Oncofusion

start / graph / ana-kmt2a-r-line-panel-monocytic-audit

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q-cells-saturation

The monocytic confound is worse than recorded: two of the six rearranged lines are one patient, and the four-line control group contains a HOXA-driven fusion and two NPM1c/myelomonocytic lines

also called monocytic line panel audit · Cellosaurus audit of the DepMap monocytic split

This project's iron-and-haem finding depends on comparing leukaemia cell lines that carry the fusion against ones that do not, and the earlier check said the comparison simply had too few cells lines to settle. Looking up every line's record one by one shows the problem is not only the count: two of the fusion lines came from the same patient, and one of the four comparison lines carries a different fusion that drives the same genes.

The same thing, for a scientist

A line-by-line Cellosaurus audit (2026-09-19) of the panel used by `ana-depmap-haem-lineage-control`. Three findings. (1) **MOLM-13 and MOLM-14 are the same patient** — Cellosaurus records an OI cross-reference CVCL_2119 <-> CVCL_7916, and the founding report establishes both from the peripheral blood of one patient at relapse of FAB M5a acute monocytic leukaemia [E1][E2]. The monocytic arm of the rearranged panel therefore represents five patients, not six. (2) **The four-line monocytic comparison group is not lineage-matched and is not fusion-clean.** U-937 (CVCL_0007) carries **PICALM-MLLT10 (CALM-AF10)**, whose transformation is DOT1L-dependent and which drives HOXA5/HOXA9/HOXA10 and MEIS1 to levels indistinguishable from MLL-translocated leukaemia [E3][E4]; OCI-AML-3 (CVCL_1844) carries NPM1 p.Trp288Cysfs*12 with DNMT3A p.Arg882Cys, a genotype that is itself menin-inhibitor sensitive [E5][E6]; and OCI-AML-3 and MUTZ-8 are annotated **myelomonocytic (FAB M4)**, not monocytic (M5) [E5]. (3) **Mono-Mac-1 shares an individual with Mono-Mac-6** (OI CVCL_1425 <-> CVCL_1426), which does not reduce n here but constrains any future expansion of the group [E5]. Net: the group labelled 'four monocytic non-rearranged leukaemia lines' contains **one** line that is monocytic, non-KMT2A-rearranged, and free of a HOXA-axis fusion or a menin-dependent genotype. The earlier verdict — inconclusive — stands, but the reason should be recorded as contamination of the control arm, not only its size.

Go deeper
What was audited and how3 deeper

analyses/depmap-haem-lineage-control/run.py excluded the 11 KMT2A-rearranged leukaemia lines and split the remaining non-rearranged leukaemia lines by monocytic identity, yielding 4 monocytic (OCI-AML-3, U-937, Mono-Mac-1, MUTZ-8) versus 35 other and 4 unresolved. No haem gene separated the groups (p 0.19-0.88) and the GAPDH control itself moved (p = 0.031), so the run was recorded as inconclusive.

This audit fetched the full Cellosaurus text record for every line on both sides of that split on 2026-09-19 via https://api.cellosaurus.org/cell-line/<accession>?format=txt, and read the DI (disease), OI (originate from same individual), CC Sequence variation and CC Problematic cell line fields rather than relying on a lineage label 5. Identifiers were resolved by accession, not by name.

The rearranged monocytic arm, line by line
lineaccessionfusion (Cellosaurus)disease termnote
MV4-11CVCL_0064KMT2A-AFF1childhood acute monocytic leukaemia; acute monoblastic/monocytic leukaemiamale, 10y
THP-1CVCL_0006KMT2A-MLLT3childhood acute monocytic leukaemiamale, 1y
NOMO-1CVCL_1609KMT2A-MLLT3adult acute monocytic leukaemiafemale, 31y
SHI-1CVCL_2191KMT2A-AFDNadult acute monocytic leukaemiamale, 37y
MOLM-13CVCL_2119KMT2A-MLLT3adult acute monocytic leukaemiamale, 20y; OI CVCL_7916
MOLM-14CVCL_7916KMT2A-MLLT3adult acute myeloid leukaemia (no monocytic term)male, 20y; same patient as MOLM-13

