Oncofusion

start / graph / finding-q149-two-of-three-kmt2a-eps15-infant-all-cases-carry-irx1-at-aff1-level

Analysisweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q149-irx1-expression-across-partners

Two of three KMT2A::EPS15 infant ALL cases carry IRX1 at KMT2A::AFF1 magnitude, while the only KMT2A::EPS15 patient with occupancy data has no oncoprotein at IRX1 at all

also called q149 EPS15 IRX1 · t(1;11) IRX1-high

Two of the three babies in a public dataset whose KMT2A fusion joined the EPS15 gene had IRX1 switched on as strongly as any t(4;11) baby. The one EPS15 patient whose chromatin was profiled - a different child - had no fusion protein sitting on IRX1 at all.

The same thing, for a scientist

**n = 3 and n = 1. This entry names values and computes nothing, and it is recorded because it is the sharpest available instance of the beat's central ambiguity.** **Expression side, GSE68720 infant ALL, `mll_par: EPS15`:** | sample | IRX1 | per-sample floor | HOXA10 | IRX2 | PROM1 | karyotype | |---|---:|---:|---:|---:|---:|---| | GSM1679800 | **11.07** | 6.75 | 5.77 | <floor | 3.68 | `46,XY,T(1;11;3)(P32;Q23;P21)[4]/46,XY[20]` | | GSM1679782 | **10.11** | 6.41 | 6.63 | <floor | 6.81 | `46,XY,T(1;11)(P32;Q23)[8]/46,XY[12]` | | GSM1679759 | 3.24 | - | - | <floor | - | - | The highest KMT2A::AFF1 value in the same cohort is **11.37**. So these two are **at AFF1 magnitude**, not marginally above a floor. **Occupancy side, GSE252378, `KMT2A-EPS15`:** samples **R3595** and **R3652**, IRX1 within-sample shares **0.109x** and **0.047x** of an average promoter, against HOXA at **5.74x** and **2.94x** - deeply unoccupied in libraries that demonstrably report occupancy. **And q147's own manifest reading records R3595 and R3652 as the same patient sampled twice, so the EPS15 occupancy arm is ONE patient.** **Three things that make this more than an anecdote:** 1. **IRX1 is at 5p15.33 and EPS15 is at 1p32.** The translocation does not touch the IRX1 locus, so this is not a positional artefact of the rearrangement. 2. **Both cases are IRX2-negative**, unlike the AFF1 IRX1-high cases where IRX2 is ON in 9 of 21. So IRX1-high-without-IRX2 may be a different state from the AFF1 iALL-IRX1 state. 3. **HOXA10 in GSM1679782 is 6.63, above its 6.41 floor** - i.e. this is a HOXA-positive *and* IRX1-positive case, one of only 3 double-positives in the whole 97-sample cohort. The mutual exclusivity that holds in AFF1 does not hold here. **What it is not.** It is not a demonstration that IRX1 is expressed without the oncoprotein in the same cell. Those are different children, different institutions, different assays and different decades. It is a reason to run one experiment.

Go deeper
Why EPS15 and not another partner

Across both assays EPS15 is the only partner where the two sides point in opposite directions with non-trivial magnitude:

partnerIRX1 expression, infant ALLIRX1 occupancy, GSE252378
AFF121 of 48 ON3 of 8 occupied
MLLT1 (ENL)1 of 16 ON, max 6.23 marginal0 of 5-6
MLLT3 (AF9)0 of 60 of 3
EPS152 of 3 ON at 10.11 / 11.070 of 2 samples = 1 patient
AFDN, MLLT10, ELLno cases in GSE687200 of 4, 0 of 3, not in deposit

MLLT1 and MLLT3 agree across the two assays: off and unoccupied. EPS15 is the one disagreement, and it is also the partner with the thinnest occupancy arm in the entire deposit. That conjunction - the largest expression signal outside AFF1 sitting on the smallest occupancy arm - is exactly where a bounded absence is least informative.

KMT2A::EPS15 is a t(1;11)(p32;q23) fusion. fusion-kmt2a-eps15 exists in this corpus. Nothing in it records IRX1.

