Oncofusion

start / graph / ana-q120-in-the-same-mouse-nuclei-mll1-diverges-from-flag-at-myb-and-agrees-at-meis1

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q120-h3k79me2-second-line-and-the-mll1-epitope

Q117's prediction, run: an MLL1 N-terminal ChIP and a FLAG ChIP of the tagged fusion in the same murine cells diverge 1.83-fold at Myb (0.6645 against 0.3623) and agree to within 4% at Meis1 (0.112 against 0.1171) - the direction is exactly as predicted and the pre-registered numeric threshold for Myb still fails

also called the shared-epitope test · Myb diverges, Meis1 agrees

Two years of argument about whether a popular antibody hides the drug's effect can be settled at two genes, because one mouse dataset ran that antibody and a clean tag on the cancer protein in the same cells. At one gene the two disagree by nearly two-fold; at the other they are the same number.

The same thing, for a scientist

**The prediction, quoted verbatim from `ana-q117-the-published-named-locus-pattern-reproduces-for-meis1-and-jmjd1c-and-breaks-for-mef2c-and-myb`:** *'if wild-type MLL1 masks fusion loss at MYB, then an MLL1n ChIP and a FLAG ChIP in the SAME cells should differ at MYB and agree at MEIS1. GSE260742 contains an MLL1 ChIP in the same MLLAF9 line under the same three conditions (GSM8123395/6/7) and this beat did not download it.'* This is that test [E1]. | mm10 window | n bins | FLAG (the fusion) | MLL1 (A300-086A) | MLL1 / FLAG | |---|---:|---:|---:|---:| | **Myb** chr10:21,119,930-21,162,984 | 23 | **0.3623** | **0.6645** | **1.83x** | | **Meis1** chr11:18,875,428-19,020,969 | 74 | **0.1171** | **0.1120** | **0.96x** | Floor-normalised against each track's own aggregate ratio over ~500,000 mid-occupancy bins (FLAG 0.9295, MLL1 0.9165): Myb 0.390 and 0.725; Meis1 0.126 and 0.122. **Both halves of Q117's prediction are confirmed in direction.** At Myb the shared-epitope antibody reports 1.83 times as much retained signal as the tag does. At Meis1 the two reagents are indistinguishable. **The pre-registered numeric criterion M1 nevertheless DOES NOT FIRE**, because it demanded `R_MLL1 >= 0.80` at Myb (band 0.72-0.88) and the observed value is 0.6645. That threshold was Q118's P5, not Q117's - Q117 stated a direction and named no number - and it is quoted here as it was pre-registered rather than moved. **M2 FIRES** (0.112 < 0.45). **The instrument reproduced its own controls exactly.** Q117's stored FLAG values for these two windows are 0.362 and 0.117 and this run returns 0.3623 and 0.1171; Q117's stored FLAG track totals, 528,768,301 and 532,157,522 coverage-bases, come back identical; and the mm10 refGene windows were re-fetched live from UCSC in this beat and matched Q117's stored windows to the base pair.

Go deeper
What this does to a committed conclusion of the previous cycle

ana-q118-q117s-myb-prediction-fires-in-rs4-11-and-fails-in-the-matched-fusion-line concludes in its own title that 'the MYB difference is a partner difference and not an antibody artefact', reasoning from human data alone: the MYB divergence appears in RS4;11 and not in MOLM13, which carries the mouse's own fusion.

The mouse measurement contradicts the second half of that. In the murine MLL-AF9 line - the matched fusion - an MLL1 N-terminal ChIP and a FLAG ChIP of the fusion in the same cells differ at Myb by 1.83-fold. There is an antibody artefact at Myb, measured against ground truth, in exactly the fusion where Q118 concluded there was not one.

This is reported, not edited. The two claims are also not strictly incompatible - Q118's is about why RS4;11 and human MOLM13 differ from each other, mine is about whether the reagent diverges from the tag at all - and both can hold. What cannot hold is the words 'not an antibody artefact', because the artefact is now measured at 1.83-fold.

A correction to the mouse-side floor that Q118 flagged as a stopgap

Q118's cross-species locus table states: 'the mouse floor is Q117's Gapdh at 0.508, which is two genes and is the best available mouse-side normaliser.' Q117's own memo named a proper floor as its open gap.

This beat built it. Applying Q118's own DEVIATION-2 definition - the aggregate share ratio over bins in DMSO percentile 50-90 - to the murine data gives FLAG 0.9295 over 505,306 bins and MLL1 0.9165 over 491,926 bins. The Gapdh stopgap is 1.83-fold too low, so every mouse-side floor-normalised figure in that table is inflated by 1.83x:

locusmouse FLAG rawQ118's /0.508corrected /0.9295
Meis10.1170.2300.126
Myb0.3620.7130.390
Jmjd1c0.1670.3290.180
Hoxa90.6601.2990.710
Mef2c0.8701.7130.936
Gapdh0.5081.0000.547

Two of Q118's readings change sign. Mouse Hoxa9 and Mef2c were recorded as above their floor, i.e. relative gains; corrected, both are below 1 and are modest losses. And Q118's headline agreement at MEIS1 - 'RS4;11 0.245 against mouse 0.230, agreement to 6%' - becomes 0.245 against 0.126, a 1.9-fold difference. The MEIS1 conclusion survives in direction and its precision does not. Reported for the coordinator, not edited.

