Half the KDM4C deposition is not a KMT2A fusion at all: MOZ-TIF2 8, MLL-GAS7 6, MLL-AF9 2, and the human xenograft deposited nothing
also called E-MTAB-3322 provenance · KDM4C PRMT1 system · MLL3 xenograft identity
Two entries cite a survival result in a human leukaemia sample called MLL3 whose fusion partner is never named. The paper's deposited data contains no human sample at all - it is sixteen mouse samples, and half of them carry MOZ-TIF2, which is not a KMT2A fusion, while the KMT2A arm is mostly the very rare partner GAS7.
The same thing, for a scientist
`gene-kdm4c` and `hyp-paralogue-buffering-hides-dependencies` both cite a 90-day survival result in *"a primary MLL-fusion AML sample whose partner is not specified in the text"* [E1]. Two findings follow, and the second is larger than the first. **First, the partner genuinely does not resolve.** The open-access full text says *"Primary human MLL leukemia (MLL3) was tagged with luciferase reporter and then transplanted into NSG mice"* and *"the primary AML cells carrying MLL fusion (MLL3)"* - MLL3 is the sample's label, not its partner, and the partner is named nowhere [E2]. This is one of the few caveats in the beat that survives a full-text read. **Second, the deposition is a different study from the one the graph is citing.** The paper's only accession, ArrayExpress **E-MTAB-3322**, holds **16 samples, all *Mus musculus***, all `cell type: bone marrow haematopoietic stem and progenitor cells`, transduced with a named fusion [E3]. By submitted filename the fusions are **MOZ-TIF2 (8), MLL-GAS7 (6) and MLL-AF9 (2)**, and the structured `clinical history` field splits 8 *mixed-lineage leukemia* against 8 ***non* mixed-lineage leukemia** [E3]. So the transcriptional evidence behind the KDM4C/PRMT1 axis is **half non-KMT2A**, its KMT2A half is dominated by a partner that appears in a handful of case reports, and MLL-AF9 contributes two samples out of sixteen. The human survival result the graph actually cites has **no deposited sequencing whatsoever**.
E-MTAB-3322, counted from the SDRF
| Attribute | Values |
|---|---|
| Samples | 16 |
| Organism | Mus musculus 16/16 |
| Cell type | bone marrow haematopoietic stem and progenitor cells, 16/16 |
| Fusion (submitted filename) | MOZ-TIF2 8, MLL-GAS7 6, MLL-AF9 2 |
| Clinical history | mixed-lineage leukemia 8, non mixed-lineage leukemia 8 |
| Genotype | wild type 7, KDM4C knock-down 5, PRMT1 knock-out 4 |
The study description states the design plainly: "Haematopoietic progenitors and stem cells harvested from 4-10 week old C57BL/6 mice were transduced with named fusion, immortalised, and subsequently prmt1 or kdm4c expression..." 3. Read from the SDRF on 2026-09-19.
MLL-GAS7 is the point that should change an entry. It is a t(11;17)(q23;p13) partner known from a handful of reports; it transforms multipotent progenitors and, unusually, does so without requiring Hoxa7 or Hoxa9. A KMT2A dependency mapped predominantly in MLL-GAS7 is mapped in the partner least likely to behave like the HOXA-dependent index fusion.
Two different failures in one citation
It is worth separating them because they need different fixes.
The unresolvable one. The 90-day survival result - the striking number both entries quote - is in a human primary sample whose fusion nobody has written down. GEO cannot fix this, ArrayExpress cannot fix this, and the full text does not fix it. Only the authors can. The honest entry says a primary human MLL-fusion AML xenograft of unstated partner, with no deposited sequencing.
The resolvable one, which nobody asked. The transcriptional evidence that motivates KDM4C at all is half MOZ-TIF2. That was always answerable and is arguably the more consequential fact, because it means the KDM4C/PRMT1 axis was framed as a cross-fusion mechanism from the start - chimeric transcription factors is the study's own phrase - rather than a KMT2A mechanism. An entry that imports it as KMT2A biology has narrowed a general claim and simultaneously widened its applicability to the index fusion, which are opposite errors made at once.
| script | none - BioStudies study record and SDRF |
|---|---|
| status | run |
| result summary | 16 of 16 deposited samples are murine BM HSPC. Fusions: MOZ-TIF2 8, MLL-GAS7 6, MLL-AF9 2. Half the deposition is non-KMT2A. Zero human samples. The cited human 'MLL3' xenograft has no deposited sequencing and its partner is unnamed in the full text. |
inputs
[ "PMID 26766589", "E-MTAB-3322", "PMC4712026" ]
counts — the numbers behind the chart
{
"samples_total": 16,
"samples_mouse": 16,
"samples_human": 0,
"samples_moz_tif2": 8,
"samples_mll_gas7": 6,
"samples_mll_af9": 2,
"samples_non_kmt2a": 8,
"samples_kmt2a_aff1": 0,
"xenograft_samples_deposited": 0
}entries affected
[ "gene-kdm4c", "hyp-paralogue-buffering-hides-dependencies" ]
import axes
[ "fusion: half the deposition is MOZ-TIF2, a non-KMT2A fusion; the KMT2A half is dominated by GAS7", "species: mouse not human", "lineage: myeloid not B-lymphoid" ]
- A KDM4C or PRMT1 inhibitor prioritised for infant KMT2A::AFF1 disease from this evidence is prioritised from a dataset that is half non-KMT2A and whose KMT2A half is a partner that does not need the HOXA programme the index fusion depends on.
