Oncofusion

start / graph / absence-q124-the-vtp-timecourse-in-gse260742-has-no-dot1l-inhibitor-arm-so-f-cannot-be-recomputed-kinetically

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q124-the-timecourse-and-the-pcgf1-knockout

GSE260742's VTP-50469 kinetic series carries no SGC0946 sample at any timepoint, so the denominator of F cannot be measured inside that batch and the fraction cannot be recomputed at any time - the only SGC0946 arms in the series are in different experiments at a single 72-hour or 7-day exposure

also called no SGC arm in the timecourse · F's denominator has no kinetics anywhere

The project expresses the menin drug's effect on a histone tag as a fraction of what the DOT1L drug achieves. The dataset follows the menin drug over time but never the DOT1L drug, so the fraction cannot be tracked over time at all - and the two measurements that make it come from different experiments.

The same thing, for a scientist

**`F = (1 - V)/(1 - G)` needs both drugs measured in one batch.** GSE260742 deposits a four-timepoint VTP-50469 H3K79me2 series and **no SGC0946 sample in that experiment at any timepoint** [E1]. Counted over the live records: the series contains **eight** H3K79me2 SGC0946 samples - murine F3 (GSM8123373, 72 h), F3_NT (GSM8123387, 72 h), NT (GSM8123439, 72 h), PCGF1-KO C1 (GSM8123436, 72 h), K562 (GSM8123351, 7 days), MOLM13 (GSM8549856, 96 h), HUDEP-2 (GSM8123346) and mESC (GSM8123358) - **and every one is a single exposure in an experiment that has no timecourse.** **So the DOT1L inhibitor's own kinetics on this mark are never-measured in this deposit, and `F` has no kinetic version at any timepoint.** That is why this beat's ARM 1 computed no `F` at all - by design and declared in its pre-registration rather than as a failure - and it is why closing cycle 82's 48 h/72 h asymmetry cannot be done by re-analysis even in principle from this series. **What the absence costs, precisely.** Cycle 82's next-step item 4 asked whether GSE260742 holds a 72 h VTP H3K79me2 sample that would close the asymmetry by matching the exposures. **It holds no 72 h VTP sample and no timed SGC sample**, so neither route exists: the numerator cannot be moved to 72 h and the denominator cannot be moved to 48 h. **A duration-matched `F` in this system requires a wet experiment**, and its cheapest form is a single additional SGC0946 48 h H3K79me2 ChIP in the F3 line - one sample, which would make `F` exposure-matched at 48 h and turn a lower bound into a point estimate.

Facts and figures
seriesGSE260742
consequenceF has no kinetic version at any timepoint in this deposit, and no VTP H3K79me2 sample in the series is at 72 h, so cycle 82's 48 h/72 h asymmetry cannot be closed by moving either side
cheapest wet fixone additional SGC0946 5 uM 48 h H3K79me2 ChIP in the F3 line, which would make F exposure-matched at 48 h and turn the lower bound into a point estimate
absence classnever-measured
absence class notenever-measured IN THIS SERIES: the samples do not exist, so no re-analysis can create them and no amount of compute closes it. Distinguished deliberately from the timecourse itself, which was measured-not-analysed and which this beat closed. The distinction decides the next step - one is a download, the other is a single ChIP at the bench - and a reader who saw only 'the timecourse exists' would mis-price it.
absence class assigned2026-09-23 (Q124)
vtp h3k79me2 timepoints available
{
  "4h": "GSM8123449",
  "8h": "GSM8123450",
  "16h": "GSM8123447",
  "24h": "GSM8123448",
  "vehicle_24h": "GSM8123452",
  "input": "GSM8123451"
}
sgc0946 h3k79me2 samples and their single exposures
{
  "GSM8123373": "murine F3, 72 h",
  "GSM8123387": "murine F3_NT, 72 h",
  "GSM8123439": "murine NT, 72 h",
  "GSM8123436": "murine PCGF1-KO C1, 72 h",
  "GSM8123351": "K562, 7 days",
  "GSM8549856": "MOLM13, 96 h",
  "GSM8123346": "HUDEP-2",
  "GSM8123358": "mESC"
}
How this could bite us back
  • The figure this corpus now quotes - F = 0.09 to 0.25 - has a denominator measured at 72 hours and a numerator at 48, and this entry establishes that no deposited sample anywhere in the source series can fix either side. Anyone treating F as a point estimate is treating a ratio of two different exposures as a single quantity, and the only route to fixing it is a bench experiment.
  • It also means the kinetic asymmetry will persist in every future beat that uses this series, because it is a property of what was deposited rather than of what was analysed. Recording it as never-measured rather than as an analysis gap is the difference between the next cycle spending a download and the next cycle writing an experiment proposal.
Still open
  1. Has anyone deposited a DOT1L-inhibitor H3K79me2 timecourse in any system? SGC0946 and pinometostat have longer preclinical histories than any menin inhibitor, so this is the likelier of the two timecourses to exist somewhere, and it would bound the denominator's kinetics from outside this series. No cross-series search was run.
  2. Would a single SGC0946 48 h H3K79me2 ChIP in the F3 line be worth proposing to the depositors? It is one sample and it converts a lower bound into a point estimate for a figure this corpus now quotes. That is an unusually cheap ask and nothing in this project has made it.
  3. Do the H2AK119ub and H3K27me3 timecourses, which exist at 8 h and 24 h only, have SGC0946 arms elsewhere in the series that could be paired the same way? The NTvKO and K562 H2AK119ub triplets do contain SGC0946 samples, so an H2AK119ub F is computable at a single exposure - but not kinetically, for the same reason.
Every source behind this entry (1)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    Of GSE260742's 158 per-sample records, the VTP-50469 H3K79me2 timecourse (GSM8123449 4 h, GSM8123450 8 h, GSM8123447 16 h, GSM8123448 24 h, against GSM8123452 DMSO 24 h and input GSM8123451) contains no SGC0946 sample, and the H2AK119ub and H3K27me3 timecourses of the same experiment (GSM8123441/443/442 and GSM8123444/446/445) contain none either; all eight H3K79me2 SGC0946 samples in the series sit in experiments with no timecourse and at a single exposure each - GSM8123373 (F3, 72 h), GSM8123387 (F3_NT, 72 h), GSM8123439 (NT, 72 h), GSM8123436 (PCGF1-KO C1, 72 h), GSM8123351 (K562, 7 days), GSM8549856 (MOLM13, 96 h), GSM8123346 (HUDEP-2) and GSM8123358 (mESC); and no VTP-50469 H3K79me2 sample anywhere in the series carries a 72-hour exposure.

    What this evidence cannot show

    This bounds the deposit and not the study or the field: no cross-series search for a DOT1L-inhibitor H3K79me2 timecourse was run, and one plausibly exists - SGC0946 and pinometostat have longer preclinical histories than any menin inhibitor. The count of eight SGC0946 H3K79me2 samples is from a grep over !Sample_title for SGC across the 158 records combined with the K79 channel, so a sample titled with a different spelling of the compound would be missed; SGC0946, SGC and 0946 were all tried. Exposures are read from the treatment: field of each record.

    Evidence that argues against this (1)
Connected to