The first same-condition H3K79me2 noise floor measured anywhere in this thread: 11.17% at D1 and 26.44% at the TOP 0.1% stratum, against 45.72% and 62.63% for H2AK119ub in the same arm, the same condition and the same pairing shape - 4.1x tighter. A mark floor transferred from H2AK119ub does not describe H3K79me2.
also called the first per-mark H3K79me2 floor · a transfer replaced by a measurement · marks do not share a floor
Until now the repeat-to-repeat wobble for one chemical tag on DNA was assumed to apply to the others. Measuring it directly for a second tag gives a number four times smaller, so the assumption was making that tag's results look far less reliable than they are.
The same thing, for a scientist
**Every mark floor in this thread has been transferred from H2AK119ub or imported from a factor ChIP.** This replaces one transfer with a measurement. **H3K79me2, MLLAF9 Control arm, rep1 (GSM8123373, `.filt.bam.tdf`) against rep2 (GSM8123387, bare `.tdf`), both Mar-2024 submission batch** - the `P_marker` shape, which is cycle 107's pairing shape: | stratum | TOP | D1 | D2 | D3 | D4 | D5 | D6 | D7 | D8 | D9 | D10 | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | **H3K79me2 SGC** | 26.44 | **11.17** | 7.19 | 4.91 | 1.86 | 0.62 | 1.66 | 3.68 | 7.45 | 11.85 | 14.51 | | H2AK119ub SGC, same shape | 62.63 | **45.72** | 13.40 | 2.56 | 5.08 | 9.56 | 14.32 | 18.04 | 23.06 | 31.36 | 26.30 | | ratio | 0.42x | **0.24x** | | | | | | | | | | **Shape- AND condition-matched**, which is the strict comparison: H3K79me2's D1 floor is **0.24x** H2AK119ub's, i.e. 4.1x tighter. The pre-registered M1 outcome - mark-specific if any two shape-matched floors differ by more than 2x - **fires**. **The floor is U-shaped in coverage rank** for both marks: high at TOP and D1, minimal in the middle deciles (0.62% at D5 for H3K79me2), rising again at D10. A single global floor is the wrong instrument for either mark, which is cycle 107's P1 confirmed a third time.
Bounded by n = 1, and said plainly
Only one of the three H3K79me2 conditions survived the universe gate. H3K79me2/DMSO aborted at 93.76% track agreement and H3K79me2/VTP at 92.62%; both abort under the pre-registered gate too. So this floor is a single matched pair, not a distribution, and no spread, median or interval is reported for it. The comparison against H2AK119ub is made at the same condition (SGC0946 5 uM) so the n=1 does not also import a condition mismatch.
A second condition would be worth more than any reanalysis of this one. The cheapest route is the coverage-restricted universe question in the open questions below.
What this changes for the thread
Cycles 100-107 score H3K79me2 claims against a floor borrowed from H2AK119ub or from a factor ChIP. In the one place both are now measured, the borrowed floor is 4.1x too wide, which means H3K79me2 effects previously written off as within-noise may not be. That is the opposite direction from cycle 107's correction, and it is the direction that re-opens claims rather than closing them.
It does not mean H3K79me2 measurements are 4.1x more trustworthy in general - the floor is arm-dependent for H2AK119ub and nothing here shows it is not for H3K79me2.
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 | H3K79me2 |
|---|---|
| arm | MLLAF9 Control |
| condition | SGC0946 5 uM |
| rep1 gsm | GSM8123373 |
| rep2 gsm | GSM8123387 |
| pairing shape | P_marker (.filt.bam vs bare, same submission batch) |
| d1 ratio to h2ak119ub | 0.24 |
| n conditions scored | 1 |
| n conditions refused | 2 |
| preregistered outcome fired | M1 - mark-specific |
floor PCT — the numbers behind the chart
{
"TOP": 26.44,
"D1": 11.17,
"D5": 0.62,
"D10": 14.51
}h2ak119ub same shape same condition
{
"TOP": 62.63,
"D1": 45.72
}refusals
{
"DMSO": "track agreement 93.76%",
"VTP": "track agreement 92.62%"
}- Does the H3K79me2 floor stay 4x tighter than H2AK119ub's in the NTvKO arm, where the H2AK119ub floor is itself 4x tighter? If the ratio holds across arms it is a property of the mark; if it collapses it is a property of this arm.
- Would restricting to bins both members cover at depth rescue the two refused conditions without changing the surviving one's floor? That is a single re-run and has not been done.
- E1weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.dataset
In GSE260742's MLLAF9 Control arm under SGC0946 5 uM, the same-condition H3K79me2 floor |R-1| measured between GSM8123373 (.filt.bam.tdf) and GSM8123387 (bare .tdf), same Mar-2024 submission batch, is 26.44% at the TOP 0.1% stratum and 11.17% at D1, against 62.63% and 45.72% for H2AK119ub in the same arm, condition and pairing shape - a D1 ratio of 0.24x. The other two H3K79me2 conditions failed the universe gate at 92.62% and 93.76% track agreement and were refused.
What this evidence cannot show
n = 1 CONDITION. This is a single matched pair and not a distribution; two of three conditions were refused and the refusals are recorded with their percentages. THE H3K79me2 ANTIBODY IS NOT RESOLVED - the token `antibod` occurs zero times in GSE260742's 158 sample records - so the mark identity rests on the depositor's titles and filenames, which cycle 107 verified are self-consistent across all 30 H3K79me2-titled samples but which is self-consistency, not correctness. SYSTEM: murine MLL-AF9 (KMT2A::MLLT3) myeloid blood cells, Mus musculus, mm10, one lab, one deposit - NOT the infant B-lineage KMT2A::AFF1 disease this graph is about. n = 3 conditions per arm, one cell system; the spread across conditions is reported so stability is visible and IS NOT A CONFIDENCE INTERVAL. 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 (2)
H3K79me2 is DOT1L's product; this is the floor for measuring it
- related to →Pinometostat (EPZ-5676) — the DOT1L inhibitor, and the most instructive failure in this field
the DOT1L inhibitor class whose H3K79me2 readouts this floor bounds
a second mark measured in the arm whose H2AK119ub floor is the widest
the mark whose floor WAS measured, showing why the H3K4me3 transfer is unsafe
the project H3K79me2 floor is a floor on shares and cannot be carried to an absolute scale