Oncofusion

start / graph / hyp-q124-h3k79me2-reports-the-dot1l-axis-and-not-the-prc1-1-axis-of-menin-inhibitor-biology

Hypothesisspeculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.Q124-the-timecourse-and-the-pcgf1-knockout

H3K79me2 reports only the DOT1L arm of a menin inhibitor's action and is blind to its PRC1.1 arm - so the mark's response is unchanged in a PCGF1 knockout even though PCGF1 loss is a published resistance mechanism, and a cell resistant through PRC1.1 loss would look fully drug-engaged by this readout

also called the mark reports one arm of two · H3K79me2 cannot see PRC1.1-mediated resistance · hyp-q124

The menin drug does at least two things to chromatin: it takes the fusion off the DNA, and a repressive complex moves into the vacated spots. The H3K79 tag tracks the first of these and not the second. That matters because losing a component of that repressive complex is a published way for the drug to stop working - and cells that fail that way would still look like the drug was engaged if you measured the tag.

The same thing, for a scientist

**Statement.** A menin inhibitor's chromatin action in KMT2A-rearranged leukaemia has at least two arms: **subtraction** of the fusion from its sites, with the DOT1L-dependent H3K79me2 that its recruitment supports, and **local substitution** by PRC1.1 in the vacated territory. **H3K79me2 reports the first arm and is invariant to the second.** Therefore the mark cannot distinguish a drug-sensitive cell from one made resistant by loss of a PRC1.1 subunit. **Origin: data-pattern plus published literature.** This beat measured that in murine MLL-AF9, knocking out PCGF1 changes the menin inhibitor's H3K79me2 share ratio at the fusion's own sites by **0.0166** and the DOT1L inhibitor's by **0.0101**, both below the **0.0584** that two control cultures of the same line already differ by [E1]. Independently, a published chromatin-focused CRISPR screen shows that **loss of PCGF1 or BCOR confers menin inhibitor resistance** by reactivating bivalent noncanonical menin targets including MYC, with a reduced monocytic signature and BCL2-inhibitor susceptibility [E2]. **Put together: PCGF1 loss changes what the drug does to the cell without changing what it does to the mark.** **Prior probability: 0.55.** The measurement is a clean null in a pre-registered criterion with a measured rather than invented threshold, and it is reinforced by the DOT1L-inhibitor arm, which needs no fusion and no menin and gives the same answer at 0.0101. What holds the prior down is that **the knockout is unverified in the files analysed** - one guide, one clone, no rescue, no deposited PCGF1 measurement - and that the two-arm decomposition itself is this corpus's model rather than an established one. **The competing hypothesis this must beat** is that the two arms are not separable: that PRC1.1 in-fill is downstream of, and proportional to, the eviction the mark reports, so a PRC1.1-deficient cell still shows normal eviction and normal H3K79me2 loss for the good reason that both are upstream of the missing step. **That model predicts exactly the null this beat measured**, and it is not distinguished by the H3K79me2 arm at all. **What distinguishes them is the H2AK119ub arm**, which is deposited for this exact 2 x 3 design and unanalysed: under the separable-arms model the VTP-induced H2AK119ub gain should be abolished in the knockout while H3K79me2 is unchanged; under the downstream model the gain is abolished and the mark is unchanged too - **so the discriminating measurement is not the knockout's H2AK119ub but whether anything in the knockout's chromatin responds to the drug at all.**

Go deeper
CORRECTION MADE BEFORE THIS ENTRY WAS FINISHED: the published model already predicts this null, so the novel content is narrower than the title suggests

Written after running scripts/check-citation-reuse.py on this entry's own PMIDs, which is what rule 82 is for. PMID 39158067 is cited 54 times in this corpus and PMID 36635503 43 times, and reading what they were cited for changed what this hypothesis can claim.

mechanism-menin-inhibitor-silencing-is-prc1-1-deposition-not-only-eviction already records the published ordering: the PRC1.1-dependent H2AK119ub that silences fusion targets under a menin inhibitor arises from progressive loss of DOT1L-mediated H3K79 methylation. That is H3K79me2 loss upstream of PRC1.1 in-fill. A downstream perturbation cannot change an upstream quantity, so removing PCGF1 SHOULD leave H3K79me2's drug response untouched - which is exactly what this beat measured.

