Oncofusion

start / graph / absence-no-histone-mark-readout-exists-for-any-nurd-or-sin3-subunit-in-a-kmt2a-rearranged-cell

Analysismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q193-the-nurd-subunits

No histone-mark readout has ever been measured after perturbing ANY NuRD or Sin3 subunit in a KMT2A-rearranged cell - the two complexes whose entire job is acetylation have zero acetylation data in this disease

also called the NuRD epigenetics absence · Q193 epigenetics-dimension absence

NuRD and Sin3 exist to take acetyl marks off histones. In leukaemia driven by a KMT2A fusion, nobody has ever removed one of their parts and then measured what happened to any histone mark. The complexes are discussed as chromatin machines in a disease where their chromatin output has never been looked at.

The same thing, for a scientist

This graph's standing instruction names **epigenetics** as one of five saturation dimensions, and the corpus has measured itself as thin on it - 13.2% of gene entries by a dedicated field, 46.6% by any specific mark named anywhere. This entry bounds the dimension for an entire pair of complexes. Across NuRD (14 canonical subunits) and Sin3/HDAC (17), **no published experiment perturbs any subunit in a KMT2A-rearranged cell of any lineage and reports a histone-mark measurement** [E1]. The nearest approaches all fall short in a specific, nameable way: the CRISPR screen that made MBD3 and ARID4B vulnerabilities in KMT2A::AFF1 lines measured **proliferation only** and tabulated no chromatin readout [E2]; the CHD4 work in MLL/AF4 lineage switching measured gene expression and myeloid phenotype, not marks; and the mark-level results that do exist for these subunits - p66-alpha's histone-tail binding being blocked by tail acetylation, the RUNX1/PRMT6/Sin3a complex's H3R2me2a with H3K27me3 and H3K4me2 but not H3K4me3 - are in non-leukaemic or non-fusion systems [E3][E4]. The class is **never-measured**: no instrument has been pointed at it, and closing it needs wet work, not an unopened file. That makes it the opposite of this beat's interactome absences, which are `measured-not-reported` and closable by compute - and stating both classes side by side is the point of the corpus's absence-class convention.

Go deeper
What was checked, and what would count as closing it

The claim being bounded: a published measurement of any histone mark - acetylation, methylation, anything - in a KMT2A-rearranged cell, after perturbing any subunit of NuRD or Sin3/HDAC.

Literature: zero records tie MTA2 or GATAD2A to a KMT2A fusion under current or retired symbols; MBD3 returns one (the CRISPR screen); SIN3A returns four, three of which fail a title check; the complex-level pair returns 0 for NuRD with KMT2A and 6 for NuRD with MLL, none of which is a mark measurement after a subunit perturbation 1.

Archives: GEO was NOT searched for ChIP-seq after NuRD- or Sin3-subunit perturbation in a KMT2A-rearranged line. This is recorded as not done rather than passed over, because a beat that names two archives and skips a third produces a stronger-looking bound than it earned. It is the single cheapest thing that could weaken this absence, and it is the first thing a follow-up should run.

What would close it: an H3K27ac or H3K9ac ChIP-seq (or a CUT&RUN) in SEM or RS4;11 after MBD3 knockout or CHD4 degradation, compared against the KMT2A::AFF1-bound promoter set this project already carries. Days of wet work, not years - never-measured here means nobody looked, not that looking is hard.

Why this absence matters more than most

Because a deacetylase complex with no acetylation data invites exactly the inference this graph has already warned itself against. An existing entry's bite_back records that treating complex membership as transferring a chromatin readout to the scaffold "would credit large parts of the corpus on the epigenetics dimension without any new measurement." NuRD and Sin3 are the highest-risk case for that error in the entire corpus: their names contain their enzymatic activity, so an entry can sound like it carries an epigenetic readout while carrying only a membership claim.

The entries written by this beat therefore each carry an explicit epigenetic_readout field describing what has been measured and in which system, plus an epigenetic_readout_absence_class of never-measured where the answer is nothing. That is a deliberate choice against the easier route of writing 'histone deacetylation' into the field and moving on.

