H3K4me3 is the one mark in GSE260742 with no declared second channel at all - six samples, no replicate suffix anywhere - so no same-condition pairing exists and its noise floor remains a transfer from another mark whatever else is measured.
also called H3K4me3 has no replicate · the transfer that cannot be replaced
Of the four chemical tags this dataset measures, one was never repeated at all. Its reliability therefore has to be guessed from the others, and no amount of reanalysis can change that.
The same thing, for a scientist
**Registered in advance, before any number was computed**, because a negative found after the fact is weaker than one predicted. Cycle 107's parse of all 158 GSE260742 sample records records H3K4me3 with `samples_titled: 6`, `rep_suffixes_declared: []` and `declares_a_second_channel: false` - the only one of the four marks for which all three hold. H2AK119ub declares suffixes `1, 2, 3, paired`; H3K27me3 and H3K79me2 each declare `2`. **Consequence.** H3K4me3 has no same-condition pair in this deposit, so its floor cannot be measured here by any method, and every H3K4me3 verdict in cycles 100-107 - including the internal control that carried the most weight, H3K4me3 running **opposite** in sign to the two repressive marks across the same bins - is scored against a floor imported from a different mark. **And this beat has just shown that such an import is unsafe**: the one place two marks' floors are both measured, they differ 4.1-fold. The H3K4me3 sign-reversal result does not depend on a floor, which is exactly why it is the most robust thing in the thread; the magnitude statements around it do.
Absence class, and what would close it
absence_class is not-measured - the strongest class, because no analysis can recover it. Closing it needs either a second H3K4me3 ChIP from the same depositors, or a different public deposit with H3K4me3 same-condition replicates in a comparable system, binned and scored the same way. The second is a compute task and nobody has scoped it.
THE RENORMALISATION BOUND, AND IT APPLIES TO EVERY NUMBER IN THIS ENTRY. Within-sample shares sum to 1 and GSE260742 has NO USABLE SPIKE-IN (Q117 verified zero hits for the entire spike-in vocabulary across all 419,824 characters of its 158 sample records, and no scale factor is deposited), so a purely local change leaks into every other stratum by arithmetic. A FLOOR COMPUTED THIS WAY CANNOT DISTINGUISH 'the bulk moved' from 'the peaks moved and the shares renormalised'. A TIGHTER NUMBER IS NOT THEREBY A CLEANER ONE.
| mark | H3K4me3 |
|---|---|
| deposit | GSE260742 |
| samples titled | 6 |
| declares a second channel | false |
| absence class | never-measured |
| floor status | transferred from another mark |
| absence class note | FOLDED 2026-09-24: this entry originally carried the non-vocabulary token 'not-measured', which is the truer description and is preserved here; 'never-measured' is the nearest controlled token. |
rep suffixes declared
[]
other marks
{
"H2AK119ub": [
"1",
"2",
"3",
"paired"
],
"H3K27me3": [
"2"
],
"H3K79me2": [
"2"
]
}- Which public deposit has H3K4me3 same-condition replicates in a KMT2A-rearranged myeloid or B-lineage system, and what would binning it cost?
- How many cycles-100-107 H3K4me3 statements are magnitude claims (floor-dependent) rather than sign claims (floor-independent)? Nobody has counted.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
GSE260742 contains six H3K4me3-titled samples and declares no replicate suffix for any of them, so it contains no same-condition H3K4me3 replicate pair. H2AK119ub declares suffixes 1, 2, 3 and 'paired'; H3K27me3 and H3K79me2 each declare '2'.
What this evidence cannot show
ABSENCE OF A DECLARED SUFFIX IS NOT ABSENCE OF A REPLICATE - cycle 107's own U3, and this deposit is the reason it exists, since it names its first replicate by saying nothing. Two H3K4me3 samples of one condition could exist under two titles that do not look like a replicate pair. What is established is that none is DECLARED. THE RENORMALISATION BOUND, AND IT APPLIES TO EVERY NUMBER IN THIS ENTRY. Within-sample shares sum to 1 and GSE260742 has NO USABLE SPIKE-IN (Q117 verified zero hits for the entire spike-in vocabulary across all 419,824 characters of its 158 sample records, and no scale factor is deposited), so a purely local change leaks into every other stratum by arithmetic. A FLOOR COMPUTED THIS WAY CANNOT DISTINGUISH 'the bulk moved' from 'the peaks moved and the shares renormalised'. A TIGHTER NUMBER IS NOT THEREBY A CLEANER ONE.
Evidence that argues against this (1)
the mark whose floor WAS measured, showing why the H3K4me3 transfer is unsafe
- related to →KMT2A / MLL1 — the gene that breaks
H3K4me3 is the wild-type KMT2A SET domain's product