Experiment: a 6 / 24 / 72 hour / 7 day VTP-50469 time course in SEM reading menin occupancy and H3K4me3 together, to decide whether the 23% is a partial effect or a slow one
Measure the same thing at four time points instead of one. If the effect keeps growing, the mark is just slow to change; if it stops at a quarter, something is genuinely holding the rest in place. These are different biology and currently nobody can tell them apart.
The same thing, for a scientist
**Design.** SEM, 30 uM VTP-50469 vs DMSO, harvest at 6 h, 24 h, 72 h and 7 days, n=3. At each point: menin CUT&Tag (does menin leave, and when), H3K4me3 CUT&Tag with spike-in, IgG, plus nascent transcription (TT-seq or PRO-seq) at the same loci and a viability curve. **The discriminating measurement** is the menin-bound/menin-free H3K4me3 class ratio as a function of time, with menin occupancy at the same sites on the same axis. **Scoreable outcomes, pre-declared:** if the ratio is still falling between 72 h and 7 days and reaches below 0.5, the 72-hour 23% is a lag and H3K4me3 is fusion-dependent at these sites. If it plateaus at 0.7-0.8 by 24 h and is unchanged at 7 days while menin occupancy has already collapsed by 6 h, the residual is menin-independent and hypothesis option (ii) or (i) is live.
Why the transcription arm is not optional
The point of the chromatin measurement is that it is supposed to explain the transcriptional shutdown. If HOXA9 and MEIS1 nascent transcription collapses by 6 h while the H3K4me3 class ratio has barely moved, H3K4me3 is downstream of or parallel to the effect rather than on the path, and that single ordering would do more to settle this project's menin-interface question than another occupancy dataset. The time course is the cheapest way to get an ordering, and ordering is what a single 72-hour snapshot can never give.
The falsifier this experiment cashes
It scores the falsifier attached to hyp-q140-h3k4me3-survives-menin-eviction-at-fusion-targets-because-a-second-complex-holds-it: if H3K4me3 at menin-bound sites falls by more than 60% at any time point under 30 uM VTP-50469 in SEM, the residual the hypothesis is about does not exist. Named interval: menin-enriched bins within the top-0.1% DMSO H3K4me3 set, matched on H3K4me3 level. Named system: SEM, human, infant-derived KMT2A::AFF1. Named effect size: 60%.
| design | SEM, 30 uM VTP-50469 vs DMSO, harvest 6 h / 24 h / 72 h / 7 d, n=3 per point |
|---|---|
| readout | menin CUT&Tag occupancy; spike-in H3K4me3 CUT&Tag; menin-bound/menin-free class ratio vs time; nascent transcription at HOXA9, MEIS1, CDK6, JMJD1C; viability |
| controls | IgG at every time point; DMSO at every time point; pre-specified housekeeping promoters; viability-sorted cells at 7 d |
| cost tier | high |
| timeline | 3-4 months |
| cashes falsifier of | hyp-q140-h3k4me3-survives-menin-eviction-at-fusion-targets-because-a-second-complex-holds-it |
| scoreable threshold | class ratio below 0.5 at any time point refutes the residual; a plateau at 0.7-0.8 by 24 h with menin already gone at 6 h supports it |
- If H3K4me3 turns out to lag transcription badly, every chromatin-mark pharmacodynamic assay in this drug class is reading a signal that arrives after the decision point.
- A 7-day arm in a drug that kills selects for the cells that survive. Any chromatin state measured there may be the resistant population's, not the drug's effect - which is interesting, but it is a different experiment and must not be reported as this one.
- Should the perturbation be the drug or a MEN1 degron? The drug is what patients receive; the degron gives clean kinetics. Running both in the same design would separate drug-specific effects from menin loss, at roughly double the cost.
- Is 7 days too long for SEM under 30 uM, i.e. will there be cells left to profile? A viability pilot is the first step and would cost a week.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
The existing measurement is a single 72-hour snapshot giving 0.766 (95% CI 0.667-0.870), which cannot distinguish a plateau from a point on a descending curve.
sem-arm-identity-from-signal: run3.py and run4_bootstrap.py (this repository) · Q140-sem-menin-arm · Oncofusion knowledge graph 2026What this evidence cannot show
One time point, two libraries. The inability to distinguish lag from partiality is a property of the design, not of the analysis.
Evidence that argues against this (1)
- A time course adds a dimension but also adds a confound: prolonged exposure to a cytotoxic agent changes the cell population, so a 7-day sample is a different set of cells. Sorting on a viability marker at each point, or using a degron instead of the drug, is the mitigation. Krivtsov AV et al. 2019
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
VTP-50469 causes differentiation and apoptosis in MLL-rearranged cells, so the cell population at 7 days differs from that at 6 hours and the time course must control for it.
A Menin-MLL Inhibitor Induces Specific Chromatin Changes and Eradicates Disease in Models of MLL-Rearranged Leukemia · Krivtsov AV, Evans K, Gadrey JY, Eschle BK, Hatton C, Uckelmann HJ, et al., Armstrong SA · Cancer Cell 36:660-673.e11 2019What this evidence cannot show
Establishes the confound the design must handle; does not specify the kinetics of menin displacement in SEM, which is precisely what is unknown.
Evidence that argues against this (2)
- If the population changes over seven days, a time course does not merely need to 'control for' it - the confound may be unremovable, because differentiation and chromatin change are the same process here rather than signal and nuisance. A design that cannot separate them should say so rather than propose a covariate.
- Sorting or gating to hold the population fixed would itself select cells that resist the drug, which is the subpopulation whose chromatin is least representative of the drug's mechanism. Both the controlled and uncontrolled versions of this experiment are biased, in opposite directions, and the proposal does not say which bias it prefers.
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Oncoprotein and menin levels on chromatin are dynamic and can be reduced together under a menin inhibitor in a patient sample, with cofactors persisting at target loci - evidence that a time-resolved occupancy measurement is informative in this system.
MLL oncoprotein levels influence leukemia lineage identities · Janssens DH, Duran M, Otto DJ, Wu W, Xu Y, Kirkey D, Mullighan CG, Yi JS, Meshinchi S, Sarthy JF, Ahmad K, Henikoff S · Nature Communications 15:9341 2024What this evidence cannot show
A single lineage-switching sample under revumenib rather than VTP-50469, and not a time course.
Evidence that argues against this (1)
- The cited evidence is one patient sample in an unusual clinical state, used here to argue that time-resolved occupancy is informative in general. A single lineage-switching case is thin support for a design choice, and the same cycle's q141 beat found no MENIN ChIP anywhere between 3 hours and 4 days in the best-characterised degron deposit - so the field's actual record on intermediate timepoints is near-empty rather than encouraging.
The time course decides whether the residual is a plateau or a lag
Re-measures the 0.766 class ratio at four time points
Reads menin occupancy on the same axis as the mark, which no existing dataset does in this line
A 6 h to 7 d exposure series in SEM
The DOT1L arm's unverified target engagement is one reason the comparator needs an H3K79me2 readout