Facts and figures
scriptanalyses/q120-h3k79me2-second-line-and-the-mll1-epitope/arm2-mouse-epitope.py
statusrun
result summaryMyb: FLAG 0.3623, MLL1 0.6645, divergence 1.83x. Meis1: FLAG 0.1171, MLL1 0.1120, agreement to 4%. M1 does not fire at its pre-registered 0.80 threshold; M2 fires.
cell linemurine FLAG-MLLAF9
fusionMLL-AF9 (Kmt2a::Mllt3), FLAG-tagged, retroviral
drugVTP50469 500 nM, 48 h, against a 72 h DMSO control
genome buildmm10, verified from each file's own chromosome extents
bin bp2000
n bins universe1227708
antibodyMLL1 Bethyl A300-086A; FLAG NEB 14793S DYKDDDDK D6W5B
mouse floor aggregate flag0.9295
mouse floor aggregate mll10.9165
shown for fusionMLL-AF9 in a murine FLAG-tagged model - the only system in this project's evidence base where a KMT2A fusion's own occupancy is measured without an antibody that also sees wild-type MLL1
absence classmeasured-not-analysed
absence class noteQ117 named this exact test as its own next step and recorded that it did not download the files. GSM8123395/96/97 have been public in GSE260742 since November 2025. Two downloads and one script closed it; no wet work.
absence class assigned2026-09-23 (Q120)
inputs
[
  "GSM8123395 murine MLLAF9 MLL1 DMSO",
  "GSM8123397 murine MLLAF9 MLL1 VTP50469 500 nM",
  "GSM8123359 FLAG DMSO and GSM8123361 FLAG VTP (binned by Q117, reused)",
  "GSM8549823 input",
  "ENCODE mm10 blacklist v2",
  "UCSC mm10 refGene, live"
]
How this could bite us back
  • If an MLL1-N ChIP over-reports retention by 1.83-fold at a locus where a fusion is genuinely being lost, then any clinical pharmacodynamic assay built on that antibody will under-call target engagement at exactly the loci where the drug is working.
  • And the failure is locus-specific rather than global, which is worse for an assay: a single correction factor cannot fix it, and which loci are affected depends on where wild-type MLL1 sits in that patient's cells.
Still open
  1. Why Myb and not Meis1? If wild-type MLL1 occupies Myb and not Meis1 in these cells, a wild-type-specific readout would show it directly - an MLL1 C-terminal ChIP, or the MLL1 track in a fusion-free cell of the same lineage. GSE260742 has neither.
  2. Does the divergence scale with the wild-type MLL1 share at a locus? The 23-bin Myb window and the 74-bin Meis1 window are two points. Running the FLAG-versus-MLL1 ratio across all ~500,000 mid-occupancy bins would give the distribution and was not done here.
Every source behind this entry (1)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    In murine FLAG-MLLAF9 cells (GSE260742), within-sample share ratios VTP50469/DMSO over mm10 gene-body -5 kb/+2 kb windows in a universe of 1,227,708 2 kb bins (INPUT_B2 > 0, ENCODE mm10 blacklist v2 removed) are: Myb FLAG 0.3623 and MLL1 0.6645 (23 bins); Meis1 FLAG 0.1171 and MLL1 0.1120 (74 bins). Genome-wide mid-occupancy floors are FLAG 0.9295 (505,306 bins) and MLL1 0.9165 (491,926 bins). The FLAG values and track totals reproduce Q117's stored figures exactly and the mm10 windows were re-fetched live from UCSC and matched Q117's stored windows.

    What this evidence cannot show

    VTP50469 here is **500 nM**, against **330 nM** in every GSE127507 sample - these are NOT matched doses. Within GSE260742 the VTP arms are 48 h and the DMSO controls 72 h, in BOTH the FLAG and the MLL1 arm, so the FLAG-versus-MLL1 comparison is exposure-matched to itself while neither arm is internally symmetric. The FLAG files are `*_FLAG_F3_*.filt.bam.tdf` and the MLL1 files `*_MLL1_F3_NT_*.tdf`: both are batch-1 low-artefact files (1.6-1.9% of coverage in blacklist intervals) but the naming implies different processing and possibly a different culture, and the depositors' titles call both 'MLLAF9 Control'. The murine fusion is a RETROVIRALLY OVER-EXPRESSED FLAG-MLLAF9 construct, so its fusion-to-wild-type ratio has no reason to match an endogenous human fusion. NEB 14793S DYKDDDDK (D6W5B) reads the FLAG-tagged MLLAF9 construct only, so it is ground truth for the fusion and blind to wild-type MLL1. Neither it nor the MLL1 antibody can see a fusion molecule whose FLAG tag has been proteolysed, which would inflate the MLL1/FLAG ratio and mimic masking. Bethyl A300-086A, deposited as 'anti-MLL1n', an MLL1 N-TERMINUS epitope. The N-terminus is retained in every KMT2A fusion, so this track is the sum of the fusion and wild-type MLL1 in unknown proportion and CANNOT distinguish them. The immunogen residue range was not verified: three vendor routes were tried this session (fortislife.com two URL forms, bethyl.com/product/A300-086A) and all returned HTTP 404, so 'N-terminus' rests on the depositors' own target label and on the standard fact that MLL-FPs lack the C-terminal SET domain. n = 1 per antibody per condition per line in both series. No number in this beat carries a spread and none is precise.

    Evidence that argues against this (1)
Connected to