Oncofusion

start / graph / resistance-q130-a-polycomb-deficient-cell-loses-h3k79me2-normally-where-the-fusion-leaves

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

A PCGF1-null cell - a published resistance genotype - loses H3K79me2 at the most-evicted sites exactly as a control cell does (0.6123 against 0.5860), so the new stratified readout scores it as fully drug-engaged, and stratification does NOT close the resistance blind spot it was hoped to close

also called the blind spot survives stratification · H3K79me2 cannot see this resistance · Q130's bite-back

A cell engineered to carry a change that should make it resist the drug still loses the activating tag just as completely as a normal cell where the drug works. A test based on that tag would report the drug as working perfectly in a cell that is not responding.

The same thing, for a scientist

**The resistance mechanism.** The source paper's model is that DOT1L provides transcriptional memory through PRC1.1 antagonism; PCGF1 is the PRC1.1-defining paralog, and Q125 measured that a PCGF1 knockout **reduces but does not abolish** the repressive-mark gain at the sites the menin inhibitor vacates. A cell that cannot silence the vacated sites should be evicted normally and respond poorly [E1]. **What this beat adds, and it is bad news for the readout.** Stratifying H3K79me2 by eviction depth was the obvious way to make the resistant genotype visible. It does not [E2]: | VTP/DMSO, H3K79me2 | Q1 most evicted | Q2 | Q3 | Q4 least | |---|---:|---:|---:|---:| | control | 0.5860 | 0.8188 | 0.9280 | 1.1865 | | **PCGF1-KO** | **0.6123** | 0.7955 | 0.8928 | 1.0464 | | difference | **+0.0263** | -0.0233 | -0.0352 | **-0.1401** | **At the sites the drug actually clears the fusion from, the two genotypes are indistinguishable at the resolution this deposit supports** - 0.0263 on a scale where the drug effect itself is 0.41, i.e. 6% of the signal, in a single unreplicated library whose span reproducibility is a factor of 1.99. The whole genotype effect sits at Q4, the quartile the drug barely touches, and runs the other way [E2]. **So the resistance is invisible on the most-used readout and visible on a less-used one.** Q125 found the genotype clearly on the repressive marks - H2AK119ub Q1 falls from 1.4450 (paired; rep2 gives 1.0151) to 1.1325 and H3K27me3 Q1 from 2.0007 to 1.0134 - a difference an order of magnitude larger than H3K79me2's [E3]. **A resistance that a H2AK119ub assay would catch and an H3K79me2 assay would not is still a trap, because H3K79me2 is the readout this field reaches for first.**

Go deeper
What would have to be true for this to be a false alarm

The knockout may not be resistant. GSE260742 deposits no viability, no growth curve and no transcriptional readout for the NTvKO cultures at all - Q125 recorded that as a bounded absence - so 'resistance genotype' is imported from the source paper's model, not measured in these cells. If the PCGF1-KO clone responds normally to the drug, an assay scoring it as engaged is correct.

The knockout may not be a knockout. One CRISPR guide, one clone, no second clone, no rescue, and no deposited PCGF1 protein or transcript measurement anywhere in the series.

Neither of these makes the readout safe. They make the specific experiment weak. The general shape - a genotype that changes silencing without changing eviction is invisible to an eviction-following readout - follows from the mechanism regardless of whether this clone is the right example of it.

The practical consequence for anyone building an assay

Do not build a single-mark pharmacodynamic assay on H3K79me2 for this drug class. It reports eviction, and the resistance mechanisms this corpus knows about - BCOR, BCORL1, KDM2B, PCGF1 loss, all of which occur in leukaemia - act downstream of eviction. An H3K79me2 assay, stratified or not, reports full target engagement in exactly the cells that will not respond.

A two-mark readout is the minimum the data supports: an eviction-following mark (H3K79me2 or H3K4me3) to confirm the drug reached the target, and a repressive mark (H2AK119ub or H3K27me3) to confirm the cell executed the consequence. The gap between them is the resistance signal. Nothing in this corpus has tested that two-mark readout as a unit, and it is the obvious follow-up.

