Oncofusion

start / graph / ana-unstated-system-med23-mediator-resolved

Analysisestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.Q18b-unstated-systems

The Mediator/BET deposition is 48 of 48 mouse MLL-AF9/NrasG12D samples: resolving the KMT2A status behind the MED23 verdict

also called GSE74651 provenance · MED23 system · Mediator BET RN2

This graph's verdict on the Mediator subunit MED23 rests on a study whose abstract never says whether its cells were KMT2A-rearranged. All forty-eight deposited samples are mouse leukaemia cells labelled MLL-AF9/NrasG12D, so they are KMT2A-rearranged - but mouse, myeloid, and the wrong partner.

The same thing, for a scientist

`ana-med23-verdict-kmt2a-b-lineage` carries the caveat *"KMT2A rearrangement status of the cells is not stated in the abstract, and no comparison against non-KMT2A leukaemia is made"* [E1]. Both halves are answerable and both answers matter. Five GEO series are linked to PMID 27068464 - GSE78224 (4), GSE78221 (8), GSE74651 (21), GSE74650 (6) and GSE74536 (9), **48 samples in total, every one *Mus musculus*, strain C57BL/6, `cell type: Acute myeloid leukemia`, `genotype: MLL-AF9/NrasG12D`** [E2]. So the cells *are* KMT2A-rearranged, which the graph did not know; and the deposition contains **zero non-KMT2A comparator samples**, which confirms the second half of the caveat as a real absence rather than an unknown [E2]. The MED23 shRNA arms (shMed23.678, shMed23.4061) sit in the RNA-seq subseries GSE74650 and GSE78224 alongside Med12 and Med8 hairpins, with Brd4 and Med1 ChIP under JQ1 at 500 nM for 2 h [E3]. The import is triple - partner MLLT3, species mouse, lineage myeloid - in an entry whose title asserts a verdict about **B-lineage** KMT2A biology.

Go deeper
The five series, counted
SeriesnOrganismGenotypeAssay content
GSE782244Mus musculusMLL-AF9/NrasG12DRNA-seq, shRen vs shMed8
GSE782218Mus musculusMLL-AF9/NrasG12DBrd4 ChIP +/- dox, shMed12/shMed23
GSE7465121Mus musculusMLL-AF9/NrasG12DBrd4/Med1 ChIP, DMSO vs 500 nM JQ1 2 h
GSE746506Mus musculusMLL-AF9/NrasG12DRNA-seq, shMed12/shMed23 hairpins
GSE745369Mus musculusMLL-AF9/NrasG12DBrd4/Med1 ChIP, DMSO vs JQ1
Total4848/48 mouse48/48 MLL-AF9/NrasG12D

The shared protocol names the system outright: "we performed PolyA+ (illumina TruSeq) in the murine AML RN2 cell line" 2. The line is the same RN2 that carries the RNF20 result (ana-unstated-system-rnf20-h2bub-resolved), so two independent entries in this graph rest on the same mouse cell line without either saying so.

Why the second half of the caveat is the load-bearing one

Resolving "is it KMT2A-rearranged?" to yes looks like good news for the entry. It is not, because the same deposition resolves "is there a non-KMT2A comparator?" to no - 48 of 48 samples share one genotype. A dependency measured in a single genotype with no comparator cannot distinguish a KMT2A-specific requirement from a general AML or general proliferation requirement, and ana-med23-verdict-kmt2a-b-lineage reaches a verdict about B-lineage KMT2A leukaemia from it.

This is the shape the standing brief calls an absence of testing rather than a negative result: nobody looked at a non-KMT2A line, so the absence of a difference is not evidence of specificity.