And there is a specific claim at risk. This beat's own measurement finds HDAC1 and HDAC2 absent from every column of the pan-MLL-fusion interactome. If any entry in this graph ever states that deacetylation happens at a KMT2A fusion, it will not be able to source that to the interactome dataset, and - on this bound - it will not be able to source it to anything else either.

Facts and figures
statusrun
absence classnever-measured
absence class noteNEVER-MEASURED, and the contrast with this beat's other absences is the reason to state it. The interactome absences (MTA2, MBD3, GATAD2A, SIN3A and nine more in the MLL-AF4 AP-MS column) are measured-not-reported, upgradeable to measured-not-analysed against public raw files - closable by compute. THIS one needs an experiment: no ChIP, CUT&RUN, mass-spectrometric histone panel or accessibility assay has ever followed a NuRD or Sin3 subunit perturbation in a KMT2A-rearranged cell. Three orders of magnitude apart in cost, and a beat that used one class for both would misstate what it would take to fix.
what would close itH3K27ac or H3K9ac ChIP-seq or CUT&RUN in SEM or RS4;11 after MBD3 knockout or CHD4 degradation, scored against the KMT2A::AFF1-bound promoter set this project already carries. Wet work, but days rather than years - never-measured here means nobody looked, not that looking is hard.
shown for fusionNO for every subunit of both complexes - nobody has looked in any fusion, for this dimension, in the strict sense this corpus reserves for that phrase.
relationship to existing entriesThe epigenetics-dimension counterpart to absence-no-experiment-has-ever-asked-whether-a-nurd-scaffold-subunit-binds-a-kmt2a-fusion, which bounds the BINDING question and carries a different absence_class. Each gene entry written by this beat carries its own epigenetic_readout and epigenetic_readout_absence_class fields, using the corpus convention rather than a new spelling; this entry states the complex-wide bound once so those fields do not each have to argue it.
scope
{
  "complexes": "NuRD (14 canonical subunits) and Sin3/HDAC (17 canonical subunits)",
  "perturbation": "any - knockout, knockdown, degradation, mutation",
  "cell": "any KMT2A-rearranged cell of any lineage, human or mouse, line or primary",
  "readout": "any histone mark"
}
nearest approaches and how each falls short
{
  "Schneider 2023 CRISPR screen": "RIGHT cells (SEM and ALL-PO, KMT2A::AFF1), RIGHT perturbation (MBD3 and ARID4B knockout) - WRONG readout: proliferation only, no chromatin measurement of any kind.",
  "Tirtakusuma 2022 lineage switching": "RIGHT fusion and right disease (human MLL/AF4 pro-B ALL relapsing as AML, CHD4 perturbation in MLL/AF4 cell models) - readout is gene expression and myeloid phenotype, not marks. Cited in gene-chd4, not re-read here.",
  "Brackertz 2006": "RIGHT mark relationship (histone-tail acetylation blocks p66-alpha binding) - WRONG system: in vitro tails, non-haematopoietic cells, no fusion.",
  "Herglotz 2012": "RIGHT compartment (human haematopoietic progenitors) and real marks (H3R2me2a, H3K27me3, H3K4me2/3, H3K9ac) with Sin3a and HDAC1 in the complex - WRONG lineage (megakaryocytic) and no fusion; the MLL in it is wild-type and on the coactivator side."
}
archive checks
{
  "GEO": "NOT RUN. Recorded as not done rather than skipped. It is the cheapest thing that could weaken this bound and is the named first step of any follow-up.",
  "PRIDE_ProteomeXchange": "Not applicable to a mark readout; PXD009338 is checked in the sibling interactome absence entry.",
  "this_corpus": "The project's own deposited-data analyses (promoter matrices, GSE210480, GSE260742 work) contain no NuRD- or Sin3-subunit perturbation arm."
}
measurements
[
  {
    "quantity": "published_mark_measurements",
    "subject": "any histone mark after perturbing any NuRD or Sin3/HDAC subunit in a KMT2A-rearranged cell",
    "value": 0,
    "units": "published measurements",
    "n": 31,
    "universe": "14 canonical NuRD subunits plus 17 canonical Sin3/HDAC subunits, bounded by PubMed title/abstract on 2026-09-24 under current and retired symbols with a positive control of 55; GEO NOT searched",
    "null": null,
    "note": "rule of three over 31 subunits gives an upper bound under 9.7% on the per-subunit rate of such a measurement EXISTING IN THE LITERATURE - and that bound is on reporting, with the archive check unrun"
  }
]
How this could bite us back
  • This is the absence most likely to be closed on paper rather than at the bench: writing 'histone deacetylation' into an epigenetic_readout field on the strength of complex membership would make the corpus's epigenetics coverage figure rise without a single new measurement. That is the exact failure an existing entry's bite_back already names, and NuRD is its highest-risk instance because the activity is in the complex's name.
  • If the measurement is eventually made and shows no mark change, this graph must not quietly drop the NuRD axis - a null would be informative about where the complex is NOT acting, and would be the first real epigenetic datum in this beat.
Still open
  1. Run the GEO query. Is there any deposited ChIP-seq or CUT&RUN after CHD4, MBD3, MTA2 or SIN3A perturbation in a KMT2A-rearranged line? One query either weakens this absence or upgrades it from literature-only to archive-checked, and it was not run this cycle.
  2. Does H3K27ac at KMT2A::AFF1-bound promoters change at all when CHD4 is degraded? If the deacetylase module is not in the complex, it should not - and a null there would be a real result rather than a failed experiment.
  3. Is the Tirtakusuma lineage-switching work's CHD4 perturbation accompanied by any chromatin data in its supplementary material? That paper is the closest existing experiment in the right fusion and the right disease, and this beat read only its abstract via another entry.
Every source behind this entry (4)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Bounded literature queries run on 2026-09-24 with one field tag per term, both current and retired gene symbols, and a working positive control return no record reporting a histone-mark measurement after perturbation of any NuRD or Sin3 subunit in a KMT2A-rearranged cell.

