Oncofusion

start / graph / ana-q152-collapsing-p279-and-p318-to-one-patient-removes-nop58-from-the-recurrent-core

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q152-aff1-only-recurrent-core

p279 and p318 are one patient in two lineage states, and collapsing them removes exactly one gene - NOP58 - from the recurrent core

also called q152 C1 patient-level collapse

Two of the eight KMT2A::AFF1 samples come from the same child, sampled twice as the disease changed lineage. Counting them as one patient, the eight samples are seven patients, and one of the seventeen genes turns out to have been counted twice in the same child.

The same thing, for a scientist

Pre-registered escape condition C1. Treating p279 and p318 as one patient (occupied = occupied in either state) reduces the 8-sample arm to 7 patients and the tier table to 269 / 47 / 9 / 6 / 1 / 0 / 0 for 1 through 7 patients. The recurrent core falls from 17 to 16: **NOP58** drops out, because its three sample-level hits are p279, p318 and p454, i.e. two patients. SYNE3 (p247, p279, p318, p454) and MTFR1 (TB051939, p247, p318, p454) also involve p279 or p318 but retain three or more patients. C1 does not fire as a verdict-changing condition - 16 is not below the pre-declared threshold of 10 - but the count that should be quoted is 16 of 7 patients, not 17 of 8 samples.

Go deeper
Which genes touch the shared patient

p279 hits 13 of the 17 core promoters and p318 hits 3 (SYNE3, MTFR1, NOP58). The only gene whose recurrence depends on counting p279 and p318 separately is NOP58, at p279 + p318 + p454 = 2 patients 1. NOP58 is also the gene whose AFF1-exclusivity rests entirely on one library's noise floor (ana-q152-the-per-library-floor-spans-zero-to-four-times-an-average-promoter-and-nop58-rests-entirely-on-it), so the same promoter fails two independent sanity checks.

The number to quote

16 promoters occupied in >=3 of 7 KMT2A::AFF1 patients and 0 of 19 non-AFF1 samples. Any count phrased as 'of 8' invites the reader to treat two lineage states of one child as two independent observations, which is the specific error this project's q142 work already documented for this pair.

Facts and figures
scriptanalyses/q152-the-aff1-only-recurrent-promoter-core/step2.py
statusrun
result summary17 of 8 samples becomes 16 of 7 patients; NOP58 is the only exit.
measurements
[
  {
    "quantity": "count_recurrent_partner_exclusive_promoters_patient_level",
    "subject": "promoters occupied in >=3 of the 7 KMT2A::AFF1 PATIENTS and 0 of the 19 non-AFF1 samples",
    "value": 16,
    "units": "promoters",
    "n": 19404,
    "universe": "19,404 refGene 2 kb promoters, p279 and p318 collapsed to one patient by OR",
    "null": 9.1,
    "note": "sample-level count is 17 of 8; the null quoted is the sample-level permutation mean and was not recomputed at patient level"
  }
]
Still open
  1. What is the patient-level permutation null? The 10,000-draw permutation was run at sample level only, so the 16 is compared against a sample-level null of 9.1, which is not exact.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    Collapsing p279 and p318 to one patient reduces the AFF1 arm to 7 patients and the >=3 recurrent core from 17 promoters to 16, with NOP58 the sole exit because its hits are p279, p318 and p454 - two patients; the patient-level tier table is 269 / 47 / 9 / 6 / 1 / 0 / 0 for 1 through 7 patients.

    Q152 escape-condition C1, patient-level collapse of p279 and p318 · Iqbal Lab oncofusion knowledge graph, beat Q152 · this project 2026
    What this evidence cannot show

    The collapse uses an OR (occupied in either lineage state counts as occupied in that patient). An AND version would be stricter and was not run.

    Evidence that argues against this (1)
    • Collapsing by OR is the conservative choice for a recurrence count but it is not obviously the right biology: q142 measured that p279 and p318 share no more fusion-target promoters than two different patients do, which is an argument that they should be counted as two observations, in which case 17 of 8 is the correct figure and this correction is over-cautious. Iqbal Lab oncofusion knowledge graph et al. 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    p279 and p318 are two lineage states of a single patient in GSE252378, and this project has previously measured that their fusion-target promoter sets overlap no more than two different patients' do.

    What this evidence cannot show

    That earlier measurement means the two states are not redundant observations either; collapsing them is the conservative choice for a recurrence count, not a statement that they are identical.

    Evidence that argues against this (1)
    • The q142 result cuts against the collapse it is cited beside. If two lineage states of one patient are as different as two patients, the patient-level count understates recurrence rather than correcting it, and the honest position is that neither 17 nor 16 is the single right number. Iqbal Lab oncofusion knowledge graph et al. 2026
Connected to