Facts and figures
scriptnone - GEO accession-display text endpoint
statusrun
result summary48 of 48 deposited samples are murine MLL-AF9/NrasG12D AML (RN2). The cells ARE KMT2A-rearranged, resolving the first half of the caveat, and there is NO non-KMT2A comparator in the deposition, confirming the second half.
counts
samples_human0.000samples_human: 0samples_non_kmt2a_compara…0.000samples_non_kmt2a_comparator: 0series_count5.000series_count: 5samples_total48.0samples_total: 48samples_mouse48.0samples_mouse: 48samples_kmt2a_rearranged48.0samples_kmt2a_rearranged: 48
inputs
[
  "PMID 27068464",
  "GSE78224",
  "GSE78221",
  "GSE74651",
  "GSE74650",
  "GSE74536",
  "PMC4838499"
]
counts — the numbers behind the chart
{
  "samples_total": 48,
  "samples_mouse": 48,
  "samples_human": 0,
  "samples_kmt2a_rearranged": 48,
  "samples_non_kmt2a_comparator": 0,
  "series_count": 5
}
entries affected
[
  "ana-med23-verdict-kmt2a-b-lineage"
]
import axes
[
  "fusion: KMT2A::MLLT3",
  "species: mouse",
  "lineage: myeloid",
  "cooperating lesion: NRAS G12D"
]
How this could bite us back
  • A MED23 or Mediator-directed intervention selected from this evidence inherits a single-genotype measurement with no comparator, so a selectivity claim made from it is unsupported in the direction that matters for a therapeutic window.
  • Mediator is required for essentially all RNA polymerase II transcription, so an apparent leukaemia dependency at 48 h of knockdown may be a general essential-function effect that longer exposure would reveal as intolerable.
  • The JQ1 arm is a 2-hour acute exposure; designing a combination schedule from it assumes the chromatin displacement persists under chronic dosing, which the deposition does not test.
Still open
  1. Is Mediator release under BET inhibition a KMT2A-specific behaviour at all? The deposition contains no non-KMT2A comparator, so the question is untested rather than answered.
  2. Does MED23 knockdown do anything in a human KMT2A::AFF1 B-lineage line? Nothing in this deposition bears on B-lymphoid cells.
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The BET/Mediator paper reports that BET bromodomain inhibition releases the Mediator complex from select cis-regulatory elements, and its abstract does not state the KMT2A rearrangement status of the cells or compare against non-KMT2A leukaemia.

    BET Bromodomain Inhibition Releases the Mediator Complex from Select cis-Regulatory Elements. · Bhagwat AS, Roe JS, Mok BYL, Hohmann AF, Shi J, Vakoc CR · Cell Rep 2016
    PMID 27068464DOI
    What this evidence cannot show

    Identifiers taken from the NCBI esummary record fetched on 2026-09-19. Full text was read from NCBI efetch db=pmc for PMC4838499 after Europe PMC fullTextXML returned HTTP 500; the recovered text names RN2 22 times and MLL-AF9 twice, consistent with the deposition.

    Evidence that argues against this (1)
    • The paper's headline claim is about BET inhibition and Mediator generally; MED23 appears as one subunit in a knockdown panel, so an entry reaching a MED23-specific verdict is citing a sub-result the study was not powered around. Morgan M et al. 2022
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    All 48 samples across the five GEO series linked to PMID 27068464 are Mus musculus, strain C57BL/6, cell type 'Acute myeloid leukemia', genotype 'MLL-AF9/NrasG12D', with source name 'MLL-AF9/NrasG12D murine cell line'; no human and no non-KMT2A comparator sample is deposited.

    What this evidence cannot show

    Read on 2026-09-19 from the GEO accession-display text endpoint and counted programmatically across GSE78224, GSE78221, GSE74651, GSE74650 and GSE74536. Submitter-entered fields. Series were located by elink dbfrom=pubmed&db=gds and cross-checked against the accession list in the paper's data-availability statement; the two agree exactly for this paper, which is not true of every paper in this beat.

    Evidence that argues against this (1)
    • A deposition's silence about comparators is not the study's silence: viability or colony assays in non-KMT2A lines leave no sequencing trace, so the no-comparator finding is bounded to deposited sequencing and could understate the paper's controls. Rogic S et al. 2026
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    The MED23 arm of the deposition consists of two independent shRNAs (shMed23.678, shMed23.4061) profiled by RNA-seq in GSE74650 and GSE78224, alongside Med12 and Med8 hairpins, with Brd4 and Med1 ChIP-seq under DMSO versus 500 nM JQ1 for 2 hours.

    What this evidence cannot show

    Sample titles and treatment characteristics read on 2026-09-19. Two hairpins per subunit is the minimum for an off-target control and there is no rescue arm in the deposition. The JQ1 exposure (500 nM, 2 h) is an acute chromatin-displacement condition, not a growth-inhibition condition, so transcriptional effects here are immediate-early and not a viability phenotype.

    Evidence that argues against this (1)
    • This graph has separately established that MED23 is unreachable from subunit-symbol PubMed queries because the literature frames this screen at the level of the Mediator complex - so the same paper that fixes the system here is also the paper a symbol-keyed enumeration would have missed entirely. Bhagwat AS et al. 2016
Connected to