Oncofusion

start / graph / mechanism-q130-h3k79me2-does-not-only-leave-the-evicted-sites-it-rises-19-percent-at-the-sites-the-fusion-keeps

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q130-does-the-eviction-gradient-exist-in-h3k79me2

The menin inhibitor's effect on H3K79me2 is a REDISTRIBUTION within the fusion's own site set, not a loss from it: the share falls 41.4% at the quartile the fusion is most evicted from and RISES 18.7% at the quartile it is least evicted from, in murine MLL-AF9

also called the Q4 gain · H3K79me2 moves towards the sites the fusion keeps

The drug does not simply strip the activating tag off the fusion's landing sites. Where most of the fusion is pulled off, the tag drops by about 40 percent; where least of it is pulled off, the tag actually goes up by about 20 percent. The tag is moving within the fusion's own territory, not just disappearing from it.

The same thing, for a scientist

**In murine MLL-AF9 (KMT2A::MLLT3) myeloid cells**, the four eviction-depth quartiles of `S_FLAG` carry H3K79me2 VTP/DMSO within-sample share ratios of 0.5860 / 0.8188 / 0.9280 / **1.1865** in the non-targeting control [E1]. **Q4 is above 1.** The least-evicted quarter of the fusion's own strongest sites gains 18.7% of H3K79me2 share under the drug. **This is a per-bin fact, not an artefact of summing.** The MEDIAN bin in Q4 has a log2 share ratio of **+0.2673**, i.e. 1.204-fold, and only 24.7% of Q4 bins are individually below 1 - against 89.3% at Q1 [E2]. **It is also visible in the knockout, weaker.** The PCGF1-KO clone gives 1.0464 at Q4 against the control's 1.1865, so the gain is 0.14 smaller in the knockout while the Q1 loss is essentially identical (0.6123 against 0.5860) [E1]. **Where the displaced share goes is not determined here, and this entry does not claim it is.** A within-sample share sums to one, so a rise at Q4 requires a fall elsewhere, and Q124 measured the mid-occupancy floor of the same track falling to 0.723 under the same drug in the same culture [E3]. So both the most-evicted sites AND the mid-occupancy compartment lose share while the least-evicted sites gain it. That accounting does not balance on the bins this beat looked at, and the remainder is an open question rather than a conclusion.

Go deeper
Why 'redistribution' and not 'loss' is the right word, and what it costs to say it

The word is forced by the normalisation, not chosen for effect. There is no spike-in anywhere in GSE260742, so every number is a within-sample share and a uniform loss of H3K79me2 is invisible by construction. What a share ratio can see is one compartment gaining relative to another, which is what 0.5860 at Q1 against 1.1865 at Q4 is.

The cost of that precision is that this entry cannot say the drug reduces H3K79me2. It says the drug moves H3K79me2 out of the sites it evicts the fusion from and towards the sites it does not. Those are different claims and the corpus has conflated them before. PMID 27875550 is the worked demonstration: standard normalisation sees no H3K27me3 change under an EZH2 inhibitor, and the real change appears only with a chromatin spike-in.

What would make this mechanistic rather than descriptive

Three things, none available in this deposit:

  1. A spike-in-normalised H3K79me2 arm, which would say whether Q4's 1.1865 is an absolute gain or the arithmetic shadow of Q1's loss.
  2. A DOT1L occupancy track in the same samples. If DOT1L itself redistributes to the retained sites, the mark is following its writer; if DOT1L falls everywhere, the Q4 gain is turnover-limited and reflects where the mark is not being diluted by replication.
  3. Matched RNA-seq. No transcriptional readout is deposited for the NTvKO cultures at all, which Q125 recorded as a bounded absence, so 'the sites the fusion keeps gain an activating mark' cannot be connected to output.
Facts and figures
mechanism classchromatin mark redistribution under pharmacological eviction
systemmurine MLL-AF9 (KMT2A::MLLT3) myeloid blood cells, GSE260742 NTvKO arm
q4 ratio control1.1865
q4 ratio pcgf1 ko1.0464
q1 ratio control0.586
q1 ratio pcgf1 ko0.6123
absence classmeasured-not-analysed
absence class noteThe Q4 value was computable from tracks on disk since 2026-09-23 08:37 and from quartiles committed since cycle 68. Nobody had asked whether any quartile of S_FLAG went UP under the drug.
absence class assigned2026-09-23 (Q130)
epigenetic readoutH3K79me2 within-sample share, quartile-resolved, under VTP-50469
median per bin log2 by quartile control
[
  -0.7213,
  -0.168,
  -0.0158,
  0.2673
]
fraction of bins below 1 by quartile control
[
  0.8929,
  0.6503,
  0.5114,
  0.2467
]
q124 mid occupancy floor vtp
{
  "NT": 0.723,
  "PCGF1_KO_C1": 0.762
}
How this could bite us back
  • If a menin inhibitor concentrates an activating mark onto the fusion sites it fails to clear, the residual disease compartment may be more, not less, transcriptionally competent at exactly the sites the drug left behind. That is a mechanism for a plateau in response, and it predicts that adding a DOT1L inhibitor should matter most in the Q3/Q4 regime.
  • A pharmacodynamic assay that samples the fusion's sites without knowing which quartile it sampled can return a number above 1 in a fully drug-exposed cell. That is not a failure of the drug; it is a failure of the sampling, and it would read as non-engagement.
Still open
  1. Does the share that leaves Q1 arrive anywhere this beat looked? It does not arrive in mid-occupancy chromatin (Q124's floor falls to 0.723) and only part of it can be accounted for by Q4's gain, because Q4 is a quarter of 1,228 bins out of 1,227,708.
  2. Is the Q4 gain an absolute gain or a relative one? Only a spike-in-normalised arm separates them and none exists in this series.
  3. Does DOT1L itself redistribute the same way? No DOT1L occupancy track is deposited in the NTvKO arm, so the mark's movement cannot be attributed to its writer's movement.
Every source behind this entry (3)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    In murine MLL-AF9 (GSE260742 NTvKO), H3K79me2's VTP/DMSO within-sample share ratio at the LEAST-evicted quartile of S_FLAG is 1.1865 in the non-targeting control and 1.0464 in the PCGF1-KO clone - above 1 in both - while the MOST-evicted quartile is 0.5860 and 0.6123.

