IRX1 sits in the 2-of-8 tier - two different patients, not the same one twice - and is not in the recurrent core
also called q152 IRX1 tier placement
IRX1, the gene this project has been chasing for several cycles, is switched on in two of the eight KMT2A::AFF1 samples at promoter scale. Those two are two different children, which is the good news. But forty-seven other genes share the same two-of-eight pattern, so IRX1 is not among the seventeen most recurrent.
The same thing, for a scientist
At promoter scale under the q147/q151 rule, IRX1 is occupied in exactly 2 of 27 GSE252378 samples: **p279 and p454**, both KMT2A::AFF1 and **two distinct patients** (p279's partner sample p318 is the same child and is negative for IRX1). IRX1 therefore sits in the 2-of-8 tier of the AFF1-only set, which contains 47 promoters, and is not in the >=3-of-8 recurrent core of 17. The full tier table is 268 / 47 / 9 / 7 / 1 / 0 / 0 / 0 for 1 through 8 samples: 80.7% of the 332 are singletons, and the single 5-of-8 promoter is GBA.
The tier table, complete
| occupied in N of the 8 AFF1 samples | promoters | notable |
|---|---|---|
| 5 | 1 | GBA |
| 4 | 7 | BLNK, CCP110, MAGI3, MTFR1, POU2AF1, SYNE3, WT1 |
| 3 | 9 | AMZ2, EPHA7, MATR3, MED28, NOP58, PBLD, SPRED1, TBX21, UHRF1BP1L |
| 2 | 47 | includes IRX1 (p279, p454) |
| 1 | 268 | singletons, 80.7% of the 332 |
The dispatch brief for this beat listed eight names for the nine-promoter 3-of-8 tier; the ninth is UHRF1BP1L 1.
What the two-patient detail is worth
It is worth more than the tier position. This project has repeatedly had to correct counts that treated p279 and p318 as two observations. IRX1's two hits are p279 and p454, which are different children, so IRX1's 2-of-8 is genuinely 2 patients - unlike NOP58, whose 3-of-8 is 2 patients 1. Against the q151 expectation that an arbitrary promoter is called occupied in 2.73 of 27 samples, IRX1 at 2 of 27 is below the per-promoter chance rate in total, with both of its hits happening to fall in one arm.
| status | run |
|---|
tier table samples — the numbers behind the chart
{
"1": 268,
"2": 47,
"3": 9,
"4": 7,
"5": 1,
"6": 0,
"7": 0,
"8": 0
}measurements
[
{
"quantity": "n_aff1_samples_occupied",
"subject": "IRX1 2 kb promoter in the KMT2A::AFF1 arm",
"value": 2,
"units": "samples of 8 (2 of 7 patients)",
"n": 27,
"universe": "27 GSE252378 samples, two-antibody 2.0x plus per-library floor, 2 kb TSS-centred windows",
"null": 2.73,
"note": "p279 and p454, two distinct patients; null is q151's per-promoter expectation across all 27 samples"
},
{
"quantity": "singleton_share_of_the_AFF1_only_set",
"subject": "the 332 AFF1-only promoters",
"value": 0.807,
"units": "fraction occupied in exactly 1 of 8 AFF1 samples",
"n": 332,
"universe": "the 332 promoters occupied in >=1 AFF1 and 0 non-AFF1 samples",
"null": null,
"note": "268 of 332"
}
]- Do the 47 members of the 2-of-8 tier include any other gene with published KMT2A::AFF1 evidence, as the 3-of-8 tier does via EPHA7? The tier was not annotated against the literature in this beat.
- Is IRX1's promoter-scale 2-of-8 versus gene-body-scale 3-of-8 a window-size effect or a different sample? The sample identities at gene-body scale were not compared here.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
At 2 kb promoter scale under the two-antibody 2.0x plus per-library floor rule, IRX1 is occupied in exactly p279 and p454 of the 27 GSE252378 samples - two distinct KMT2A::AFF1 patients - placing it in the 47-promoter 2-of-8 tier of the 332 AFF1-only set; the complete tier table is 268 / 47 / 9 / 7 / 1 / 0 / 0 / 0 and the ninth member of the 3-of-8 tier, absent from the dispatch brief's example list, is UHRF1BP1L.
Q152 tier table for the 332 AFF1-only promoters · Iqbal Lab oncofusion knowledge graph, beat Q152 · this project 2026What this evidence cannot show
Promoter scale only. q147 reports IRX1 fusion occupancy in 3 of 8 KMT2A::AFF1 samples at gene-body scale (5,572 bp), so the 2-of-8 here is a statement about the 2 kb window, not a contradiction of q147.
Evidence that argues against this (1)
- Tier position is a function of window size. q147 reports IRX1 occupancy in 3 of 8 AFF1 samples at 5,572 bp gene-body scale, which would put it in the 3-of-8 tier alongside EPHA7. Whether IRX1 belongs in the recurrent core therefore depends on a scale choice this beat did not make and cannot adjudicate. Iqbal Lab oncofusion knowledge graph et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
q147 reports IRX1 fusion occupancy confined to KMT2A::AFF1 across six partners at gene-body scale, and q149 subsequently scoped that finding by noting the deposit holds no megakaryoblastic sample.
ana-q147-irx1-fusion-occupancy-is-confined-to-kmt2a-aff1 and absence-q149-... (this graph) · Iqbal Lab oncofusion knowledge graph, beat Q152 · this project 2026What this evidence cannot show
Cross-reference to other beats' entries, not re-derived prose; the numeric IRX1 occupancy count was independently re-derived here and matches q151's 2 of 27 at promoter scale.
Evidence that argues against this (1)
- q149 recorded that this deposit holds no megakaryoblastic sample, so the 'confined to KMT2A::AFF1' finding IRX1's tier position sits inside is already scoped by an absent subgroup. A tier table built on 27 samples cannot speak to a subgroup none of them belongs to. Iqbal Lab oncofusion knowledge graph et al. 2026
- supports →IRX1 — Iroquois homeobox 1, the gene that names the HOXA-negative half of infant t(4;11) ALL
places IRX1 in the 2-of-8 tier, in two distinct patients, outside the recurrent core
- related to →IRX1 fusion occupancy is confined to KMT2A::AFF1: 0 of 18 samples across five other partners
promoter-scale 2 of 8 beside q147's gene-body-scale 3 of 8; a window-size difference, not a contradiction
- related to →HOXA occupancy is the partner-invariant: occupied in 17 of 18 non-AFF1 KMT2A-r samples, 2.94x to 29.1x
HOXA is the partner invariant and IRX1 is not; this adds the tier position
IRX1's tier position is what the recurrence structure of the 332 looks like
IRX2 is occupied in 0 of 27 samples under the same rule