Oncofusion

start / graph / hyp-reciprocal-product-as-enforceable-remainder

Hypothesisspeculative confidence 1 of 5: Our own inference or analogy. Not shown to be true.Q12-analogy

Hypothesis: the reciprocal AFF1::KMT2A product protects only when it is structurally complete, so its prognostic effect is a breakpoint class in disguise

In half of these leukemias the swap produces a second, mirror-image gene as well as the cancer-causing one, and children whose leukemia makes it do far better. We think that is not luck: the mirror gene only helps when the break left it intact, which means the real prognostic variable is where the break fell.

The same thing, for a scientist

Reciprocal AFF1::KMT2A transcript expression is present in about 45% of t(4;11) infant B-ALL and predicted 4-year event-free survival of 62.4% versus 11.7%, retaining significance in a Cox model adjusted for risk stratification [E1]. The recombinome separately reports that about half of KMT2A::AFF1 cases break upstream of AFF1 exon 1, which produces no reciprocal allele at all [E2]. Contract law supplies the frame: the severed remainder is enforceable only if it is itself a complete agreement. The hypothesis is that reciprocal expression is not a stochastic transcriptional property but a deterministic consequence of partner-side breakpoint position, and that its protective effect requires the remainder to retain a functional KMT2A C-terminus (PHD-bromodomain, FYRN/FYRC, SET). If so, the prognostic variable the field should be measuring is the partner-side breakpoint, which is cheaper and more stable to assay than transcript expression. This is our inference from an analogy plus two published observations; it has not been tested.

Go deeper
The two published facts that have never been put in the same model1 deeper

Reciprocal expression predicts outcome in 124 uniformly treated infants 1. The argument this entry originally built on that - that roughly half of KMT2A::AFF1 cases break upstream of AFF1 exon 1 and so cannot make a reciprocal allele 2 - was wrong, and wrong by about a factor of 600. In AFF1 that break happens in 1 of 1,262 cases. The structural population the hypothesis needed does not exist in this partner.

The hypothesis survives by moving house. The population does exist - in MLLT1, at 241 of 460 (about 50%), which is where the misremembered number came from. So the test is available, just not in t(4;11): KMT2A::MLLT1 splits into roughly 241 reciprocal-negative and roughly 219 reciprocal-positive patients with the partner gene held constant. That is the one comparison in this field free of the partner, breakpoint-class and direct-product confounds, and nobody has run it.

What remains true of t(4;11) is narrower and still unexplained: 45% express the reciprocal transcript while about 80-94% carry the allele. The gap is therefore mostly transcriptional or clonal, not structural - a different mechanism from the one this entry proposed, and one no cohort has resolved.

Why that matters practically

Transcript expression is a fragile assay: it depends on sample quality, blast fraction and RT-PCR sensitivity. Genomic breakpoint position is stable, is already sequenced in Interfant-06 as an MRD target for 249 patients, and does not degrade. If the two are equivalent, the prognostic marker becomes essentially free and retrospectively available.

The mechanistic half of the claim, and why it might be wrong

The reciprocal product places an AFF1 5' segment in front of the KMT2A C-terminus, which carries the SET methyltransferase and the PHD-bromodomain block lost from the oncogenic fusion. Under the severability frame, a remainder is enforceable only if complete, so the protective effect should depend on retention of that C-terminus and on an in-frame junction.

The honest alternative is more mundane and at least as likely: reciprocal-positive cases may simply be those with a simple, balanced translocation, and simple translocations may mark a less genomically disturbed leukemia that does better for reasons having nothing to do with the reciprocal protein. That confound is not separable in existing outcome data and must be built into any analysis as a covariate, not assumed away.

