Oncofusion

start / graph / ana-q118-menin-occupancy-in-rs4-11-falls-to-one-fifth-and-reproduces-q117s-murine-numbers

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q118-menin-at-the-human-index-fusion

In human RS4;11 (KMT2A::AFF1), menin occupancy at its own top 0.1% of 2 kb bins falls 14.9x to 2.0x over the track median and its share of total coverage falls to 0.200 - reproducing Q117's murine FLAG-MLLAF9 figures (15.9x to 3.0x, ratio 0.179) to within 1.3-fold on all five statistics

also called menin eviction in RS4;11 · the 80% menin occupancy collapse · Q117 replication, menin arm

In cells from the leukaemia this lab studies, the menin drug strips about four fifths of menin off the places on DNA where menin normally sits hardest. The same measurement made last cycle in a mouse leukaemia, using a completely different tag and dataset, gave almost exactly the same number.

The same thing, for a scientist

**In human RS4;11 cells (KMT2A::AFF1, t(4;11), an ADULT-derived B-ALL line), at 330 nM VTP-50469 for 3 days, menin's own top 0.1% of 2 kb genomic bins lose about 80% of their relative occupancy.** Measured on 2 kb bins re-derived from the deposited `.deduped.tdf` files, universe = 1,389,514 bins with input > 0 and not in the ENCODE hg19 blacklist v2, held set = the top 1,390 bins of MENIN DMSO [E1]: | statistic | RS4;11 MENIN (this beat) | murine FLAG-MLLAF9 (Q117, GSE260742) | agreement | |---|---:|---:|---:| | enrichment over the track's own genome median, DMSO -> VTP | **14.91x -> 1.96x** | 15.9x -> 3.0x | 1.07x / 1.53x | | share of total coverage, DMSO -> VTP | **1.4818% -> 0.2962%** | 1.975% -> 0.354% | — | | **share ratio** | **0.1999** | 0.179 | **1.12x** | | fraction still in the track's own top | **10.58%** | 12.8% | 1.21x | | fraction losing within-sample rank | **91.51%** | 92.3% | 1.01x | | per-bin share-ratio median | **0.1333** | 0.174 | 1.31x | | per-bin fraction below 0.5 | **92.9%** | 95.5% | 1.03x | **Two independent deposits, two species, two different antibodies, two genome builds, and six statistics that agree within 1.53-fold and five of them within 1.31-fold.** This is the closest replication this project has recorded. **What it does and does not establish.** It establishes that the drug's on-target chromatin action - removing menin from menin's own strongest sites - is quantitatively the same in the human index fusion as in murine MLL-AF9. **It does not establish that the fusion is evicted**, because the antibody is anti-menin and menin also occupies wild-type MLL1 complexes and non-KMT2A sites. The corresponding MLL1-N measurement in the same cells, same universe, is 0.9069 - see `ana-q118-the-mll1-n-antibody-masks-the-occupancy-collapse-four-fold-in-rs4-11`.

Go deeper
The construction, and why it is the same construction Q117 used

Deliberately identical so the numbers are comparable:

  • 2 kb fixed bins, coverage-bases (value x (end - start) summed into the bin) from each TDF's own tiles, so the aggregation does not depend on the tile representation.
  • Universe = input > 0 AND not blacklisted. RS4;11: 1,389,514 bins of 1,518,163 (Q117: 1,227,708 of 1,279,476).
  • Held set = the top 0.1% of the track's own DMSO values, n = 1,390 (Q117: n = 1,228).
  • Within-sample share as the cross-condition statistic, because no scale factor exists (see ana-q118-gse127507-ran-a-drosophila-spike-in-measured-it-and-deliberately-did-not-use-it).
  • Class-matched null matched on the DMSO percentile decile AND the input percentile decile, exactly Q117's matcher. The null's own share ratio is 0.4531, so the held set at 0.1999 falls 2.3-fold further than occupancy- and input-matched bins do, and E2 is 21.14 percentile points below the null.

And the CpG-island signature of the held set reproduces too: 700 of 1,390 menin-held bins (50.36%) overlap an hg19 CpG island against a universe background of 2.52% - a 20-fold enrichment 2. Q117's fusion-held set was 65.6% against 1.62%.