- KDM4C and PRMT1 are broadly used chromatin enzymes; a requirement shared by MOZ-TIF2 and MLL fusions alike is more consistent with a general transformed-progenitor dependency, which predicts a narrow therapeutic window in normal haematopoiesis.
- The 90-day survival figure is the number most likely to be quoted from these entries and it comes from a single primary sample of unknown genotype, so any go/no-go decision anchored on it is anchored on an n of one with no recorded fusion.
- What fusion does the primary sample 'MLL3' carry? This is the one question in the beat that only the authors can answer, and two entries quote its survival result.
- Is the KDM4C requirement present in MLL-AF9 at all, given that MLL-AF9 contributes two of sixteen deposited samples? The axis's transcriptional support is mostly MOZ-TIF2 and MLL-GAS7.
- Does KDM4C matter in a HOXA-dependent KMT2A::AFF1 B-lineage cell, given that MLL-GAS7 transforms without Hoxa7 or Hoxa9?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The KDM4C/PRMT1 paper reports that inhibiting KDM4C and PRMT1 suppresses transformation by MLL fusions and MOZ-TIF2, and it reports a 90-day survival result in a primary human MLL-fusion AML sample labelled only 'MLL3'.
Targeting Aberrant Epigenetic Networks Mediated by PRMT1 and KDM4C in Acute Myeloid Leukemia. · Cheung N, Fung TK, Zeisig BB, Holmes K, Rane JK, Mowen KA, Finn MG, Lenhard B, Chan LC, So CW · Cancer Cell 2016What this evidence cannot show
Identifiers taken from the NCBI esummary record fetched on 2026-09-19. Full text read from Europe PMC PMC4712026 (open access) on the same date. The label 'MLL3' in this paper denotes a primary patient sample and is unrelated to the KMT2C gene, which is also called MLL3 - a collision that would mislead any automated reader and is worth flagging in any entry that quotes the result.
Evidence that argues against this (1)
- MOZ-TIF2 leukaemia has its own well-characterised chromatin mechanism, running through KAT6-dependent histone acylation and CBP recruitment rather than through H3K9 demethylation, so a shared KDM4C requirement across MOZ-TIF2 and MLL fusions is more plausibly a general transformed-progenitor requirement than a fusion-specific one. Smolko AE et al. 2024
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The partner of the primary human sample 'MLL3' is not named anywhere in the paper's full text, which describes it only as 'Primary human MLL leukemia (MLL3)' and 'the primary AML cells carrying MLL fusion (MLL3)' transplanted into NSG mice.
Targeting Aberrant Epigenetic Networks Mediated by PRMT1 and KDM4C in Acute Myeloid Leukemia. · Cheung N, Fung TK, Zeisig BB, Holmes K, Rane JK, Mowen KA, Finn MG, Lenhard B, Chan LC, So CW · Cancer Cell 2016What this evidence cannot show
Established by exhaustive search of the flattened PMC full text on 2026-09-19: the string MLL3 occurs six times and never beside a partner gene. A negative over a full text is stronger than a negative over an abstract but is still bounded by what the XML contains - supplementary tables and a patient-characteristics file are not part of the JATS body and were not retrieved, and a partner could be recorded there.
Evidence that argues against this (1)
- Rare-partner leukaemias are routinely reported with the partner named, so an unnamed partner in a primary sample more often reflects incomplete karyotyping at the source institution than deliberate omission; the answer may not exist rather than be withheld. Panagopoulos I et al. 2006
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
ArrayExpress E-MTAB-3322, the paper's only deposited dataset, contains 16 Mus musculus samples of bone-marrow haematopoietic stem and progenitor cells whose submitted filenames give the fusions as MOZ-TIF2 (8), MLL-GAS7 (6) and MLL-AF9 (2), with the structured clinical-history field splitting 8 'mixed-lineage leukemia' against 8 'non mixed-lineage leukemia', and genotypes wild type (7), KDM4C knock-down (5) and PRMT1 knock-out (4).
Targeting epigenetic networks mediated by Prmt1 and Kdm4c in acute myeloid leukemia (E-MTAB-3322) · ArrayExpress submitters · EMBL-EBI ArrayExpress / BioStudies 2016What this evidence cannot show
Read on 2026-09-19 from the BioStudies study record and its SDRF file, and counted column by column. The fusion assignment comes from the submitted FASTQ filename prefix, not from a structured genotype field - the SDRF's genotype column records the knockdown, not the fusion - so the fusion counts are a filename-derived inference. The clinical-history column corroborates it: exactly 8 samples are 'non mixed-lineage leukemia', matching the 8 MOZ-TIF2 filenames.
Evidence that argues against this (1)
- MLL-GAS7 transforms multipotent progenitors without requiring Hoxa7 or Hoxa9, which makes it the KMT2A partner least representative of the HOXA-dependent index fusion - so the KMT2A half of this deposition is not merely a different partner but a mechanistically divergent one. So CW et al. 2004
the deposition is half MOZ-TIF2 and its KMT2A half is dominated by MLL-GAS7
the 90-day survival result cited here is a primary human sample with no deposited sequencing and no named partner
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
zero deposited samples carry the index fusion
resolved source in the unstated-system worklist