What would kill this
  1. The mll_par: EPS15 calls being wrong for either sample. Both carry a karyotype naming T(1;11)(P32;Q23), which is the independent handle and it agrees - so this is the least likely failure.
  2. The two IRX1-high cases being array artefacts. A single probe, 230472_at, at 10-11 log2 in two samples out of 97. A second probe or an RT-PCR would settle it and neither exists.
  3. Blast purity. A bulk sample at 20% or 40% blasts (both karyotypes show a substantial normal metaphase population - [4]/46,XY[20] and [8]/46,XY[12]) could be reporting IRX1 from non-blast cells. GSM1679800's karyotype shows the abnormal clone in only 4 of 24 metaphases. That is a serious concern for exactly this sample and it is the objection to answer first.
  4. n = 3. Two of three is 66.7% with a Wilson 95% interval of roughly 20-94%. It is compatible with anything.
Facts and figures
partnerEPS15
fusionfusion-kmt2a-eps15
cytogeneticst(1;11)(p32;q23)
irx1 locus5p15.33 - not on either chromosome involved in the translocation
aff1 max in same expression cohort11.37
absence classnever-measured
expression arm
{
  "cohort": "GSE68720",
  "n": 3,
  "values": {
    "GSM1679800": 11.07,
    "GSM1679782": 10.11,
    "GSM1679759": 3.24
  },
  "floors": {
    "GSM1679800": 6.75,
    "GSM1679782": 6.41
  },
  "irx2": "below floor in all three",
  "hoxa10": {
    "GSM1679800": 5.77,
    "GSM1679782": 6.63
  },
  "karyotypes": {
    "GSM1679800": "46,XY,T(1;11;3)(P32;Q23;P21)[4]/46,XY[20]",
    "GSM1679782": "46,XY,T(1;11)(P32;Q23)[8]/46,XY[12]"
  }
}
occupancy arm
{
  "deposit": "GSE252378",
  "n_samples": 2,
  "n_patients": 1,
  "samples": [
    "R3595",
    "R3652"
  ],
  "irx1_share": {
    "R3595": 0.1091,
    "R3652": 0.0472
  },
  "hoxa_share": {
    "R3595": 5.7389,
    "R3652": 2.9423
  },
  "irx1_call": "NOT_OCCUPIED in both"
}
measurements
[
  {
    "quantity": "IRX1_log2",
    "subject": "KMT2A::EPS15 infant ALL GSM1679800",
    "value": 11.07,
    "units": "log2 GPL570 series-matrix units",
    "n": 3,
    "universe": "GSE68720; this sample's own silent-panel max floor is 6.75; cohort AFF1 maximum is 11.37",
    "null": null,
    "note": "NAMED, NOT COMPUTED - n=3. Second case GSM1679782 at 10.11 (floor 6.41), third at 3.24"
  },
  {
    "quantity": "IRX1_within_sample_share",
    "subject": "KMT2A::EPS15 occupancy, GSE252378 R3595 / R3652",
    "value": 0.1091,
    "units": "x an average promoter in that library",
    "n": 2,
    "universe": "GSE252378 KMT2A-N AutoCUT&RUN, 19404 promoters, hg38; occupancy requires >=2.0x in both antibodies AND above the width-matched floor",
    "null": 0.2579,
    "note": "the null is that library's max cannot-be-occupied floor; second sample 0.0472. These two samples are ONE patient"
  }
]
How this could bite us back
  • KMT2A::EPS15 is a rare partner. If IRX1 status is used for enrolment, the children most likely to be IRX1-positive outside t(4;11) sit in a partner group too small for any trial to stratify on - so they would be enrolled or excluded by a partner-based proxy that is wrong for them specifically.
  • If the IRX1 in these samples is not blast-derived, an IRX1-directed agent selected on this signal would be aimed at a normal population in the marrow.
Still open
  1. Is IRX1 transcript and oncoprotein occupancy measurable in the SAME KMT2A::EPS15 blast? That is the single experiment this entry exists to motivate and it is written up as exp-q149-paired-rna-and-oncoprotein-occupancy-in-a-single-kmt2a-eps15-blast.
  2. Is the IRX1 signal in GSM1679800 blast-derived, given that its abnormal clone is 4 of 24 metaphases? A blast-count-matched re-analysis, or any sorted-population measurement, would decide whether this entry survives.
  3. Does IRX1-high-without-IRX2 define a different state from the AFF1 iALL-IRX1 subgroup, where IRX2 is ON in 9 of 21 IRX1-high cases?
Every source behind this entry (2)
  1. E1weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.dataset

    Of three KMT2A::EPS15 infant ALL cases in GSE68720, two carry IRX1 transcript at 11.07 and 10.11 log2 against per-sample silent-tissue floors of 6.75 and 6.41, values comparable to the highest KMT2A::AFF1 case in the same cohort (11.37), and both are IRX2-negative.

    PMID 22210879DOI
    What this evidence cannot show

    n=3, confidence weak, NAMED AND NOT COMPUTED - no rate, no test. Single array probe. GSM1679800's karyotype shows the t(1;11;3) clone in only 4 of 24 metaphases, so a bulk IRX1 signal in that sample may not be blast-derived; GSM1679782's is 8 of 20. Neither sample has a blast count. The partner call is the submitter's mll_par field, corroborated by the free-text karyotype.

    Evidence that argues against this (1)
    • Three KMT2A-germline infant ALL cases in the same cohort are also IRX1-ON, one at 10.37 - i.e. at the same magnitude as these EPS15 cases, with no KMT2A rearrangement at all. So a high IRX1 in an infant ALL sample is not by itself evidence of anything fusion-related, and the EPS15 cases may simply be two draws from that background. Kang H et al. 2012
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    The two KMT2A::EPS15 samples in GSE252378 carry IRX1 within-sample occupancy shares of 0.109x and 0.047x of an average promoter against HOXA shares of 5.74x and 2.94x, and are the same patient sampled twice, so the EPS15 occupancy arm is one patient.

    GSE252378 - AutoCUT&RUN profiling of KMT2A-rearranged paediatric leukaemia patient samples · GEO depositors, series GSE252378 · NCBI Gene Expression Omnibus 2024
    What this evidence cannot show

    Values re-read from analyses/q147-hoxa-irx1-across-partners/results_gene_scale147.json, which the king independently re-derived in cycle 131; not recomputed from the bigWigs in this beat. The same-patient status of R3595 and R3652 comes from q147's reading of the GEO sample descriptions ('paired with'), which I did not re-verify against the deposit this session. HOXA at 2.94x in R3652 is the lower of the two, so this pair is not the deposit's strongest occupancy demonstration either.

    Evidence that argues against this (1)
    • If R3652's HOXA at 2.94x indicates a weaker library, then a real but modest IRX1 occupancy could sit below the detection floor in that sample. The q147 floor rule requires >=2.0x in BOTH KMT2A-N antibodies AND above a width-matched cannot-be-occupied floor, and IRX1 at 0.047x is far from that - but a reader should know the EPS15 pair is at the weak end of the deposit. GEO depositors et al. 2024
Connected to