Two drugs, two spatial mechanisms on one mark: catalytic DOT1L inhibition de-concentrates H3K79me2 out of its peaks and into mid-occupancy chromatin, while menin-interface inhibition lowers the mark's share at the peaks AND at mid-occupancy, so its effect is relocation rather than a smaller version of the same change
also called relocation versus de-concentration · the two drugs do not act on one axis · why F is a ratio of two different things
Both drugs lower the H3K79 tag where the fusion protein sits, but blocking the writing enzyme spreads the tag out of its peaks into the middle of the genome while blocking the menin interface lowers it in the peaks and the middle alike. One drug's effect is therefore not a fraction of the other's - they are different kinds of change.
The same thing, for a scientist
**Statement.** In murine MLL-AF9 (KMT2A::MLLT3) myeloid cells, 5 uM SGC0946 for 72 h and 500 nM VTP-50469 for 48 h both reduce H3K79me2's within-sample share at the tagged fusion's own top-0.1% sites - to 0.13-0.15 and 0.77-0.83 respectively - **and they do so with opposite compartmental signatures.** SGC0946 **raises** the mark's mid-occupancy share ratio to **1.89-2.01**; VTP-50469 **lowers** it to **0.69-0.78**. Four independent cultures, two genotypes, and the two ranges are 2.4-2.9 fold apart with no overlap. **Origin: data-pattern**, from a quantity computed by three consecutive beats and read by none of them - the mid-occupancy floor, Q118's DEVIATION-2 definition, pre-registered in section 3 of both Q122's and Q124's protocols. **The mechanistic reading, stated with its own alternative.** A catalytic block on a writer whose deposition is turnover-coupled should hit the highest-turnover territory hardest, so the residual mark becomes relatively enriched in the slower mid-occupancy fraction - **de-concentration**, which is what the SGC floor of ~2.0 looks like. An interface inhibitor that removes the fusion from chromatin removes the recruitment platform at specific sites, so the mark's share should fall where the fusion was; **the observation that it also falls at mid-occupancy means the vacated share does not land in that compartment**, and must go to the low-occupancy tail or to bins carrying no signal in vehicle. **The rival reading is that VTP-50469 produces a small global loss of the mark that a within-sample share renders as a broad share decrease**, and a share-based instrument cannot distinguish those - so the *sign difference* is the robust part and the word 'relocates' is the interpretation. **Three independent statistics support the separation, from three different bin populations** - Q122's pre-registered per-bin criterion **R4 fired** (Spearman between the two drugs' log share ratios over the mark's own sites = **-0.026** against a threshold of 0.10); this beat's enrichment residual separates them at **-0.16/-0.20 against -0.004/-0.025** in four cultures each; and the floor separates them by **sign**. **And the separation is channel-specific**: in the same universe with the same definition, the FLAG track's floor is 0.9288/0.9051 and the MLL1-N track's is 0.9151/0.9226 - all four within 0.10 of unity under both drugs. **What it costs cycle 82's headline.** `F = (1 - V)/(1 - G) = 0.0947-0.2500` remains a valid arithmetic ratio of two share-ratio deficits and **no committed number changes.** What this mechanism removes is the word *achievable*: expressing one drug's effect as a share of the other's presumes a common axis, and the compartmental signatures say there is not one.
What would distinguish the two readings of the VTP signature
The floor's fall under VTP-50469 has two explanations and this instrument cannot separate them.
| reading | predicts | how to test |
|---|---|---|
| relocation - the vacated share lands in the low-occupancy tail or in vehicle-zero bins | the share deficit at peaks plus mid-occupancy is matched by a share SURPLUS below percentile 50 | decompose the universe into percentile strata below 50 and check the share sums; no new download, and this beat did not do it |
| global loss rendered as a broad share decrease | no compartment gains; the apparent surplus is spread thinly across the whole low tail in proportion to its size | the same decomposition, with the surplus per bin compared against a proportional-fill null |
Both need one script over bins already on disk. A spike-in would settle it outright and the series has none - Q117 verified zero spike-in vocabulary across all 419,824 characters of the 158 records - and the published literature has demonstrated that standard normalisation fails to detect exactly this class of genome-wide change for H3K27me3 under an EZH2 inhibitor.