Two things follow. First, MOLM-14's disease annotation in Cellosaurus (NCIt C9154, ORDO Orphanet_519) carries no monocytic term at all, so the '6 of 7 monocytic' count cannot be reproduced from the database alone — it requires the founding paper, which does establish FAB M5a for both lines and monocyte-specific esterase expression in both 2. Second, and more consequentially, MOLM-13 and MOLM-14 are not independent: same patient, same peripheral blood at relapse, differing only in subclonal antigen expression (MOLM-13 CD34+CD13-CD14-; MOLM-14 CD4+CD13+CD14+) 2.

One further observation worth its own line: MV4-11 carries KMT2A-AFF1 — the archetypal infant B-ALL fusion — and is nonetheless annotated as acute monocytic leukaemia. The same fusion in this panel produces both the B-lymphoid and the monocytic phenotype, which is a small piece of direct support for the lineage-plasticity picture in mechanism-lineage-switch-kmt2a.

The control arm is contaminated in three separate ways

U-937 is a HOXA-axis fusion line. Cellosaurus records PICALM-MLLT10 (CALM-AF10) in U-937 5. MLLT10/AF10 is a canonical KMT2A fusion partner in this graph, and CALM-AF10 is not merely adjacent biology: targeted disruption of Dot1l abolishes CALM-AF10 transformation in vitro and CALM-AF10 leukaemia initiation and maintenance in vivo, exactly as it does for MLL-AF10, and the DOT1L inhibitor EPZ004777 suppresses Hoxa cluster genes and Meis1 in both 3. In patients, CALM-AF10-positive T-ALL over-expresses HOXA5, HOXA9, HOXA10 and MEIS1 to levels similar to MLL-translocated acute leukaemia, and the authors propose grouping MLL-t, CALM-AF10 and HOXA-t leukaemias together 4. Placing U-937 in a 'non-KMT2A-rearranged' comparison arm therefore places a positive in the control group for anything that runs through the HOXA/MEIS1/DOT1L axis.

OCI-AML-3 is menin-dependent by genotype. It carries NPM1 p.Trp288Cysfs*12 (NPM1c) with DNMT3A p.Arg882Cys 5. NPM1c/DNMT3A-mutant AML is explicitly among the genotypes sensitive to menin-KMT2A inhibition in a large-scale primary-AML drug screen 6. So the control arm also contains a line that is menin-dependent for reasons other than a KMT2A rearrangement.

Two of the four are not monocytic in the same sense. OCI-AML-3 and MUTZ-8 are annotated acute myelomonocytic leukaemia (NCIt C7962, ORDO Orphanet_517, FAB M4), whereas U-937 and Mono-Mac-1 are acute monocytic/monoblastic (NCIt C8263, ORDO Orphanet_514, FAB M5) 5. A haem-handling phenotype that tracks the degree of monocytic commitment would be diluted by mixing M4 and M5.

Removing the HOXA-axis line and the two myelomonocytic lines leaves Mono-Mac-1 alone as a monocytic, non-rearranged, non-HOXA-axis, non-NPM1c comparator. No permutation test on n=1 can settle anything, which is a more specific and more actionable diagnosis than 'too few lines'.

What this changes, and what it does not

It does not rescue the haem finding. The monocytic explanation remains live and is not excluded by anything here.