    What this evidence cannot show

    NAMES TRIED: MTA2, MTA1L1, MBD3, GATAD2A, p66alpha, SIN3A, ARID4B, SAP180, BRCAA1, NuRD. One tag per term throughout - never the grouped form, which silently expands to All Fields - with the querytranslation recorded for every count, and run through pubmed_safe so an empty translation raises rather than returning a fabricated zero. THE SIX NuRD-WITH-MLL RECORDS WERE TITLE-CHECKED AND NONE IS A MARK MEASUREMENT AFTER A SUBUNIT PERTURBATION: they are an NT2 differentiation proteomics paper, an LSD1-in-AML review, an H3R2me2s histone-mark paper in which NuRD and MLL appear only as complex names in a SILAC interactor list, and three I did not read past the title. THE QUERY SHAPE CANNOT SEE A MARK MEASUREMENT BURIED IN A PAPER ABOUT SOMETHING ELSE - a figure panel in a lineage-switching paper would not surface here, and that is this bound's main weakness.

    Evidence that argues against this (1)
    • THE ARCHIVE CHECK THAT WOULD MOST WEAKEN THIS BOUND WAS NOT RUN. GEO was not queried for ChIP-seq or CUT&RUN after NuRD- or Sin3-subunit perturbation in any KMT2A-rearranged line. This corpus's own counting found that of 26 absence entries only 2 named a deposited accession, and that the archive check repeatedly turned weak absences into refuted ones - including a case where a PRIDE query returned a dataset in which an MLL fusion was the bait. A literature-only bound on a mark measurement is the weak kind, and this one is stated as such. ProteomeXchange/PRIDE 2018
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    The only study that perturbs a NuRD or Sin3 subunit in KMT2A::AFF1 cells - a CRISPR-Cas9 epigenome and kinome library screen identifying ARID4B, MBD3 and BMPR2 as vulnerabilities in KMT2A-rearranged ALL - measured growth and proliferation only, attributing the MBD3 effect in SEM to inhibition of cell proliferation rather than cell death, and reports no histone-mark, chromatin-accessibility or occupancy readout of any kind.

