Oncofusion

start / graph / hyp-cell-type-is-set-before-the-fusion-and-the-fusion-does-not-move-it

Hypothesisspeculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.Q60-cellstate2

Hypothesis (first clause REFUTED 2026-09-21): the fusion's core dependencies may be cell-type-invariant, but the cell type itself is not fixed

also called cell-type-invariant core dependency

The idea is that what kind of cell the leukaemia came from is decided first, and the fusion then runs the same core machinery regardless. If true, the central drug targets would work the same way in myeloid and B-cell versions of the disease.

The same thing, for a scientist

**FIRST CLAUSE WITHDRAWN 2026-09-21 (Q68).** This hypothesis was stated as two claims joined by an 'and': (i) the cell type is fixed before the fusion arrives, and (ii) the fusion's core dependency set does not change with cell type. **Claim (i) is contradicted by the published lineage-switch literature and is withdrawn.** A 2024 review of 104 published lineage-switch cases (PMID 38775354) reviewed 104 published lineage-switch cases: 49 (46.2%) carried a KMT2A rearrangement, 75.5% switched from a lymphoid to a myeloid phenotype, most commonly B-ALL to AML (59.4%), and 37.7% had received lineage-specific immunotherapy, which the authors propose as a trigger acting on leukaemic stem cells capable of either differentiation programme. Lineage switch is CONCENTRATED in the fusion class this graph is about. A cell type that can be moved by an antibody is not fixed before the fusion arrives. **Claim (ii) survives and is what this entry is now for.** This beat's DepMap contrast holds the fusion constant at KMT2A::AFF1 and varies cell type (MV4-11 against SEM, SEMK2 and RS4;11). MEN1 moves -0.168 and DOT1L -0.143 across that contrast, against KMT2A-rearranged-versus-all-lines differences of -0.901 and -0.457 respectively [E1]. The residual hypothesis is that **the fusion's core dependency set is cell-type-invariant**, and the cell-type-dependent part of the phenotype (surface immunophenotype, lineage-switch DESTINATION, niche response) is carried by genes outside that core. The rival remains that the contrast is n = 1 against n = 3 - and MV4-11's founding paper calls it biphenotypic B-myelomonocytic (PMID 3500218), so the n = 1 is also not a clean myeloid arm. **And the withdrawal makes claim (ii) MORE testable, not less.** If cell type moves under therapy while the fusion stays, then a lineage switch is a natural experiment in exactly this design: the same fusion, the same patient, two cell types, sequential in time.

Go deeper
What would distinguish the two

The contrast that would settle it is not more cell lines. It is an isogenic pair: the same fusion knocked into two starting cell types from the same donor, with the dependency measured in both. The graph already contains exp-monocytic-isogenic-haem-dependency and exp-menin-inhibitor-lineage-destination, which are neighbours of this design.

The cheaper intermediate is to add every KMT2A-rearranged line DepMap has, split by BOTH fusion partner and lineage, and ask whether MEN1's flatness survives a larger panel.

Why it matters if it is true

Menin inhibitors are being given across KMT2A-rearranged AML and ALL on the assumption that the mechanism is the same. If the core dependency set is cell-type-invariant, that assumption is licensed and the lineage-switch escape route becomes the main worry. If it is not, then response rates should differ by lineage in a way the trials can already be interrogated for.

Correction, 2026-09-21 (Q68): the hypothesis's first clause is already refuted

What was withdrawn. The title clause 'the cell type is fixed before the fusion arrives' and the framing that made cell type a variable to be held constant.

Why. PMID 38775354. Of 104 published lineage-switch cases, 46.2% carried a KMT2A rearrangement and 59.4% of switches were B-ALL to AML; 37.7% had received lineage-specific immunotherapy. The entry's own first open question already said 'the lineage-switch literature argues the cell type is not fixed at all under therapy' - so the graph knew, and the title asserted the opposite anyway. That is the 'entry arguing with itself' pattern the brief names, and it had been sitting in the open-questions array rather than the claim.

What is unchanged. Confidence remains speculative. Evidence items E1 and E2 and their claims stand. The dependency-invariance half is untouched by the lineage-switch data, because a switch that changes immunophenotype says nothing directly about whether MEN1 and DOT1L requirements move with it - and measuring that in paired pre- and post-switch material is now the cleanest available test of the surviving claim.