It does change three things:

  1. The power calculation. The rearranged monocytic arm is 5 patients, not 6, and the usable control arm is 1 line, not 4. Any future attempt to run this comparison inside DepMap will fail for the same reason unless the panel is extended outside DepMap.
  2. The design of the replacement experiment. The right control is not 'a monocytic AML line without a KMT2A rearrangement' — that set is almost empty and contaminated. It is an isogenic one: the same monocytic KMT2A-rearranged line with the fusion degraded, with haem-gene dependency read out before lineage identity has had time to drift. See exp-monocytic-isogenic-haem-dependency.
  3. A standing rule for this graph's DepMap work. A 'non-KMT2A-rearranged' control group should be screened for pathway-equivalent lesions — PICALM-MLLT10, NUP98 fusions, NPM1c, and other HOXA/MEIS1 drivers — not only for KMT2A rearrangement. That screen was not applied here and would have caught two of the four lines.
Correction: the one clean comparator shares a patient with a KMT2A::MLLT3 line

This audit already recorded that Mono-Mac-1 (CVCL_1425) shares an individual with Mono-Mac-6 (CVCL_1426) and judged that this "does not reduce n here but constrains any future expansion of the group". That judgement missed the consequence.

Mono-Mac-6 is annotated as carrying KMT2A-MLLT3 / MLL-AF9, cited to PMID 7833491 ("Molecular analysis of the Mono Mac 6 cell line: detection of an MLL-AF9 fusion transcript", Blood 1995). Verified by fetching both records by accession on 2026-09-20: CVCL_1425 lists same-origin-as: CVCL_1426, and CVCL_1426's sequence-variation list carries the KMT2A-MLLT3 gene fusion.

So the sole surviving non-rearranged monocytic comparator comes from a patient whose leukaemia is KMT2A-rearranged. Mono-Mac-1's fusion-negative status is an absence of annotation, not a negative assay - Cellosaurus records no fusion for it, which is not the same as a fusion having been looked for and not found.

The effective clean-control count is therefore 1 or 0, and which it is has not been established. Two lines from one patient can genuinely differ - they were derived at different times and subclones diverge - so this is not a demonstration that Mono-Mac-1 carries the fusion. It is a demonstration that nobody has checked, in the one line the whole control arm reduces to.

The check is cheap and the material is in hand: Mono-Mac-1 is in DepMap (ACH-001129), so an RNA-seq fusion call settles it without touching a bench.

Sharpened: the two lines share a TP53 mutation, and the fusion assay was never run

The previous correction said Mono-Mac-1's fusion-negative status is absence of annotation, not a negative assay, and left open whether the two autologous lines genuinely differ. Fetching both full records on 2026-09-20 narrows it in both directions.

They share a driver mutation. Mono-Mac-1's sequence-variation list contains exactly one entry: TP53 p.Arg273His (c.818G>A). Mono-Mac-6 carries the same variant, alongside KMT2A-MLLT3 and RUNX1-ATP8A2 fusions. A shared clonal TP53 mutation is strong corroboration that these are the same patient's leukaemia rather than two lines that merely share a donor.

And the assay that would find a fusion was never run on Mono-Mac-1. Its omics inventory lists whole-exome sequencing, shRNA screening, quantitative proteomics and transcriptomics: microarray. A microarray cannot call a fusion transcript. So the missing annotation is not evidence of absence at all - it is the footprint of an assay that could not have detected one, on a line whose sibling from the same patient carries the fusion.

The honest position on the control arm is therefore worse than 'one clean comparator'. The one line the arm reduces to comes from a KMT2A::MLLT3 patient, shares that patient's TP53 mutation, and has never been tested for the rearrangement. It should be treated as unknown-status, not fusion-negative, until someone calls fusions from its RNA.

It is in DepMap as ACH-001129, so the call is a download and a query rather than an experiment.