So the null is predicted by the published model and is not evidence for a novel mechanism. What this hypothesis actually adds, stated without the inflation:

  1. A quantitative bound where there was a qualitative expectation. The published ordering predicts 'unchanged'; this measurement says the change is 0.0166 for the menin inhibitor and 0.0101 for the DOT1L inhibitor, against a 0.0584 two-culture spread. An ordering that predicts a null is confirmed more strongly by a number than by an absence.
  2. The biomarker consequence, which nobody has drawn. If H3K79me2 sits upstream of the PRC1.1 step, and PRC1.1 loss is the published resistance mechanism, then the mark reports engagement in the genotype where the drug fails. This corpus proposed H3K79me2 as a pharmacodynamic readout for a menin inhibitor on the strength of one cell line's number; that proposal is what this entry opposes.
  3. The complement of an existing hypothesis rather than a rival to it. hyp-prc1-1-is-effector-and-gatekeeper-of-one-activity proposes that H2AK119ub should be the menin inhibitor's pharmacodynamic biomarker. This entry supplies the other half of that argument by measuring that H3K79me2 cannot be, and the two should be read together.

The prior stays at 0.55, and the reason is now different from the reason given below: not because the mechanism is uncertain - the published ordering supports it - but because the measurement that establishes it rests on a knockout this project cannot verify, and because the corpus's own model already implies it, so a confirmation adds less than a surprise would.

What follows if this is right

A biomarker consequence, and it is the sharp end. H3K79me2 has been proposed in this corpus as a better pharmacodynamic readout for a menin inhibitor than an MLL1 ChIP, on the strength of a single cell line's number. If the mark reports only the DOT1L arm, then a patient whose disease is resistant through PRC1.1 loss - a recurrent lesion class in myeloid malignancy, since PCGF1 and BCOR loss-of-function are recurrent - would show normal H3K79me2 pharmacodynamics on drug. The readout would say 'engaged' and the disease would progress.

A combination consequence. If the two arms are separable, then an agent acting on the PRC1.1 arm and one acting on the DOT1L arm are not redundant, and the published resensitisation of PRC1.1-deficient cells by MYC inhibition or venetoclax is acting on the arm the mark cannot see. That argues for pairing a menin inhibitor with a PRC1.1-arm agent in exactly the genotype where the menin inhibitor alone fails.

A measurement consequence for this project. Every entry here that treats H3K79me2's fall as a readout of 'the drug working' is reading one arm of at least two, and the corpus contains several.

Falsifiers

Each names its population, its instrument and what that instrument cannot distinguish, because a falsifier that names only a threshold has failed once in this corpus already.

F1 - THE MARK DOES DEPEND ON PRC1.1 AT THE LOCI WHERE PRC1.1-MEDIATED RESISTANCE ACTS. Population: murine MLL-AF9 (cell line: MLLAF9, myeloid, blood, taxid 10090), the non-targeting culture NT (GSM8123438/440) against PCGF1-KO clone C1 (GSM8123435/437), restricted to the 2 kb bins overlapping the bivalent noncanonical menin targets named in PMID 39158067 rather than the genome-wide fusion set, 500 nM VTP-50469 48 h against DMSO 72 h. Fires if the absolute difference in H3K79me2 share ratio between the genotypes at that restricted set is >= 0.2089 (Q120's cross-partner delta, the effect size this corpus treats as meaningful for this mark), band +/-20%. Consequence: the mark then does report the PRC1.1 arm, at the loci where that arm matters, and its genome-wide invariance is an averaging artefact - which would make the biomarker concern disappear and replace it with a locus-selection problem. Instrument and blindness: ABCAM ab3594 rabbit polyclonal, which reads a mark and cannot distinguish halted deposition from demethylation from replicative dilution, cannot separate me2 from me3, and cannot attribute anything to the fusion; and the bins are already on disk, so this costs one script and a gene-coordinate resolver.

F2 - NOTHING IN THE KNOCKOUT'S CHROMATIN RESPONDS TO THE DRUG, so the null is about the clone and not about the mark. Population: the same two genotypes, H2AK119ub channel, GSM8123399/401/402 (KO) against GSM8123403/405/406 (control) in the paired convention, at the fusion's own FLAG-defined top-0.1% set. Fires if the VTP-induced H2AK119ub share ratio in the knockout is within 0.05 of the control's, i.e. the knockout shows the same PRC1.1 in-fill as its control. Consequence: the clone did not lose PRC1.1 function, criteria P1 and P3 measured nothing, and this hypothesis has no evidence. This is the internal validation the deposit can supply for its own genotype and it is twelve downloads. Instrument and blindness: the H2AK119ub antibody is NEB 8240S clone D27C4 per the key-resources sheet and cannot distinguish PRC1.1 from canonical PRC1 deposition, nor deposition from loss of a deubiquitinase, so F2 not firing shows a drug response and not specifically a PRC1.1 response.