Facts and figures
statementFor a given KMT2A fusion, the core dependency set (MEN1, DOT1L, PSIP1, KMT2A itself) is invariant with respect to the cell type the fusion arose in; cell-type-dependent phenotype is carried by genes outside that core.
origindata-pattern
prior probability0.45
predictions
[
  "Adding further KMT2A::AFF1 lines of either lineage will not increase the MEN1 myeloid-minus-B-lineage gap beyond noise.",
  "Menin-inhibitor response rate in KMT2A-rearranged disease will not differ materially by lineage once partner and prior therapy are accounted for.",
  "An isogenic fusion knock-in into two starting cell types will show the same MEN1 and DOT1L dependency and different surface-marker and niche-response genes."
]
experiments
[
  "exp-same-fusion-two-cell-types-allele-selective-degron"
]
analyses
[
  "ana-q60-cell-type-versus-fusion"
]
falsifiers
[
  "A larger KMT2A::AFF1 panel in which the myeloid-minus-B-lineage MEN1 gap exceeds the within-lineage spread.",
  "An isogenic knock-in pair in which DOT1L or MEN1 dependency differs by more than the assay's replicate variance.",
  "A menin-inhibitor trial readout showing a lineage-associated response difference that survives adjustment for partner and line of therapy."
]
How this could bite us back
  • If the core set is cell-type-invariant, then so is the toxicity: menin inhibition would hit normal cells of both lineages equally and there is no lineage-based sparing to exploit.
  • An invariant core is also an invariant resistance surface - the same MEN1 mutations would be expected to arise in AML and ALL, removing any hope that one lineage is harder for the tumour to escape in.
  • The hypothesis is built on a single myeloid line and would be embarrassing to act on before the isogenic experiment is done.
Still open
  1. In a documented KMT2A-rearranged lineage switch, do the fusion's core dependencies (MEN1, DOT1L) move with the immunophenotype? Paired pre- and post-switch material is the natural experiment for the surviving half of this hypothesis and nobody has run ex vivo menin-inhibitor sensitivity on such a pair.
  2. Does the invariance hold for the partner as well as for the cell type, or is KMT2A::AFDN different from KMT2A::AFF1 in ways the core set can see?
  3. If lineage-specific immunotherapy triggers switching in 37.7% of reported cases, is the switch destination itself predictable from the fusion partner? The 104-case series does not stratify that way.
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    With the fusion held constant at KMT2A::AFF1, the myeloid-minus-B-lineage gene-effect difference is -0.168 for MEN1 and -0.143 for DOT1L, against KMT2A-rearranged-versus-all-lines differences of -0.901 and -0.457.

    What this evidence cannot show

    n = 1 myeloid line (MV4-11) against n = 3 B-lineage lines. This is a look, not a test, and it is the principal weakness of the hypothesis as stated. Cancer lines in culture.

    Evidence that argues against this (1)
    • KAT8/MOF is a direct counterexample to cell-type invariance in the surrounding biology: its acetyltransferase activity is required for adult but not for early fetal haematopoiesis, so at least one chromatin requirement in this system is decided by developmental context. The hypothesis is confined to the fusion's CORE dependency set and would be false if stated more broadly. Valerio DG et al. 2017
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    The repository's earlier lineage split already reported MEN1 as lineage-independent (-2.29 myeloid, -2.28 B-lineage in comparator standard deviations) while CEBPA and SPI1 were myeloid-identity artefacts, establishing that the invariant-versus-lineage-specific distinction is measurable in this panel.

    What this evidence cannot show

    That analysis is expressed in standard deviations of the comparator arm rather than in raw gene effect, and its panel omitted MV4-11, so lineage and fusion partner were confounded in it.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT against the premise that a cell-line lineage label is a fixed variable to hold constant. Cell type in KMT2A-rearranged leukaemia demonstrably moves in patients: in a review of 104 published lineage-switch cases, 49 (46.2%) carried a KMT2A rearrangement, 75.5% switched from a lymphoid to a myeloid phenotype, most commonly B-ALL to AML (59.4% of cases), and 37.7% had received lineage-specific immunotherapy, which the authors propose as a trigger acting on leukaemic stem cells capable of either differentiation programme. A lineage-invariance statistic computed across DepMap lines with fixed lineage annotations is measuring something that in this disease is not fixed, and the entry's title claim - that the fusion does not move the cell type - is contradicted for the fusion class where lineage switch is most concentrated. Bełdzińska-Gądek K et al. 2024
Connected to