Facts and figures
scriptmanual audit; no new script. Source panel from analyses/depmap-haem-lineage-control/run.py and results.json
statusrun
run date2026-09-19
result summaryRearranged monocytic arm = 5 independent patients, not 6 (MOLM-13/MOLM-14 same individual). Control arm of 4 reduces to 1 clean comparator (Mono-Mac-1) after removing U-937 (PICALM-MLLT10, DOT1L/HOXA axis), OCI-AML-3 (NPM1c + DNMT3A R882C, menin-sensitive genotype; FAB M4) and MUTZ-8 (FAB M4, JAK2 V617F). The earlier 'inconclusive' verdict is unchanged; its reason is revised from insufficient n to a contaminated control arm.
standing rule proposedA 'non-KMT2A-rearranged' control group in this graph's DepMap work must additionally be screened for pathway-equivalent lesions: PICALM-MLLT10, NUP98 fusions, NPM1c, and other HOXA/MEIS1 drivers.
inputs
[
  "https://api.cellosaurus.org/cell-line/<accession>?format=txt for CVCL_0064, CVCL_0006, CVCL_1609, CVCL_2191, CVCL_2119, CVCL_7916, CVCL_0007, CVCL_1844, CVCL_2130, CVCL_1425",
  "analyses/depmap-haem-lineage-control/results.json"
]
haem genes under test
[
  "ALAD (rank 14 of 17728)",
  "IREB2 (15)",
  "UROS (41)",
  "ALAS1 (47)"
]
rearranged monocytic lines
[
  {
    "line": "MV4-11",
    "accession": "CVCL_0064",
    "fusion": "KMT2A-AFF1",
    "fab": "M5 (monocytic/monoblastic)"
  },
  {
    "line": "THP-1",
    "accession": "CVCL_0006",
    "fusion": "KMT2A-MLLT3",
    "fab": "M5 (monocytic/monoblastic)"
  },
  {
    "line": "NOMO-1",
    "accession": "CVCL_1609",
    "fusion": "KMT2A-MLLT3",
    "fab": "M5 (monocytic/monoblastic)"
  },
  {
    "line": "SHI-1",
    "accession": "CVCL_2191",
    "fusion": "KMT2A-AFDN",
    "fab": "M5 (monocytic/monoblastic)"
  },
  {
    "line": "MOLM-13",
    "accession": "CVCL_2119",
    "fusion": "KMT2A-MLLT3",
    "fab": "M5a",
    "same_individual_as": "CVCL_7916"
  },
  {
    "line": "MOLM-14",
    "accession": "CVCL_7916",
    "fusion": "KMT2A-MLLT3",
    "fab": "M5a (per founding paper; Cellosaurus DI lacks a monocytic term)",
    "same_individual_as": "CVCL_2119"
  }
]
control arm disqualifications
[
  {
    "line": "U-937",
    "accession": "CVCL_0007",
    "reason": "carries PICALM-MLLT10 (CALM-AF10), a DOT1L-dependent HOXA/MEIS1 driver of the same axis as KMT2A fusions"
  },
  {
    "line": "OCI-AML-3",
    "accession": "CVCL_1844",
    "reason": "NPM1 p.Trp288Cysfs*12 + DNMT3A p.Arg882Cys, a menin-inhibitor-sensitive genotype; also FAB M4 myelomonocytic rather than M5"
  },
  {
    "line": "MUTZ-8",
    "accession": "CVCL_2130",
    "reason": "FAB M4 myelomonocytic rather than M5 monocytic; JAK2 p.Val617Phe"
  },
  {
    "line": "Mono-Mac-1",
    "accession": "CVCL_1425",
    "reason": "retained as the only clean comparator; note OI with Mono-Mac-6 (CVCL_1426) constrains future expansion of this group"
  }
]
effective n
{
  "rearranged_monocytic_patients": 5,
  "clean_control_lines": 1
}
How this could bite us back
  • If the four haem genes turn out to be a monocytic-lineage dependency after all, any haem-pathway inhibitor developed from this finding would hit normal monocytes and macrophages at the same dose, and the therapeutic index would be near zero.
  • The standing rule proposed here would shrink the control arm of several existing comparisons in this graph, so applying it will reduce, not increase, the apparent strength of previously reported selective dependencies.
  • Correcting the panel to 5 independent patients lowers the power of every future analysis that uses it, which makes an isogenic degron design the only affordable route to an answer.
Still open
  1. Call fusions from Mono-Mac-1's DepMap RNA-seq (ACH-001129). Cellosaurus records only a microarray for this line, so the fusion has never been looked for with an assay that could find it, and the whole monocytic-lineage control rests on the answer.
  2. Are there monocytic, non-KMT2A-rearranged, non-HOXA-axis, non-NPM1c leukaemia lines outside the DepMap panel that could be CRISPR-screened for the four haem genes, or is the required control group empty in every public panel?
  3. Does re-running the original split with U-937, OCI-AML-3 and MUTZ-8 reassigned change the GAPDH control offset of p = 0.031, or is that offset driven by something other than lineage composition?
  4. How many other comparisons in this graph's DepMap work used a control group that was screened only for KMT2A rearrangement and not for pathway-equivalent lesions?
  5. Does the KMT2A-AFF1-positive but monocytic phenotype of MV4-11 represent a lineage switch in the original patient, or a fusion-independent monocytic drift in culture? The Cellosaurus record does not say, and the answer bears on how MV4-11 is used as a B-ALL model.
Every source behind this entry (8)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Cellosaurus records MOLM-13 (CVCL_2119) and MOLM-14 (CVCL_7916) as originating from the same individual, via a reciprocal OI cross-reference.