    CRISPR-Cas9 Library Screening Identifies Novel Molecular Vulnerabilities in KMT2A-Rearranged Acute Lymphoblastic Leukemia. · Schneider P, Wander P, Arentsen-Peters STCJM, Vrenken KS, Rockx-Brouwer D, Adriaanse FRS, Hoeve V, Paassen I, Drost J, Pieters R, Stam RW · International Journal of Molecular Sciences 2023
    PMID 37686014DOI
    What this evidence cannot show

    THE EXPECTED_ABSENT_TERMS CARRY THE CLAIM: this is an absence assertion about a paper's contents, and I verified it by a targeted read of the PMC full text on 2026-09-24 for cell lines, hit identities and validation method rather than by an exhaustive term sweep - so a mark measurement in a supplementary figure would not have been caught. The MCP PubMed tool's rendering of this abstract deletes italicised gene symbols (QUEEN_BRIEF, cycle 46); PubMed itself does not - the raw efetch XML fetched 2026-10-03 contains them, and MBD3[tiab] AND ARID4B[tiab] AND BMPR2[tiab] returns PMID 37686014 alone. The hit names came from the full text and the keyword list, and are present in the raw abstract too..

    Evidence that argues against this (1)
    • A screen not reporting a mark readout is not evidence that the authors did not generate one, and a follow-up from the same group would be the obvious place for it. More importantly, this item bounds ONE paper: the claim that no such measurement exists anywhere rests on E1's literature bound, which cannot see a figure panel inside a paper about something else. NCBI PubMed E-utilities 2026
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Mark-level results DO exist for these subunits outside leukaemia: p66-alpha and p66-beta bind the tails of all octamer histones in vitro and acetylation of those tails interferes with binding, with conserved region 2 of p66-alpha required for the interaction - supporting a model in which stable chromatin association follows deacetylation.

    p66alpha and p66beta of the Mi-2/NuRD complex mediate MBD2 and histone interaction. · Brackertz M, Gong Z, Leers J, Renkawitz R · Nucleic Acids Research 2006
    PMID 16415179DOI
    What this evidence cannot show

    IN VITRO binding to recombinant histone tails, in unnamed human and mouse cell lines for the localisation half; nothing leukaemic, nothing with a fusion, and no mark measured IN chromatin. The expected_absent_terms are the point of citing it here: it is the nearest thing to an epigenetic readout for any subunit in this beat, and it is three systems away from the disease.

    Evidence that argues against this (1)
    • An acetylation-sensitive interaction with recombinant histone tails is a property many basic disordered nuclear proteins display, and the direction of this one is awkward for the graph: it makes GATAD2A's chromatin binding DOWNSTREAM of the deacetylase module that this beat finds absent from the MLL-fusion interactome. Cited as the best available and not as adequate. This repository 2026
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The one haematopoietic system in which a Sin3 subunit's mark profile has been measured is a RUNX1 corepressor complex containing Sin3a and HDAC1, in which PRMT6 deposits H3R2me2a at megakaryocytic genes and holds RUNX1 targets in an intermediate state carrying H3K27me3 and H3K4me2 but not H3K4me3; on differentiation, PRMT6 binding is lost, H3R2me2a falls and a WDR5/MLL and p300/pCAF coactivator complex is recruited, raising H3K4me3 and H3K9ac.

    Histone arginine methylation keeps RUNX1 target genes in an intermediate state. · Herglotz J, Kuvardina ON, Kolodziej S, Kumar A, Hussong H, Grez M, Lausen J · Oncogene 2012
    PMID 22777353DOI
    What this evidence cannot show

    K562 and human haematopoietic progenitors undergoing MEGAKARYOCYTIC differentiation - haematopoietic but not lymphoid and not fusion-driven. THE 'MLL' IN THIS PAPER IS WILD-TYPE MLL IN A WDR5-CONTAINING COACTIVATOR COMPLEX, on the activating side of the switch; it is not a fusion protein and this item must never be cited as a Sin3 subunit interacting with an MLL fusion. Read from the abstract. It is included to show precisely how close the field gets to the missing measurement and how it still misses: right cell compartment, real marks, wrong lineage, no fusion.

    Evidence that argues against this (1)
    • This result argues that the missing measurement might come out the other way. Sin3a/HDAC1 is the REPRESSIVE arm of a switch whose activating arm contains MLL and WDR5, and the two are mutually exclusive in time. If a KMT2A fusion behaves like the coactivator arm, then a Sin3 or NuRD subunit is something it DISPLACES - in which case the mark readout after perturbing one would be small or absent, and the absence bounded here would turn out to be an absence of effect as well as of data. Herglotz J et al. 2012
Connected to