F3 - THE ARMS ARE NOT SEPARABLE BECAUSE PRC1.1 IN-FILL IS PROPORTIONAL TO EVICTION. Population: murine MLL-AF9 F3, the FLAG-defined top-0.1% set stratified into eviction quartiles as Q117 did, H2AK119ub under 500 nM VTP-50469 48 h. Fires if the H2AK119ub gain per eviction quartile is monotone in eviction depth with a Spearman >= 0.80 across quartiles in the control genotype and the same monotone relation holds in the PCGF1 knockout. Consequence: in-fill is then a passive consequence of vacancy rather than a PRC1.1-gated step, the two-arm decomposition collapses, and this hypothesis is a restatement of 'the mark reports eviction' with no extra content. Instrument and blindness: as F2; and Q117 measured the control-genotype half already (RING1B 1.38x in the most-evicted quartile against 0.81-0.88x elsewhere), so only the knockout half is new.

F4 - THE MARK IS BLIND TO EVERYTHING, NOT SELECTIVELY BLIND. Population: murine MLL-AF9, all four cultures measured by Q122 and this beat, the mark's response to any genetic perturbation of any chromatin complex in this series. Fires if the H3K79me2 drug response is invariant to a perturbation that demonstrably DOES change the drug's effect on another mark in the same samples - i.e. if the mark's invariance is a property of the readout's dynamic range rather than of the biology. Consequence: the hypothesis's specific claim (reports the DOT1L arm, blind to the PRC1.1 arm) becomes the unremarkable claim that a within-sample share at a fixed set has poor sensitivity, and this beat's own measurement that such a share is confounded with library enrichment at Spearman -1.000 over four timepoints makes that live. Instrument and blindness: the confound is measurable and reportable - the drug-to-vehicle non-zero bin count ratio, 0.986-1.110 in the four 48 h contrasts here - and the only genetic perturbation available in this series is the PCGF1 knockout, so F4 cannot be scored from GSE260742 and needs a second knockout arm that does not exist.

