Oncofusion

start / graph / exp-two-dot1l-inhibitors-at-matched-global-depth-with-replicates-and-a-spike-in

Experimentmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q138-absolute-h3k79me2

Six to twenty-four libraries in RS4;11: two DOT1L inhibitors titrated to equal spike-in-measured global H3K79me2 loss, three replicates per arm, plus the no-primary control that no public deposit contains - the smallest experiment that would convert four moderate entries in this beat into testable ones

also called the six-library fix · matched-depth two-compound H3K79me2 ChIP-Rx · the missing IgG library

Run the same chromatin experiment three times instead of once, with two different drugs given at doses that remove the same total amount of the mark, and include a control that shows how much signal the antibody produces when there is nothing to find. That last control does not exist in any public dataset we could find, and without it we cannot tell real leftover mark from background.

The same thing, for a scientist

**Design.** RS4;11 (CVCL_0093), the index fusion's line. Dm6 spike-in chromatin at a fixed ratio per cell, added before immunoprecipitation. - **Stage 1, titration (4-6 libraries, single).** EPZ5676 and a second, chemically distinct DOT1L inhibitor, each at 3 concentrations, day 9. Read global H3K79me2 loss against dm6. Pick the two concentrations whose global loss agrees within 20%. - **Stage 2, the experiment (12 libraries).** DMSO, inhibitor A at matched depth, inhibitor B at matched depth, **three biological replicates each**, day 9. Optionally repeat at day 6 (12 more). - **The control that is missing everywhere (3 libraries).** A no-primary-antibody or IgG library from the day-9 DMSO and both treated arms, with the same spike-in. This is what fixes the detection floor and it is the reason the low-baseline end of this beat's retention gradient is uninterpretable. **Readout.** Spike-in-scaled per-gene H3K79me2, scored as this beat's script scores it: median residual across the ten canonical target loci against a baseline-decile-matched permutation null from the same library, with the observed same-condition replicate floor computed from the three DMSO libraries and reported beside it. **What each arm settles.** | arm | question it closes | |---|---| | 3 DMSO replicates | the first MEASURED absolute-scale same-condition floor this project would own; all existing floors are floors on shares | | matched-depth two compounds | whether the residual landscape is compound-specific or a dose artefact | | no-primary libraries | whether the ~30% retention in the least-marked decile is real or the antibody floor | | day 6 as well as day 9 | whether the EPZ day-6 heavy tail seen in both lines of GSE149183 is real |

Go deeper
Why RS4;11 and not MOLM13

Because this project has imported nearly all its chromatin biology from KMT2A::MLLT3 systems, and this beat measured how far that import is safe: the per-gene ordering of H3K79me2 loss transfers between the two lines at Spearman 0.75-0.79, but the magnitude differs by 2.5x in one direction under one drug and 1.4x in the other under the other. Magnitudes are what a pharmacodynamic threshold is made of, so the magnitude has to be measured in the index fusion's own line.

Adding MOLM13 as a second line doubles the cost and would test partner-dependence - but the two lines differ in lineage, donor, sex, age and an NSD2 allele as well as in partner, so it would not isolate the partner and should not be sold as doing so. If a partner comparison is wanted, the right design is an isogenic pair. 1

Cost, and what makes this the cheapest item in the queue

Twelve to twenty-four ChIP-seq libraries in one cell line with one antibody and a commercial spike-in, no animals, no primary material, no new chemistry - except that a second DOT1L inhibitor must be sourced, and the obvious second compound, the 'compound 10' of the source study, has no resolvable public chemical identity. Any structurally distinct DOT1L inhibitor would serve for the compound-specificity question.