The blacklist mattered, and it mattered ASYMMETRICALLY

ENCODE hg19 blacklist v2 regions carry 7.23% of the RS4;11 MENIN DMSO track and 10.14% of the VTP track - so the artefact load differs between the two conditions being compared 3. Across all nine files the load is 5.47-10.14%.

This is a milder version of what Q117 found in GSE260742, where one 2 kb bin held 18.1% of a BCOR track. Here the worst single bin genome-wide is chr1:121,484,000 at 0.15-0.26% of each track - pericentromeric chr1, blacklisted, and present in all nine files. There is no catastrophic single-bin artefact in this series, and no evidence of the two-processing-batch split Q117 found; the 26 filenames are uniform (*.deduped.tdf).

Blacklisting is applied to everything in this beat, so the numbers above are post-filter.

Facts and figures
scriptanalyses/q118-menin-at-the-human-index-fusion/binsB.py
statusrun
result summaryRS4;11 MENIN held-set enrichment 14.91x -> 1.96x, share ratio 0.1999, 10.58% retained in own top 0.1%, 91.51% losing rank, per-bin median 0.1333. Q117 murine FLAG-MLLAF9: 15.9x -> 3.0x, 0.179, 12.8%, 92.3%, 0.174.
cell lineRS4;11
fusionKMT2A::AFF1
lineagehuman B-ALL (adult-derived)
drugVTP-50469 330 nM, 3 days
genome buildhg19, verified from each file's own extents
bin bp2000
n bins universe1389514
n held1390
antibodyanti-Menin, Bethyl A300-105A Lot 8
epigenetic readoutCpG-island overlap of the menin-held set: 700/1390 = 50.36% vs a 2.52% universe background
shown for fusionKMT2A::AFF1 in human RS4;11 for the MENIN arm; the fusion itself is NOT distinguished from wild-type MLL1 by either antibody in this series
inputs
[
  "GSM3635693 RS4;11 MENIN DMSO",
  "GSM3635694 RS4;11 MENIN VTP-50469 330 nM",
  "GSM3635699 RS4;11 input DMSO",
  "ENCODE hg19 blacklist v2",
  "UCSC hg19 cpgIslandExt"
]
How this could bite us back
  • The measurement that replicates is menin's, not the fusion's. A programme that reads this as 'the drug removes the fusion' has substituted the reagent's target for the intended one, which is the error `concept-eviction-names-menin-not-the-fusion` already records in this graph.
  • About one menin site in ten survives the drug at 330 nM for 3 days. If those are the sites that matter, escalating dose or duration will not reach them, and the residual is a candidate reservoir for the transcriptional recovery that precedes clinical relapse.
  • A pharmacodynamic assay built on menin ChIP would show a clean ~80% signal drop in a patient whose fusion complex is largely still on chromatin, because the MLL1-N measurement in the same cells is 0.91. That is a false reassurance mechanism, not a monitoring strategy.
Still open
  1. Does total menin protein fall under VTP-50469 in RS4;11 at 3 days? Until an immunoblot is read, 'menin leaves chromatin' and 'there is less menin' are not distinguished by any measurement in this beat.
  2. Menin retains 10.6% of its top bins. Are those the menin sites that were never MLL1-dependent - menin's JunD, SMAD3 and non-KMT2A chromatin - or a drug-resistant subset of MLL1-family sites? The 31.4% CpG-island rate of the menin-only held set suggests the former and nobody has tested it.
  3. The murine and human numbers agree to within 1.3-fold on six statistics. Is that because the drug's action is conserved, or because a top-0.1%-of-bins share statistic is dominated by the shape of the ChIP signal distribution and would give ~0.2 for any strongly evicted factor? A published negative-control eviction dataset would settle it and none was sought here.
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    In RS4;11 at 330 nM VTP-50469 for 3 days, menin enrichment at its own top 0.1% of 2 kb bins falls from 14.91-fold to 1.96-fold over the track's own genome median, its share of total coverage falls from 1.4818% to 0.2962% (ratio 0.1999), 10.58% of those bins remain in the track's own top 0.1%, 91.51% lose within-sample rank, and the per-bin share-ratio median is 0.1333 with 92.9% below 0.5.