    What this evidence cannot show

    NO SPIKE-IN NORMALISATION IS POSSIBLE. Q117 verified zero hits for the entire spike-in vocabulary across all 419,824 characters of GSE260742's 158 sample records and no scale factor is deposited, so every cross-condition number here is a WITHIN-SAMPLE SHARE. It cancels the unknown per-sample scale factor exactly and IS BLIND TO A UNIFORM CHANGE BY CONSTRUCTION: nothing here measures H3K79me2's abundance, everything measures its distribution. EXPOSURE ASYMMETRY: DMSO 72 h and SGC0946 5 uM 72 h are matched, VTP-50469 500 nM is 48 h, so every VTP ratio is a LOWER BOUND on the 72-hour effect - identical in both genotypes, so the genotype DIFFERENCE is unaffected. REPLICATION, CORRECTED MID-BEAT: the NTvKO arm deposits ONE H3K79me2 ChIP per genotype per drug, so ITS spans carry no technical error bar and its bin-level bootstrap intervals measure sampling within one library, NOT library-to-library variation. A REPLICATE PAIR DOES EXIST for a different culture - GSM8123372/374 and GSM8123386/388, both titled MLLAF9 Control, the second carrying `rep2` - already binned by Q122 and scored in Q130 DEVIATION 4: the SIGN and the monotonicity reproduce (Spearman -1.000 in all four murine VTP contrasts in this deposit) while the SPAN reproduces only to within a factor of 1.99 (-0.2438 against -0.4847). EVERY SPAN IN THIS BEAT SHOULD BE READ WITH THAT 2-FOLD REPRODUCIBILITY. THE KNOCKOUT is one CRISPR guide, one clone C1, no second clone, no rescue, and NO deposited PCGF1 protein or transcript measurement. S_FLAG is defined by a DYKDDDDK-tag ChIP on a RETROVIRALLY OVER-EXPRESSED tagged MLLAF9 construct, so it is the over-expressed molecule's sites and not an endogenous fusion's. THE H3K79me2 ANTIBODY IS NOT RESOLVED - `antibod` occurs zero times in the 158 GEO records - and is recorded unresolved rather than guessed. SYSTEM: murine MLL-AF9 (KMT2A::MLLT3) myeloid blood cells, NOT the infant B-lineage KMT2A::AFF1 disease this graph is about.

    Evidence that argues against this (2)
    • A published methods study establishes that STANDARD ChIP-seq normalisation FAILS TO DETECT a genome-wide decrease in a histone mark under an inhibitor of its writer: standard methods see no H3K27me3 reduction under an EZH2 inhibitor, and a substantial reduction appears only with a Drosophila chromatin spike-in. Every number in this entry is exactly the kind of normalisation that paper shows to be blind. Egan B et al. 2016
    • The system is murine MLL-AF9 (KMT2A::MLLT3) myeloid cells carrying a retrovirally over-expressed FLAG-tagged construct. This graph exists for human infant B-lineage KMT2A::AFF1, and this corpus's commonest documented synthesis failure is generalising a murine MLL-AF9 myeloid result to that disease. Neville D et al. 2026
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    The Q4 gain is carried by most of Q4's bins and not by a few: the median per-bin log2 share ratio in Q4 is +0.2673 (1.204-fold) and only 24.7% of Q4 bins are individually below 1, against 89.3% at Q1.

    What this evidence cannot show

    POST HOC - this statistic is not in the pre-registration. A per-bin median over 307 bins in a single library says nothing about whether a second library would reproduce it, and no second library exists. Bins with zero signal in either track are excluded from the per-bin statistic (307 of 307 bins carried signal in both tracks in every quartile here, so nothing was dropped, but the rule matters for anyone reusing the code).

    Evidence that argues against this (1)
    • A median per-bin ratio and a ratio of sums are not independent measurements - both come from the same two tracks over the same bins - so their agreement is a consistency check and not a replication. Iqbal Lab oncofusion knowledge graph 2026
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    Q124 measured, in the same culture and under the same drug, that H3K79me2's mid-occupancy floor (DMSO percentile 50-90 of the same track) falls to 0.723 in the control and 0.762 in the PCGF1-KO under VTP-50469, while rising to 1.948 and 1.891 under SGC0946.

    What this evidence cannot show

    The floor is a different bin population from S_FLAG's quartiles and the two are not disjoint by construction. Quoted to show that the share leaving Q1 is not simply arriving in mid-occupancy chromatin; it is NOT an accounting of where the share goes, and this beat did not compute one.

    Evidence that argues against this (1)
    • Q124 itself flagged that the timecourse arm's VTP floors RISE (1.0665 / 1.0896 / 1.3606 / 1.2010) rather than falling, the opposite sign from all four 48 h cultures, and attributed that to a library-complexity confound. Iqbal Lab oncofusion knowledge graph 2026
Connected to