    What this evidence cannot show

    The OI field asserts shared individual of origin but not that the two lines behave equivalently; MOLM-13 and MOLM-14 differ in subclonal antigen expression and in DepMap identifier (ACH-000165-series versus ACH-001574). Treating them as one observation is conservative for independence but discards genuine subclonal variation.

    Evidence that argues against this (1)
    • Same-patient origin does not make two lines redundant as observations. The founding report documents interclonal phenotypic heterogeneity between them — MOLM-13 is CD34+CD13-CD14- while MOLM-14 is CD4+CD13+CD14+ — so they sample different subclones of the relapse. For a dependency that varies with monocytic maturation, the pair may carry more information than one line, and collapsing them to n=1 patient could under-count rather than correct. Read: abstract only. Matsuo Y et al. 1997
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    MOLM-13 and MOLM-14 were both established from the peripheral blood of a single patient at relapse of acute monocytic leukaemia, FAB M5a, which had evolved from myelodysplastic syndrome; both express monocyte-specific esterase and MLL-AF9 fusion mRNA arising from an occult ins(11;9)(q23;p22p23).

    PMID 9305600DOI
    What this evidence cannot show

    A 1997 cell-line establishment report; the monocytic assignment rests on FAB morphology plus monocyte-specific esterase, not on modern immunophenotyping or transcriptomics. Read: abstract only. Note also that both lines derive from AML that evolved from MDS, so they carry an antecedent-MDS genetic background (trisomy 8 present in the MDS phase) that most KMT2A-rearranged paediatric disease does not.

    Evidence that argues against this (1)
    • The database that the rest of this graph's lineage annotation relies on does not reproduce this. Cellosaurus assigns MOLM-14 only NCIt C9154 'Adult acute myeloid leukemia' and ORDO Orphanet_519, with no monocytic term, while giving MOLM-13 the monocytic terms. Anyone recomputing 'how many rearranged lines are monocytic' from Cellosaurus disease fields alone will get 5 of 7, not 6 of 7. The discrepancy is an annotation gap rather than a biological dispute, but it means the count in this graph must cite the founding paper, not the database. Bairoch A; SIB Swiss Institute of Bioinformatics 2026
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MLL-AF10 and CALM-AF10 mediated transformation are both DOT1L-dependent: conditional Dot1l knockout abolishes in vitro transformation of murine bone marrow cells and in vivo initiation and maintenance of either leukaemia, and the DOT1L inhibitor EPZ004777 suppresses Hoxa cluster genes and Meis1 and selectively impairs proliferation of both MLL-AF10 and CALM-AF10 transformed cells.

    PMID 23138183DOI
    What this evidence cannot show

    Mouse bone-marrow transformation and transplantation models with a conditional Dot1l allele; it establishes shared dependence on DOT1L, not that CALM-AF10 and KMT2A fusions share every downstream dependency, and specifically not that they share a haem-pathway dependency. Read: abstract only.

