeps reading one as the other
Saying a gene is switched on is not the same as saying a particular protein is sitting on it. The famous HOXA9-versus-IRX1 split in infant leukaemia was measured by reading RNA, not by looking at what is bound to the DNA — and those can come apart.
The same thing, for a scientist
The published iALL-HOXA9 / iALL-IRX1 dichotomy rests entirely on transcript abundance [E1]. An occupancy claim — that the oncoprotein binds locus X — is a different measurement with different failure modes, and the two dissociate routinely: a locus can be occupied and silent (poised), or transcribed with the factor long gone. This entry exists because a dispatch brief in cycle 126 described the split as though it were established at chromatin, and because the single occupancy experiment in the IRX1 literature [E2] points in the **opposite direction** from the way the split is usually paraphrased.
The two claims, side by side
| what was measured | in what system | direction | |
|---|---|---|---|
| Symeonidou 2021 1 | RNA abundance of HOXA9/HOXA10 vs IRX1, Kallisto/DESeq2 | patient blasts, two published RNA-seq cohorts | HOXA9 and IRX1 are mutually exclusive |
| Kühn 2016 2 | ChIP occupancy of IRX1 protein at HOXA9/HOXA10 promoters | HEK293T with inducible transgenes | IRX1 sits ON the fusion's targets and shuts them down |
Note what 2 is not: it is not a measurement of the fusion occupying the IRX1 locus. Nobody had made that measurement in patient material until this beat did 3.
Why the conflation is dangerous here specifically
If you read the expression split as an occupancy split, you predict that the oncoprotein has two alternative addresses — HOXA in one subgroup, IRX1 in the other — and that predicts a menin-inhibitor response in both. If instead IRX1 is high because IRX1 protein is repressing the fusion (Kühn's direction), then the IRX1 subgroup is one where the fusion's own output is already partly suppressed, and the therapeutic logic is completely different.
This beat's measurement is compatible with the first reading and does not exclude the second, because occupancy cannot tell you which way causality runs 3.
The general form, for reuse
Before importing a subgroup, a signature or a 'target' into this graph, write down which of these was measured: transcript abundance, protein abundance, factor occupancy, histone mark, accessibility, or dependency. They are six different experiments. The corpus has a standing failure of treating them as interchangeable, and rule 86 (shown_for_fusion must name which fusion) is the same disease in a different organ.
applies to
[ "gene-irx1", "gene-hoxa9", "fusion-kmt2a-aff1" ]
measurement kinds
[ "transcript abundance", "protein abundance", "factor occupancy", "histone mark", "accessibility", "genetic dependency" ]
- Treating an expression subgroup as an occupancy subgroup licenses the inference that a chromatin-directed drug will work in both halves, which is precisely the inference that has never been tested in iALL-IRX1.
- How many entries in this corpus assert a chromatin fact whose only source is an expression experiment? Nobody has counted.
- Is there any infant t(4;11) sample anywhere with matched RNA-seq AND oncoprotein CUT&RUN? That single dataset would settle whether the expression subgroup and the occupancy subgroup are the same partition.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The iALL-HOXA9 / iALL-IRX1 split was defined by RNA-seq reanalysis with Kallisto, tximport and DESeq2, with PCA, Spearman correlation and Fisher's exact test — and contains no chromatin experiment.
HOXA9/IRX1 expression pattern defines two subgroups of infant MLL-AF4-driven acute lymphoblastic leukemia · Symeonidou V, Ottersbach K · Exp Hematol 2021What this evidence cannot show
Established as a statement about the paper's methods, which were read in full this cycle. The clinical significance of the subgroups is a separate and weaker claim.
Evidence that argues against this (1)
- Kuhn 2016 DOES contain occupancy experiments about the same axis (IRX1 and MLL-C ChIP at HOXA9/HOXA10 promoters), so the IRX1/HOXA literature as a whole is not purely transcriptional - only the paper that DEFINED the subgroups is. Reading this entry as "there is no chromatin data on IRX1" would overstate it. Kuhn A et al. 2016
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
The only occupancy experiment in the IRX1/KMT2A literature measures IRX1 protein binding at HOXA9/HOXA10 promoters in HEK293T, not fusion binding at IRX1.
The IRX1/HOXA connection: insights into a novel t(4;11)-specific cancer mechanism · Kuhn A, Loscher D, Marschalek R · Oncotarget 2016What this evidence cannot show
HEK293T with transfected transgenes. The ChIP antibodies were IRX1 (Abcam ab72642) and MLL-C (Active Motif 61295) with a goat IgG control. No primary leukaemia ChIP exists in the paper.
Evidence that argues against this (1)
- Absence of a fusion-at-IRX1 experiment is bounded by title and abstract searching plus three full texts. A supporting ChIP panel in a paper whose abstract never names IRX1 would be invisible to that search, which is the standing rule-80 failure mode for measurements. Janssens DH et al. 2024
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
Fusion occupancy at the IRX1 locus in primary KMT2A::AFF1 patient material was first measured in this beat, from the GSE252378 deposited bigWigs.
q145 re-analysis of GSE252378: IRX1 and HOXA oncoprotein occupancy across seven KMT2A::AFF1 patients (analyses/q145-irx1-in-hoxa-negative-infants) · Iqbal Lab oncofusion knowledge graph, beat q145 · internal analysis over deposited GSE252378 bigWigs 2026What this evidence cannot show
A claim of priority bounded by what I searched: PubMed and the three papers' citation contexts. A supporting figure in a paper whose abstract does not mention IRX1 would be invisible to that search (rule 80). Every number is a WITHIN-SAMPLE share: an interval's mean coverage divided by that same library's mean promoter coverage, so 1.0 = an average promoter in that library. **GSE252378 has no spike-in and no drug arm**, so no between-sample absolute statement is licensed; the only cross-sample quantity used is a ratio of two loci measured inside one library, where the unknown per-library scale cancels.
Evidence that argues against this (1)
- A claim of priority is the weakest kind of claim in this corpus and the easiest to falsify with one citation; it is made here only to mark that no such measurement was found, not to assert that none exists. Janssens DH et al. 2024
- related to →IRX1 — Iroquois homeobox 1, the gene that names the HOXA-negative half of infant t(4;11) ALL
the worked example that prompted the concept
- part of →Plain-English: chromatin and epigenetics — the sticky notes on the DNA, and which ones matter here
a measurement-kind distinction inside the chromatin explainer
the clearest single datum that the expression dichotomy is not an occupancy dichotomy
whether mutual exclusivity holds at single-cell occupancy
- related to →HOXA9 — required for survival of established MLL-rearranged leukemia, and dispensable for starting it
HOXA9 expression and HOXA9 occupancy are different measurements
- part of →What is a fusion gene?
a measurement-kind caveat on fusion target claims
the two assays are in different patients and give different answers; this is the worked case
the expression subgroup here is a morphology, and the occupancy cohort contains none of it
an expression subgroup can also be a MORPHOLOGY subgroup wearing a fusion label
- ← depends onIRX2 is not fusion-occupied at promoter scale in GSE252378: 0 of 27 samples, against a null of 2.73
GSE252378 and GSE68720 are different patients; no expression conclusion is transferred
a partner-exclusive occupancy set is not an occupancy subgroup either