Facts and figures
statementH3K79me2 reports the DOT1L arm of a menin inhibitor's chromatin action and is invariant to its PRC1.1 arm, so it cannot distinguish a drug-sensitive cell from one made resistant by loss of a PRC1.1 subunit
origindata-pattern (this beat's PCGF1-knockout null) combined with published literature (PCGF1/BCOR loss confers menin-inhibitor resistance)
prior probability0.55
prior probability note0.55, and the reason CHANGED while this entry was being written - see the correction section. NOT because the mechanism is uncertain: mechanism-menin-inhibitor-silencing-is-prc1-1-deposition-not-only-eviction already records the published ordering in which PRC1.1 H2AK119ub arises FROM progressive loss of DOT1L-mediated H3K79 methylation, i.e. the mark is upstream, and a downstream perturbation cannot move an upstream quantity, so the null this beat measured is PREDICTED rather than surprising. The prior is held at 0.55 because the measurement establishing it rests on a knockout this project cannot verify - one guide, one clone, no rescue, no deposited PCGF1 measurement - and because a confirmation of the corpus's own implied model adds less than a surprise would. What is genuinely new is the QUANTITATIVE bound (0.0166 and 0.0101 against a 0.0584 two-culture spread) and the BIOMARKER consequence, which nobody has drawn.
competing hypothesisPRC1.1 in-fill is downstream of, and proportional to, the eviction H3K79me2 reports, so a PRC1.1-deficient cell shows normal eviction and normal H3K79me2 loss because both are upstream of the missing step. THIS IS NOT ONLY A COMPETING HYPOTHESIS, IT IS THE PUBLISHED MODEL AND THIS CORPUS ALREADY HOLDS IT: mechanism-menin-inhibitor-silencing-is-prc1-1-deposition-not-only-eviction records that the PRC1.1-dependent H2AK119ub arises from progressive loss of DOT1L-mediated H3K79 methylation. It predicts exactly the null measured here and is NOT distinguished by the H3K79me2 arm at all. Sparbier et al. (PMID 36635503) supplies a second, stronger version in which PCGF1-PRC1.1 and Menin-KMT2A/B are one system at bivalent promoters. The discriminating measurement is therefore NOT this beat's arm but the H2AK119ub arm (falsifier F2), and what this entry adds over the published model is a quantitative bound and a biomarker consequence rather than a new ordering.
shown for fusionKMT2A::MLLT3 in a murine FLAG-tagged over-expression model; nothing in the measurement half is human and nothing is KMT2A::AFF1. The published half (PMID 39158067) covers human and mouse KMT2A-rearranged AML models
epigenetic readoutH3K79me2's response to both VTP-50469 and SGC0946 at the fusion's own top-0.1% sites is invariant to PCGF1 knockout (differences 0.0166 and 0.0101 against a 0.0584 two-culture spread), while PCGF1 loss is a published menin-inhibitor resistance mechanism
mark writtenH3K79me2, by DOT1L
writer of that markgene-dot1l
absence classmeasured-not-analysed
absence class noteBoth of the two falsifiers that can discriminate this hypothesis from its competitor are measured-not-analysed. F2 needs twelve downloads of H2AK119ub files public since 2025-11-23; F1 needs one script over bins already on C:/Temp/q124 plus a gene-coordinate resolver. F4 is never-measured in this series because the PCGF1 knockout is the only genetic perturbation deposited. Recording the classes separately matters because F2 is the cheap one and it is also the one that could kill the hypothesis outright.
absence class assigned2026-09-23 (Q124)
predictions
[
  "the VTP-induced H2AK119ub gain at the fusion's most-evicted bins is abolished or strongly reduced in the PCGF1 knockout while H3K79me2 is unchanged - twelve downloads, accessions in ana-q124-prediction-5-of-hyp-q117-has-three-clauses-and-an-h3k79me2-arm-can-reach-one",
  "H3K79me2 at the bivalent noncanonical menin targets of PMID 39158067 does NOT distinguish the genotypes either, so the mark's blindness holds at the loci where PRC1.1-mediated resistance acts - one script, bins already on disk",
  "in a human KMT2A-rearranged line rendered resistant by PCGF1 or BCOR loss, H3K79me2 pharmacodynamics on a menin inhibitor are indistinguishable from a sensitive line's",
  "the two arms respond to different combination partners: MYC or BCL2 inhibition resensitises the PRC1.1 arm and a DOT1L inhibitor does not"
]
falsifiers
[
  "F1 - THE MARK DOES DEPEND ON PRC1.1 AT THE LOCI WHERE PRC1.1-MEDIATED RESISTANCE ACTS. **THE LOCI ARE ALREADY PUBLISHED AS COORDINATES AND THIS GRAPH ALREADY FOUND THEM**, so this falsifier needs no gene-name-to-coordinate resolution and no new sequencing download - though it does need a human-to-mouse ortholog map, see the HAZARD below: `ana-the-locus-list-existed-at-the-publisher-not-in-geo` records that Zhou supplemental Table 4, BLOOD_BLD-2023-023644-mmc4.xlsx (356 KB, via the Europe PMC supplementaryFiles endpoint for PMC11561541), is chr|start|end|gene id|group over 9,292 rows - Menin solo 3,423, Menin+H2AK119ub shared 1,733, H2AK119ub solo 4,136 - and that the first two counts reproduce the paper's own body text exactly. Population: murine MLL-AF9 (cell line MLLAF9, myeloid, blood, taxid 10090), non-targeting culture NT (GSM8123438/440) against PCGF1-KO clone C1 (GSM8123435/437), restricted to the 2 kb mm10 bins overlapping the 1,733 Menin+H2AK119ub SHARED loci of that table - the bivalent noncanonical menin targets - 500 nM VTP-50469 48 h against DMSO 72 h. Fires if the absolute difference in H3K79me2 share ratio between genotypes at that restricted set is >= 0.2089 (Q120's cross-partner delta), band +/-20%. Consequence: the mark does report the PRC1.1 arm at the loci where that arm matters, its genome-wide invariance is an averaging artefact, and the biomarker concern becomes a locus-selection problem instead. Instrument: ABCAM ab3594 rabbit polyclonal, which reads a mark and cannot distinguish halted deposition from demethylation from replicative dilution, cannot separate me2 from me3, and cannot attribute anything to the fusion. HAZARD: Zhou's coordinates are HUMAN (hg38) and this beat's bins are MURINE (mm10), so the restriction requires orthologous gene mapping rather than a coordinate intersection - which is a real extra step and is the reason this is not a one-line job. COST: one script plus an ortholog map; the bins are on C:/Temp/q124 and the table is one Europe PMC request. NOT YET RUN.",