    Evidence that argues against this (1)
    • Shared DOT1L dependence in mouse models did not translate into shared clinical behaviour. In the adult phase 1 of the DOT1L inhibitor pinometostat (n=51), H3K79 methylation was reduced but efficacy as a single agent was modest, with complete remission in only 2 patients, both t(11;19). A pathway that is genetically required in a mouse model can be pharmacologically unrewarding in patients, so 'U-937 shares the DOT1L axis' is a reason to doubt it as a control, not a demonstration that it would behave like a KMT2A-rearranged line in a dependency screen. Read: abstract only. Stein EM et al. 2018
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CALM-AF10-positive T-ALL over-expresses HOXA5, HOXA9, HOXA10 and the cofactor MEIS1 at levels similar to those observed in MLL-translocated acute leukaemia, leading the authors to propose a combined HOXA-positive leukaemia group comprising at least MLL-translocated, CALM-AF10 and HOXA-translocated acute leukaemias.

    CALM-AF10+ T-ALL expression profiles are characterized by overexpression of HOXA and BMI1 oncogenes · Dik WA, Brahim W, Braun C, Asnafi V, Dastugue N, Bernard OA, et al., Delabesse E · Leukemia 2005
    PMID 16107895DOI
    What this evidence cannot show

    Microarray comparison in T-ALL, not in monocytic AML; U-937 is a monocytic line, so the transcriptional equivalence demonstrated in T-lineage disease is being extrapolated across lineage here. Cohort sizes are small by modern standards and validation was by qRT-PCR on an independent T-ALL group. Read: abstract only.

    Evidence that argues against this (1)
    • The same study identifies a difference that separates the two groups: BMI1, which lies close to the AF10 breakpoint, was over-expressed only in CALM-AF10-positive T-ALL and NOT in MLL-translocated acute leukaemia, and the CDKN2A locus was conserved in CALM-AF10 cases rather than deleted. CALM-AF10 is therefore a partially overlapping but distinct program, so calling U-937 a positive in the control group overstates the equivalence; it is a partial positive. Read: abstract only (same paper). Dik WA et al. 2005
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Cellosaurus records for the four comparison lines show: U-937 (CVCL_0007) carries PICALM-MLLT10 and is annotated acute monocytic/monoblastic leukaemia with a documented historical partial cross-contamination by K-562; OCI-AML-3 (CVCL_1844) carries NPM1 p.Trp288Cysfs*12 and DNMT3A p.Arg882Cys and is annotated adult acute myelomonocytic leukaemia; MUTZ-8 (CVCL_2130) carries JAK2 p.Val617Phe and is annotated adult acute myelomonocytic leukaemia; Mono-Mac-1 (CVCL_1425) is annotated adult acute monocytic leukaemia and originates from the same individual as Mono-Mac-6 (CVCL_1426).

    What this evidence cannot show

    Cellosaurus aggregates depositor and literature annotation; the fusion and mutation records carry their own source PMIDs which this audit did not individually re-read. Disease terms reflect the establishing report's diagnosis, which for lines established decades ago used FAB criteria that the WHO 5th edition has since demoted.

    Evidence that argues against this (1)
    • The audit's own framing can be pushed too far. Cellosaurus notes that the U-937 contamination problem 'was corrected and subsequent stocks have been confirmed to carry only U-937', so the contamination note is historical and does not invalidate current DepMap data for that line; and an FAB M4 versus M5 distinction is a morphological gradation, not two different cell types, so excluding OCI-AML-3 and MUTZ-8 as 'not monocytic' is a defensible tightening rather than a demonstrated error in the original analysis. The original verdict of inconclusive is not overturned by this audit; it is re-explained. Bairoch A; SIB Swiss Institute of Bioinformatics 2026
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A large-scale drug screen comparing genetic alterations and quantitative proteomics with drug sensitivity in primary AML patient cells found that GMP-like AML display the greatest sensitivity to the menin-KMT2A inhibitor bleximenib (JNJ-75276617), and that NPM1c/DNMT3A-mutant AML are also sensitive.