Facts and figures
kindresistance
severityhigh for assay design, unquantified for clinical outcome - no viability or transcriptional measurement exists for the genotype in question
mitigationa two-mark pharmacodynamic readout: an eviction-following mark (H3K79me2 or H3K4me3) plus a repressive mark (H2AK119ub or H3K27me3), with the GAP between them as the resistance signal. Untested as a unit anywhere in this corpus.
genotypePCGF1 knockout, clone C1, GSE260742 NTvKO arm
q1 genotype difference0.0263
q4 genotype difference-0.1401
absence classmeasured-not-analysed
absence class noteThe stratified genotype comparison on H3K79me2 had never been run; the threat it tests has stood on an aggregate since cycle 83. The two-mark readout proposed as mitigation has never been scored as a unit, and every number needed to score it is on disk.
absence class assigned2026-09-23 (Q130)
epigenetic readoutH3K79me2 eviction-depth response, control against PCGF1-knockout, as a resistance-detection test
affects
[
  "drug-revumenib",
  "drug-ziftomenib",
  "drug-bleximenib",
  "drug-enzomenib"
]
repressive mark q1 differences
{
  "H2AK119ub_paired": -0.3125,
  "H3K27me3": -0.9873
}
How this could bite us back
  • This entry IS the bite-back and it should not be softened by the beat's positive headline. The gradient exists, it is large, it replicates across four murine cultures and a human line - and it does not detect the one resistance genotype this deposit contains. A pharmacodynamic marker that cannot see resistance is a marker for exposure, not for benefit.
  • Proposing a two-mark readout as mitigation doubles an already impractical assay: it now needs a fusion-occupancy ChIP to define the strata plus two histone ChIPs, in a marrow aspirate. The honest position is that the mechanism is clearer and the assay is further away.
  • If the resistance signal is the GAP between an eviction-following mark and a repressive mark, then a cell with normal polycomb and poor drug penetration looks the same as a cell with good penetration and broken polycomb on the eviction mark alone - and the two need opposite clinical responses, more drug against a different drug.
Still open
  1. Does the two-mark gap actually separate the genotypes? Every number needed is on disk in C:/Temp/q124 and C:/Temp/q125 and the comparison has never been made; it is the cheapest useful follow-up in this line.
  2. Do BCOR, BCORL1 or KDM2B loss behave like PCGF1 loss on these readouts? GSE260742's only genetic perturbation is the PCGF1 knockout, which Q125 recorded as a bounded absence for BCOR.
  3. Is the PCGF1-KO clone actually drug-resistant? No viability or transcriptional measurement for these cultures is deposited, so the word 'resistance' is imported from a model.
Every source behind this entry (3)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    The resistance model is that a polycomb-deficient cell is evicted normally by a menin inhibitor and silences the vacated sites less, so occupancy-based readouts score it as fully engaged - Q125's conclusion from the same cultures, resting on the source paper's PRC1.1-antagonism model.

    DOT1L provides transcriptional memory through PRC1.1 antagonism. · Neville D, Ferguson DT, Heikamp EB, et al. · Nat Cell Biol 2026
    PMID 41634341DOI
    What this evidence cannot show

    The resistance reading follows from the source paper's model rather than from a clinical observation or a viability measurement in these cells. GSE260742 deposits no viability, growth or transcriptional readout for the NTvKO cultures.

    Evidence that argues against this (1)
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    Stratified by eviction depth, the PCGF1-KO clone's H3K79me2 response to VTP-50469 differs from the control's by +0.0263 at the most-evicted quartile - about 6% of the drug effect there and below this beat's pre-registered 0.0498 threshold - while the whole genotype difference sits at the least-evicted quartile at -0.1401, so the readout does not separate the genotypes where the drug works.

    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
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    The same genotype IS clearly visible on the repressive marks in the same cultures and the same bins: Q125 measured H2AK119ub at the most-evicted quartile falling from 1.4450 in the control to 1.1325 in the knockout, and H3K27me3 from 2.0007 to 1.0134, differences an order of magnitude larger than H3K79me2's 0.0263.

    What this evidence cannot show

    Q125's H2AK119ub REP2 convention disagrees with PAIRED on the genotype contrast, so the repressive-side genotype signal is itself convention-dependent; the H3K27me3 arm has one convention and no replicate. Read from the committed JSON in code.

    Evidence that argues against this (1)
    • If the genotype is visible on the repressive marks, the blind spot is a property of one readout rather than of the cell, which is a better outcome for the therapeutic hypothesis than this entry's title suggests. Iqbal Lab oncofusion knowledge graph 2026
Connected to