Why this was invisible for three beats
The floor was introduced by Q118 as a normaliser: a denominator to divide a share ratio by, so that a value could be expressed against its own track's background. Q120 and Q122 both adopted the definition and both stored floor_pct50to90 for every contrast in their committed results.json. No criterion was ever written about the floor itself, so no verdict mentioned it, so no summary quoted it.
Q122's file contains the sign reversal in its very first culture - 0.7804 under VTP against 1.9918 under SGC0946 - and it required no computation to see. A quantity built to sit in a denominator is read as machinery rather than as data, and this is the second time this project has found a measurement hiding inside an instrument.
| mark | H3K79me2 |
|---|---|
| writer of that mark | gene-dot1l |
| epigenetic readout | the mark's mid-occupancy (DMSO percentile 50-90) within-sample share ratio is 1.89-2.01 under 5 uM SGC0946 and 0.69-0.78 under 500 nM VTP-50469 in four murine MLL-AF9 cultures, against 0.905-0.929 for the FLAG and MLL1-N channels under both drugs |
| mechanism sgc0946 | de-concentration: catalytic block on a turnover-coupled writer hits the highest-turnover territory hardest, so the residual mark is relatively enriched in the slower mid-occupancy fraction |
| mechanism vtp50469 | relocation (INTERPRETATION, not measurement): the share falls at the fusion's sites and at mid-occupancy, so the vacated share must go to the low-occupancy tail or to vehicle-zero bins; a global loss rendered as a broad share decrease is the unexcluded rival |
| origin | data-pattern, from a quantity computed by Q118, Q120 and Q122 and read as data by none of them |
| shown for fusion | KMT2A::MLLT3 in a murine FLAG-tagged over-expression model, four cultures; nothing human and nothing KMT2A::AFF1. The floor is a genome-compartment statistic and is not fusion-specific by construction |
| absence class | measured-not-analysed |
| absence class note | Half the numbers behind this mechanism required no new computation: they are in Q122's committed results.json, which has been in this repository since cycle 82. The decomposition that would separate relocation from global loss - percentile strata below 50 - also needs no download, because the bins are on C:/Temp/q122 and C:/Temp/q124. Cost to close the remaining ambiguity by re-analysis: one script. Cost to close it properly: a spike-in, which this series does not have. |
| absence class assigned | 2026-09-23 (Q124) |
- If the two drugs act by different spatial mechanisms, the combination rationale cannot be argued from 'the menin inhibitor gets part of the way and the DOT1L inhibitor finishes the job'. This corpus has argued the combination both ways from exactly that single-axis reasoning, and neither argument is supported by these floors.
- It also breaks H3K79me2 as a shared pharmacodynamic readout across the two drug classes. A locus panel sitting in mid-occupancy chromatin would report a doubling under a DOT1L inhibitor and a decrease under a menin inhibitor, from the same assay on the same mark.
- The mechanism's own weakest point is that the timecourse arm of the same series, same antibody and same instrument gives the opposite sign for VTP-50469. If the four-culture reversal is a batch property rather than a drug property, this entry is describing an artefact and the qualification it places on F is unearned.
- Where does the mark's share go under a menin inhibitor? A percentile decomposition below 50, including the bins that are zero in vehicle, would separate relocation from a global loss rendered as a broad share decrease. It needs no download and this beat did not run it.
- Does the sign reversal hold in a human line? K562 and MOLM13 both have DMSO/SGC/VTP H3K79me2 triplets in this series and Q122 binned all six files and stored their floors, so this is a file lookup.
- Is there a combination arm anywhere? If the combination's floor lands between the two single-agent floors the one-axis model survives; if outside, it does not. GSE260742 deposits no combination sample in any channel.
- Does the same compartmental split appear for H2AK119ub, the mark the depositing paper is actually about? The NTvKO and K562 H2AK119ub triplets are deposited and Q122 binned six of them, so their floors are computable from files already on disk.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
In murine MLL-AF9 (KMT2A::MLLT3), 2 kb bins, universe 1,227,708, floor = aggregate within-sample share ratio over bins in DMSO percentile 50-90 of the same track: H3K79me2's floor is 0.7804, 0.6947, 0.7230 and 0.7620 under 500 nM VTP-50469 48 h and 1.9918, 2.0094, 1.9478 and 1.8905 under 5 uM SGC0946 72 h in the F3, F3_NT, NT and PCGF1-KO C1 cultures respectively, while at the same bins the FLAG track's floor is 0.9288 and 0.9051 and the MLL1-N track's is 0.9151 and 0.9226 under the two drugs; at the fusion's own top-0.1% sites the same contrasts give H3K79me2 share ratios of 0.77-0.83 under VTP-50469 and 0.13-0.15 under SGC0946.