    Bleximenib, the novel menin-KMT2A inhibitor JNJ-75276617, impairs long-term proliferation and immune evasion in acute myeloid leukemia · Hogeling SM, Le DM, La Rose N, Kwon MC, Wierenga ATJ, Van den Heuvel FAJ, et al., Schuringa JJ · Haematologica 2025
    PMID 39704147DOI
    What this evidence cannot show

    Ex vivo drug sensitivity in primary AML cells, not clinical response; the GMP-like and NPM1c/DNMT3A classes are defined by the study's own proteomic clustering and mutation calls. Cited here for the fact that OCI-AML-3's genotype is a menin-sensitive one, not for any claim about the haem pathway. Read: abstract only.

    Evidence that argues against this (1)
    • Menin sensitivity of a genotype does not imply that the line shares the dependency being measured. The haem-pathway result under audit is a CRISPR dependency in a growth screen, not a menin-inhibitor sensitivity, and no study has shown that NPM1c AML shares the haem-biosynthesis requirement reported in KMT2A-rearranged lines. Excluding OCI-AML-3 from the control group on menin-sensitivity grounds is therefore a precaution against a shared upstream axis, not a demonstrated confound for this specific readout. Read: abstract only. Zhou X et al. 2024
  7. E-monomacstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Mono-Mac-1 (CVCL_1425), the only line surviving this audit as a clean monocytic non-rearranged comparator, originates from the same individual as Mono-Mac-6 (CVCL_1426), which Cellosaurus annotates as carrying the KMT2A-MLLT3 (MLL-AF9) gene fusion. Mono-Mac-1's fusion-negative status rests on absence of annotation rather than on a reported assay.

    What this evidence cannot show

    Both records fetched by accession 2026-09-20. **Autologous lines are not the same cells**: Mono-Mac-1 and Mono-Mac-6 were established separately and a subclone without the rearrangement is entirely possible, so this does not show Mono-Mac-1 is rearranged. The fusion annotation on Mono-Mac-6 is Cellosaurus curation citing PMID 7833491, whose full text was not retrieved here; the title states the finding directly.

    Evidence that argues against this (2)
    • A patient with a KMT2A::MLLT3 leukaemia will usually have the rearrangement in the malignant clone of every line derived from that leukaemia, which makes the contamination reading more likely than the clean one - but 'more likely' is the wrong standard for a control arm, and the DepMap RNA-seq call is available. Blood 1995
    • This does not change the audit's conclusion, which was already that the analysis is inconclusive. It changes the reason once more: from 'too few lines' to 'one clean comparator' to **'the one clean comparator may not be clean'**. Each revision has made the control arm weaker, which is a pattern worth noticing in an arm that was originally reported as four lines. Bairoch A; SIB Swiss Institute of Bioinformatics 2026
  8. E-tp53-sharedstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Mono-Mac-1 (CVCL_1425) carries exactly one recorded sequence variation, TP53 p.Arg273His, which is the same variant recorded for its autologous partner Mono-Mac-6 (CVCL_1426) alongside the KMT2A-MLLT3 fusion. Mono-Mac-1's only transcriptomic assay on record is microarray, which cannot detect a fusion transcript.

    What this evidence cannot show

    Both records fetched by accession 2026-09-20. **Cellosaurus's omics inventory records what has been reported to it, not everything ever done to a line** - Mono-Mac-1 may have been RNA-sequenced in a study Cellosaurus does not index, and it is in DepMap, whose data were not queried here. TP53 R273H is one of the commonest hotspot mutations in cancer, so sharing it is corroboration of common origin and not proof; the same-origin relation is separately asserted by both records.

    Evidence that argues against this (1)
    • None of this shows Mono-Mac-1 carries the fusion. Autologous lines established at different times can genuinely diverge, and a subclone without the rearrangement is possible. **The correct status is unknown, which is the point** - a control arm reduced to one line of unknown status supports no conclusion either way, whereas 'fusion-negative' supported one. Bairoch A; SIB Swiss Institute of Bioinformatics 2026
Connected to