    What this evidence cannot show

    NEITHER ANTIBODY IN THIS SERIES IS FUSION-SPECIFIC AND THE ANTIBODY NAME SAYS SO. MLL1n is anti-MLL1 N-terminus (Bethyl A300-086A, lots 5 or 6); the N-terminus is retained in KMT2A::AFF1, so that track is the sum of the fusion and wild-type MLL1 in unknown proportion. MENIN is Bethyl A300-105A Lot 8; menin binds the fusion, wild-type MLL1 and non-KMT2A partners. Q117 could attribute its numbers because its fusion was FLAG-tagged; nothing here can. Every occupancy statement in this beat is about MLL1-N-reactive material or menin-occupied sites, never about the fusion. No scale factor exists for any sample. The depositors state that a Drosophila spike-in was run on one batch, that fly/human ratios were 'found to be small', and that NO spike-in normalisation was applied; no number for that measurement is reported and none is deposited. Every cross-condition figure here is therefore a WITHIN-SAMPLE share or a within-sample rank, which cancel an unknown per-sample scale factor exactly, and never a fold change. n = 1 per antibody per condition per line - there is no replicate anywhere in GSE127507 - so no number here carries a spread and none is precise. RS4;11 is an ADULT-derived t(4;11) B-ALL line, not an infant-derived one. The genome median of both tracks was asserted non-zero before any ratio was formed (DMSO and VTP medians printed in the result JSON), because a ratio gate whose denominator can be zero passes by construction. There is no VTP input for RS4;11, so no input-corrected cross-condition statement is possible in this line; the input-corrected version of this analysis exists only for MOLM13.

    Evidence that argues against this (1)
    • Occupancy loss cannot be separated from protein loss by ChIP. If VTP-50469 reduced total menin protein, the same numbers would appear. The depositing paper reports menin displacement from high-molecular-weight complexes and chromatin, and a menin immunoblot would settle it, but no blot was read in this beat. The word 'collapse' in this entry is agnostic between displacement and depletion. Krivtsov AV et al. 2019
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    700 of the 1,390 menin-held bins in RS4;11 (50.36%) overlap an hg19 CpG island, against a universe background of 2.52% - a 20-fold enrichment, and the MLL1-N held set is higher still at 75.18%.

    What this evidence cannot show

    A 2 kb bin is wider than most CpG islands, so every rate is inflated equally and only the contrast is meaningful. cpgIslandExt is a sequence annotation, not a methylation measurement in RS4;11. ChIP peaks and CpG islands both concentrate at promoters, so the enrichment overstates how specifically CXXC-driven it is.

    Evidence that argues against this (1)
    • The MLL1-N held set is MORE CpG-island-enriched (75.18%) than the menin held set (50.36%), and 31.4% of the menin-only held bins are islands against 80.8% of the MLL1n-only ones. So menin occupies a substantial non-island compartment that an MLL1-N ChIP does not see, which weakens any reading of the menin track as a pure proxy for MLL1-family occupancy. Iqbal Lab oncofusion knowledge graph 2026
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    ENCODE hg19 blacklist v2 regions carry 7.23% of the RS4;11 MENIN DMSO track and 10.14% of the VTP track, so the artefact load differs between the conditions being compared; across all nine binned files the load is 5.47-10.14% and the worst single 2 kb bin genome-wide is chr1:121,484,000 at 0.15-0.26% of each track.

    What this evidence cannot show

    The blacklist is an annotation of regions that behave badly in short-read data generally, not a measurement in these cells. v2 is not exhaustive for hg19 satellite sequence, so some artefact may survive the filter.

    Evidence that argues against this (1)
    • A 3-percentage-point difference in artefact load between conditions is small compared with the 5-fold occupancy change being measured, so filtering cannot be what produces the result. Q117's experience was the opposite - there a blacklist deviation flipped a pre-registered count across its threshold - and the contrast is the point: the same deviation is load-bearing in one series and not in another. Iqbal Lab oncofusion knowledge graph 2026
Connected to