  "F2 - NOTHING IN THE KNOCKOUT'S CHROMATIN RESPONDS TO THE DRUG, so the null is about the clone and not about the mark. Population: the same two genotypes, H2AK119ub channel, GSM8123399/401/402 (KO) against GSM8123403/405/406 (control), paired convention, at the FLAG-defined top-0.1% set. Fires if the VTP-induced H2AK119ub share ratio in the knockout is within 0.05 of the control's. Consequence: the clone did not lose PRC1.1 function, P1 and P3 measured nothing, and this hypothesis has no evidence. Instrument: H2AK119ub is NEB 8240S clone D27C4 per the key-resources sheet, and it CANNOT distinguish PRC1.1 from canonical PRC1 deposition nor deposition from loss of a deubiquitinase, so F2 not firing shows a drug response and not specifically a PRC1.1 response. COST: twelve downloads. NOT YET RUN, AND IT IS THE FIRST THING THAT SHOULD BE.",
  "F3 - THE ARMS ARE NOT SEPARABLE BECAUSE PRC1.1 IN-FILL IS PROPORTIONAL TO EVICTION. Population: murine MLL-AF9 F3, FLAG-defined top-0.1% set stratified into eviction quartiles as Q117 did, H2AK119ub under 500 nM VTP-50469 48 h, in both genotypes. Fires if the H2AK119ub gain per eviction quartile is monotone in eviction depth at Spearman >= 0.80 in the control AND the same monotone relation holds in the PCGF1 knockout. Consequence: in-fill is a passive consequence of vacancy rather than a PRC1.1-gated step and the two-arm decomposition collapses. Instrument: as F2. Q117 measured the control-genotype half already (RING1B 1.38x in the most-evicted quartile against 0.81-0.88x elsewhere), so only the knockout half is new. NOT YET RUN.",
  "F4 - THE MARK IS BLIND TO EVERYTHING, NOT SELECTIVELY BLIND. Population: murine MLL-AF9, all four cultures, the mark's response to any genetic perturbation of any chromatin complex in this series. Fires if the H3K79me2 drug response is invariant to a perturbation that demonstrably DOES change the drug's effect on another mark in the same samples. Consequence: the specific claim becomes the unremarkable claim that a within-sample share at a fixed set has poor sensitivity - made live by this beat's own measurement that such a share ranks inversely to library complexity at Spearman -1.000 over four timepoints. Instrument: the confound is reportable as the drug-to-vehicle non-zero bin count ratio, 0.986-1.110 in the four 48 h contrasts. CANNOT BE SCORED FROM GSE260742: the PCGF1 knockout is the only genetic perturbation in the series, so F4 needs a second knockout arm that does not exist."
]
analyses
[
  "ana-q124-the-menin-inhibitors-h3k79me2-effect-is-unchanged-in-a-pcgf1-knockout",
  "ana-q124-prediction-5-of-hyp-q117-has-three-clauses-and-an-h3k79me2-arm-can-reach-one",
  "ana-q124-the-two-drugs-move-h3k79me2s-mid-occupancy-floor-in-opposite-directions-in-four-cultures"
]
experiments
[
  "exp-q124-the-h2ak119ub-arm-of-the-pcgf1-knockout-is-twelve-downloads-and-scores-two-falsifiers-at-once"
]
How this could bite us back
  • RESISTANCE. If this is right, the genotype class in which a menin inhibitor is published to fail - PCGF1 or BCOR loss, recurrent in myeloid malignancy - is exactly the class in which the chromatin readout this corpus has proposed as a biomarker would report normal drug engagement. A pharmacodynamic assay that is blind in the resistance genotype is worse than none, because it produces a false reassurance rather than an absence of information.
  • COMPENSATION. The hypothesis says the two arms are separable, which means a cell can lose one and keep the other. That is a compensation route and it is the published one: PRC1.1-deficient cells reactivate noncanonical menin targets and survive menin inhibition, while their DOT1L axis - the arm the mark reports - behaves normally.
  • DELIVERY AND TOXICITY. If combination design follows from this, the partner has to act on the PRC1.1 arm, and the published resensitisers are MYC inhibition and venetoclax. Venetoclax plus a menin inhibitor is already a combination under active clinical investigation, so the hypothesis's therapeutic consequence is mostly a rationale for something already being done rather than a new proposal - which is worth saying plainly rather than claiming novelty.
  • ESSENTIAL FUNCTION. PCGF1 and PRC1.1 are not leukaemia-specific, and any agent acting on the arm the mark cannot see would be acting on a complex with developmental roles. This hypothesis provides no toxicity argument either way and should not be read as supplying one.
Still open
  1. Does H3K79me2 distinguish the two genotypes at Zhou's 1,733 Menin+H2AK119ub SHARED loci? The coordinates are published (BLOOD_BLD-2023-023644-mmc4.xlsx, 9,292 rows) and this graph already located them, so falsifier F1 needs one script plus a human-to-mouse ortholog map and no new download. It is the cheapest test of this hypothesis and it was priced wrongly in the first draft of this entry, which claimed a coordinate resolver was needed.
  2. Did PMID 39158067 measure H3K79me2 in its PRC1.1-deficient cells? If it did, that number is directly comparable to this beat's and should be read before this hypothesis is developed further. Only the abstract was read here.
  3. Is the two-arm decomposition a real separation or an artefact of measuring two marks with two antibodies of different dynamic range? F4 names the test and the series cannot supply it.
  4. If H3K79me2 reports only the DOT1L arm, what is the right pharmacodynamic readout for the PRC1.1 arm in a patient sample? H2AK119ub ChIP on a marrow aspirate is not a clinical assay, and the transcriptional signature that PMID 39158067 uses - reactivation of bivalent noncanonical menin targets including MYC - may be measurable by RNA and is the more practical candidate. Nothing in this corpus has proposed it.
  5. Does the same one-arm blindness apply to the MLL1-N and MENIN ChIP readouts this corpus has used? Both were measured in a PCGF1-wild-type background only, in both series, so nobody knows.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    In murine MLL-AF9 (KMT2A::MLLT3) myeloid cells, 2 kb bins over a 1,227,708-bin mm10 universe, at the FLAG-DMSO top-0.1% set: the H3K79me2 within-sample share ratio under 500 nM VTP-50469 48 h against DMSO 72 h is 0.8156 in a non-targeting-sgRNA control culture and 0.7990 in an isogenic PCGF1-knockout clone (difference 0.0166), and under 5 uM SGC0946 72 h it is 0.1415 and 0.1314 (difference 0.0101), both below the 0.0584 absolute difference two non-knockout control cultures of the same line give at the same set.