This project's standing alternative - a spike-in-normalised histone ChIP under a menin inhibitor - remains the higher-value experiment, because no public deposit puts a menin inhibitor on an absolute scale at all. This one is cheaper and closes a different, narrower set of questions. They are complements, not substitutes. 2

Facts and figures
designRS4;11 + dm6 spike-in. Stage 1: titrate two chemically distinct DOT1L inhibitors to within 20% of the same global spike-in-scaled H3K79me2 loss at day 9 (4-6 single libraries). Stage 2: DMSO / inhibitor A / inhibitor B at matched depth, 3 biological replicates each (12 libraries), optionally repeated at day 6. Plus 3 no-primary-antibody libraries with the same spike-in.
readoutspike-in-scaled per-gene H3K79me2; median residual across HOXA5, HOXA7, HOXA9, HOXA10, MEIS1, PBX3, CDK6, MEF2C, MYB, JMJD1C against a baseline-decile-matched permutation null, with the measured same-condition replicate floor reported beside it
controls3 DMSO replicates for the floor; matched global depth rather than matched nominal concentration; no-primary-antibody libraries for the detection floor; dm6 spike-in in every library including the controls
cost tierlow
timelinestage 1 about 3 weeks including 9 days on drug; stage 2 a further 4-6 weeks
n libraries15 minimum, 27 with the day-6 arm
converts to testable
[
  "hyp-the-residual-h3k79me2-landscape-is-compound-specific",
  "ana-compound-10-spares-the-canonical-targets-and-epz5676-does-not",
  "ana-the-most-marked-genes-are-stripped-most-completely-and-every-decile-still-loses",
  "ana-day-3-already-costs-85-to-90-percent-and-the-collapse-gets-less-uniform-not-more"
]
How this could bite us back
  • If the experiment returns a null - the two compounds leave the same residual landscape - the correct conclusion is that the GSE149183 pattern was four single libraries, and this beat's compound-10 entry must be marked REFUTED rather than narrowed.
  • A measured same-condition floor could turn out to be large enough to swallow several of this project's existing share-based chromatin results, which would be an uncomfortable and valuable outcome.
Still open
  1. Which second DOT1L inhibitor? The source study's compound 10 has no public identity, so the choice has to be made on chemotype distance from EPZ5676 and nothing in this beat can guide it.
  2. Would adding earlier timepoints - 6, 12, 24 hours - be a better use of the same budget? Day 3 is already past 85% loss, so the entire fast phase is unobserved in every public dataset.
  3. Can the missing antibody control be recovered computationally from the existing deposit rather than by new libraries? If so the detection-floor question is answerable today at no bench cost.
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    In GSE149183 the per-gene ordering of H3K79me2 loss transfers between RS4;11 and MOLM13 at Spearman 0.75-0.79 while the median absolute retention differs by 2.53x under EPZ5676 and 0.71x under compound 10, so magnitude does not transfer between lines.

    GSE149183 series supplementary files: RS4_11 and MOLM13 H3K79me2 TSS-ChIPseq signal + dm6 spike-in sitepro tables · Perner F, Gadrey JY, Xiong Y, et al. · NCBI Gene Expression Omnibus 2020
    What this evidence cannot show

    One library per condition, so the between-line difference cannot be tested; the RS4;11 arm is measurable at only 19,214 of 57,794 TSS entries because two-thirds of the deposited RS4;11 table is MOLM13.

    Evidence that argues against this (1)
    • The between-line magnitude difference reverses sign between the two drugs, which is itself the strongest argument that it is technical rather than a property of the lines - and if it is technical, a replicated experiment in one line may find no transferable magnitude at all. Orlando DA et al. 2014
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    GSE149183 contains no IgG or no-primary-antibody library; its only non-ChIP samples are three chromatin inputs (GSM4491615, GSM4491644 MOLM13; GSM4491645 RS4;11), which carry no dm6 spike-in table.

    GSE149183 supplementary filelist.txt and SOFT family file (sample list as served 2026-09-23) · Perner F, Gadrey JY, Xiong Y, et al. · NCBI Gene Expression Omnibus 2020
    What this evidence cannot show

    This is an absence in the deposited sample list; an unshared control could exist. A chromatin input is not an antibody control and cannot substitute for one, because it does not report antibody background.

    Evidence that argues against this (1)
    • Antibody specificity can now be assessed directly from sequencing data without a separate IgG library, which would make the missing control recoverable in principle from the existing deposit rather than requiring new libraries. Kupai A et al. 2023
Connected to