Q124 re-analysis of GSE260742: the floor contrast across four murine cultures and the channel-specificity control. analyses/q124-the-timecourse-and-the-pcgf1-knockout/run.py, deviation1_enrichment_null.py, deviation2_common_support.py, deviation3_four_statistic_steps.py and results.json, run 2026-09-23 · Iqbal Lab oncofusion knowledge graph 2026What this evidence cannot show
THE TWO DRUG ARMS ARE NOT EXPOSURE-MATCHED in any of the four cultures - VTP 48 h against SGC and vehicle at 72 h - so the comparison is conservative for the SIGN of the reversal and not for its magnitude. THE FLOOR IS DEMONSTRABLY BATCH-SENSITIVE IN THIS SERIES: the timecourse arm's VTP floors RISE (1.0665, 1.0896, 1.3606, 1.2010 at 4/8/16/24 h), the opposite sign from all four 48 h cultures, in an arm whose tracks span 2.208-fold in non-zero bin count. The mechanism rests on reproducibility across four cultures and on channel specificity, not on any single value. THE WORD 'RELOCATES' IS AN INTERPRETATION: a within-sample share cannot distinguish relocation from a global loss rendered as a broad share decrease, and the test that would - a percentile decomposition below 50 - was not run. Two of the eight H3K79me2 floor values are read from Q122's committed JSON rather than re-derived. 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, which cancels the unknown per-sample scale factor exactly and IS BLIND TO A UNIFORM CHANGE BY CONSTRUCTION. Nothing here measures the mark's abundance; everything measures its distribution. n = 1 per antibody per condition per culture; no number carries a spread and none is precise. REAGENTS AND WHAT EACH CANNOT DISTINGUISH: H3K79me2 is ABCAM ab3594-100UG, 'Rabbit polyclonal to Histone H3 (di methyl K79) - ChIP Grade', read verbatim from the paper's key-resources sheet 41556_2025_1859_MOESM6_ESM.xlsx this session - a POLYCLONAL with no specificity document in the sheet, so it cannot distinguish me2 from me3, and it reads a MARK, so it cannot distinguish halted deposition from active demethylation from replicative dilution and cannot attribute any of them to the fusion. `antibod` occurs ZERO times in the 158 GEO records, so the spreadsheet is the only route to any reagent in this series. S_FLAG is defined by DYKDDDDK Tag (D6W5B) Rabbit mAb, NEB 14793S, on a RETROVIRALLY OVER-EXPRESSED tagged MLLAF9 construct, so it is the over-expressed molecule's sites and not an endogenous fusion's, and it cannot see a fusion whose tag has been proteolysed. VTP50469 disrupts an INTERFACE, so its H3K79me2 effect is at least four steps removed; SGC0946 inhibits DOT1L's CATALYSIS, so its effect says nothing about DOT1L occupancy. 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 (3)
the measurement this mechanism is drawn from: floors 0.69-0.78 under VTP against 1.89-2.01 under SGC0946 in four cultures, with FLAG and MLL1-N at 0.905-0.929 under both
the compartmental signature of catalytic DOT1L inhibition: the mark's mid-occupancy share roughly doubles while its share at the fusion's own sites falls to 0.13-0.15, consistent with turnover-coupled deposition failing fastest where turnover is highest
- related to →Pinometostat (EPZ-5676) — the DOT1L inhibitor, and the most instructive failure in this field
the compartmental reading generalises a prediction to the clinical DOT1L inhibitor: if catalytic inhibition de-concentrates the mark rather than removing it uniformly, a locus-panel pharmacodynamic readout in mid-occupancy chromatin would move the wrong way
the mark's compartmental distribution separates the two drug classes by sign where its aggregate level at the fusion's sites does not - 0.13-0.15 against 0.77-0.83 is a difference of degree, 1.89-2.01 against 0.69-0.78 is a difference of kind