    What this evidence cannot show

    THE KNOCKOUT IS UNVERIFIED IN THE FILES ANALYSED: one guide, one clone, no rescue, and no deposited PCGF1 protein, transcript or locus measurement, so this null cannot be separated from an incomplete edit. The VTP arm is 48 h against a 72 h vehicle. THIS IS ONE MARK IN ONE SPECIES IN A MYELOID MODEL, not the infant B-lineage KMT2A::AFF1 disease. 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 (2)
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In human and mouse KMT2A-rearranged AML models, loss of PRC1.1 subunits - PCGF1 or BCOR - confers resistance to menin inhibition by epigenetically reactivating a group of noncanonical menin targets that are bivalently co-occupied by active menin and repressive H2AK119ub and are normally downregulated after menin inhibition, including MYC; PRC1.1-deficient cells show reduced monocytic gene signatures and susceptibility to BCL2 inhibition, and venetoclax overcomes their menin-inhibitor resistance.

    PMID 39158067DOI
    What this evidence cannot show

    Read from the abstract fetched from PubMed this session; the full text was not read, so the specific loci, the effect sizes and whether the resistance was measured in the same murine MLL-AF9 background as GSE260742's NTvKO arm are not established here. The paper reports H2AK119ub and transcriptional readouts; whether it measured H3K79me2 at all in its PRC1.1-deficient cells is not known to this beat, and if it did, its result would be directly comparable to E1 and should be checked before this hypothesis is developed further.

    Evidence that argues against this (1)
Connected to