Oncofusion

start / graph / ana-crispr-chip-menin-h3k79-partition-system-resolved-to-murine-mll-af9

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q55-calm-af10

The menin/H3K79 partition this graph leans on was measured in MURINE MLL-AF9 - resolved from GEO, because the abstract does not say

also called CRISPR-ChIP system resolution

A key screen split the H3K79 chemical mark into a menin-dependent half and a menin-independent half. The paper is paywalled and its summary never says which cells were used - so we looked up the data it deposited, which names them: mouse MLL-AF9 leukaemia cells.

The same thing, for a scientist

`hyp-menin-inhibition-spares-endogenous-h3k79-unit` records as its first open question: *'Which MLL fusion was used in the CRISPR-ChIP study? The abstract does not say.'* **Resolved.** The study's four GEO depositions, linked by `Series_pubmed_id 37679565`, name the systems per sample [E1]: the **H3K79me2 CRISPR-ChIP arm and the menin-inhibitor ChIP-seq are murine MLL-AF9 (KMT2A::MLLT3) AML** - sample titles 'MLLAF9 H3K79me2 CRISPR-ChIP day 9', characteristics 'cell line: MLLAF9', 'treatment: menin inhibtor treatment 48hrs', organism *Mus musculus*; the RNA-seq arm containing the **'MLLT10 G1/G2'** guides is likewise murine MLLAF9. The **human** arms are **K562** (CML) and cover **H3K4me3 and Pol II**, not H3K79me2 [E1]. **So the partition that underwrites this graph's expectation of residual H3K79me2 under menin inhibition rests on murine MLL-AF9, and has never been shown in a human cell, in a B-lineage cell, or in KMT2A::AFF1.**

Go deeper
Route, because the obvious one returns zero

Per the brief's corrected GEO rule, esearch db=gds term=37679565[PMID] returns a well-formed count of 0. elink dbfrom=pubmed&db=gds&id=37679565 returns uids 200192562, 200192561, 200192555, 200192544 - GSE192562, GSE192561, GSE192555, GSE192544. Per-sample SOFT records via acc.cgi?acc=<GSE>&targ=gsm&form=text&view=brief then give organism, cell line and treatment for each sample 1. The paper itself is unreadable: Europe PMC reports no PMCID, and route 5 (DOI resolution to the publisher) returns a 303-character stub.

Attribution is by Series_pubmed_id, not by topic - all four series carry 37679565.

What this changes

Nothing about the partition's validity; everything about how it should be stated. Per rule 73, any claim built on it should open 'In murine MLL-AF9 AML, H3K79 methylation partitions into...'. Three consequences follow for this beat and for the histone-marks beat:

  1. The prediction that menin inhibition spares an MLLT10-directed H3K79 unit is a murine MLL-AF9 prediction.
  2. Its transfer to CALM-AF10 crosses species, lineage and fusion partner at once, and this beat's own hypothesis says so in its caveat.
  3. Its transfer to KMT2A::AFF1, this project's index fusion, is untested - which is exactly the fourth falsifier the existing hypothesis entry already lists.
Facts and figures
scriptrun ad hoc from the beat scratchpad 2026-09-20 using NCBI elink and the GEO acc.cgi SOFT interface
statusrun
result summaryThe H3K79me2 / MLLT10 / menin-inhibitor arms of the CRISPR-ChIP study are murine MLL-AF9 AML. The human arms are K562 and cover H3K4me3 and Pol II. The partition has not been demonstrated in a human cell, in B-lineage cells, or in KMT2A::AFF1.
resolution routeelink pubmed->gds (esearch db=gds with [PMID] returns a well-formed zero and must not be used)
answers open question inhyp-menin-inhibition-spares-endogenous-h3k79-unit (owned by the histone-marks beat; reported, not edited)
inputs
[
  "elink dbfrom=pubmed&db=gds&id=37679565",
  "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE192562&targ=gsm&form=text&view=brief (and the other three series)"
]
geo series
[
  "GSE192562",
  "GSE192561",
  "GSE192555",
  "GSE192544"
]
How this could bite us back
  • A system-resolution result is easy to over-apply in the other direction: 'murine MLL-AF9' does not mean the partition is wrong, only that its generality is unmeasured.
Still open
  1. Were any non-deposited arms of the study run in human or B-lineage cells? The GEO route bounds the deposited sequencing only.
  2. Does the partition hold in KMT2A::AFF1? That is the fourth falsifier of the existing hypothesis entry and remains untested.
  3. Does it hold in a fusion where MLLT10 IS the partner - KMT2A::MLLT10 or PICALM::MLLT10 - where the two units cannot be separate by construction?
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    The four GEO series carrying Series_pubmed_id 37679565 (GSE192562, GSE192561, GSE192555, GSE192544) show that the H3K79me2 CRISPR-ChIP, the MLLT10-guide RNA-seq and the menin-inhibitor ChIP-seq samples are murine MLLAF9 AML (organism Mus musculus; sample characteristics 'cell type: MLL-AF9', 'cell line: MLLAF9', 'treatment: menin inhibtor treatment 48hrs'), while the human samples are K562 CML and carry H3K4me3 and Pol II readouts.

    What this evidence cannot show

    This bounds THE DEPOSITED SEQUENCING, not the whole study: an experiment performed and reported without deposition would be invisible to this route. The sample-level annotation is the depositors' own and includes a typo ('inhibtor') which is reproduced here verbatim. Retrieved 2026-09-20 via elink and per-sample SOFT records.

    Evidence that argues against this (1)
    • The GEO route bounds the DEPOSITED SEQUENCING only, and the same study deposits human K562 samples for its H3K4me3 and Pol II arms - so the paper is not a mouse-only study, and a claim that 'the partition is murine' must be narrowed to 'the H3K79me2 and menin-inhibitor arms are murine'. Non-sequencing experiments in human cells would be invisible to this method. Gilan O et al. 2023
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The paper reports a functional partition of H3K79 methylation into a menin-dependent, MLL-fusion-directed oncogenic DOT1L complex and a separate endogenous DOT1L complex directed by MLLT10.

    PMID 37679565DOI
    What this evidence cannot show

    ABSTRACT ONLY. Europe PMC returns no PMCID and DOI resolution returns a 303-character stub, so routes 1-5 of the readability chain all fail for the full text. The abstract states the partition and does not state the system, which is what made the GEO route necessary.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT, and it is about how this partition transfers rather than about the system it was measured in. The two-arm model - a menin-dependent, MLL-fusion-directed oncogenic DOT1L complex and a separate endogenous DOT1L complex directed by MLLT10 - has no slot for a lesion in which MLLT10 IS the oncogenic partner. Reading it as 'CALM-AF10's H3K79 runs through the menin-independent arm' assigns PICALM::MLLT10 to the arm that was defined by what MLLT10 does when it is NOT fused. The one MLLT10-containing fusion whose menin dependence has actually been measured points the other way: KMT2A::MLLT10 patient-derived xenografts, described by the authors as high-risk fusion-partner AMLs, were sensitive to menin inhibition. MLLT10 in a fusion is therefore compatible with menin dependence, and this partition cannot be used to predict otherwise. Mahdavi L et al. 2026
Connected to