Facts and figures
statementReciprocal AFF1::KMT2A expression is determined by partner-side breakpoint position rather than arising stochastically, and its reported protective association holds only in cases where the reciprocal product is in frame and retains the KMT2A C-terminal PHD-bromodomain and SET modules; partner-side breakpoint position should therefore predict outcome at least as well as reciprocal transcript expression does.
originanalogy
source analogyanalogy-law-severability-domain-retention
prior probability0.4
decision valueIf confirmed, it converts an unstable transcript assay into a stable genomic one that Interfant-06 already holds for 249 patients, and it would be the first mechanistic account of the largest unexplained prognostic effect in infant KMT2A-rearranged ALL.
predictions
[
  "In any cohort with both partner-side breakpoint sequence and reciprocal transcript status, the two are near-deterministically related rather than independent.",
  "Partner-side breakpoint position replaces reciprocal transcript expression in a Cox model without loss of predictive value.",
  "Among reciprocal-positive cases, the protective effect is concentrated in those whose reciprocal product is in frame and retains SET; out-of-frame reciprocal transcripts should behave like reciprocal-negative cases.",
  "The effect should attenuate, but not vanish, after adjustment for karyotypic complexity, if the remainder itself is doing work rather than merely marking a simple translocation.",
  "The same relationship should appear in other KMT2A partners where reciprocal products are variably produced, most cleanly in MLLT3 and MLLT1."
]
experiments
[]
analyses
[
  "ana-interfant-breakpoint-prognostic-reanalysis",
  "ana-breakpoint-linker-vs-fragility-decomposition"
]
falsifiers
[
  "Reciprocal transcript status and partner-side breakpoint position are only weakly associated, showing expression is stochastic or epigenetically controlled rather than structural.",
  "The protective association is equally strong in out-of-frame and in-frame reciprocal transcripts, which would mean the remainder protein is irrelevant.",
  "Reciprocal expression loses significance entirely once karyotypic complexity is adjusted for, showing it was a marker of a simple translocation and nothing more.",
  "The original association fails to replicate in an independent Interfant cohort, in which case there is nothing to explain."
]
How this could bite us back
  • If reciprocal-negative cases are a structurally distinct and far worse-prognosis class, then trial populations that happen to be enriched for one class will give misleading estimates of menin-inhibitor benefit, and no current trial records this variable.
  • A protective reciprocal product would mean part of the benefit in some patients comes from restored wild-type-like KMT2A activity, which sits awkwardly with any future strategy that degrades KMT2A broadly rather than blocking the fusion interface.
Still open
  1. Run the comparison in KMT2A::MLLT1, not KMT2A::AFF1: about 241 reciprocal-negative against about 219 reciprocal-positive patients with the partner held constant. This is the hypothesis's real test and the numbers already exist in the Frankfurt archive.
  2. Why do only 45% of t(4;11) patients express the reciprocal transcript when 80-94% carry the allele? With the spliced-fusion explanation removed the gap has no account at all.
  3. Is the reciprocal AFF1::KMT2A protein actually produced, or only the transcript? No protein-level demonstration appears in the indexed literature, and the whole mechanistic half of this hypothesis rests on it.
  4. If the reciprocal product is protective because it restores SET activity in trans, does forced expression of a KMT2A C-terminal fragment slow KMT2A::AFF1 leukemia in a model?
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 124 infants with B-ALL treated on Interfant protocols, the reciprocal AFF1::KMT2A transcript was expressed in only 45% of t(4;11) cases, and its expression predicted 4-year event-free survival of 62.4% versus 11.7%, retaining significance in a Cox model adjusted for risk stratification.

    PMID 30679323DOI
    What this evidence cannot show

    Single cohort, n=124, retrospective; the effect size is very large for a single unreplicated marker, which is itself a reason for caution.

    Evidence that argues against this (1)
    • The Interfant-06 multivariable prognostic analysis in 651 patients identified KMT2A rearrangement, age, white-cell count and prednisone response, and did not include reciprocal expression or breakpoint position — so the effect has never been tested against the accepted prognostic model. Pieters R et al. 2019
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.journal-article

    Spliced fusions - where the KMT2A break lies upstream of the partner's exon 1, so no reciprocal allele can be made - are rare in KMT2A::AFF1: 1 of 1,262 cases (0.08%). They are common for only one partner, MLLT1, at 241 of 460 (about 50%). Other partners: EPS15 12/51, ELL 10/119, PRPF19 3/3, AFDN 2/138, MLLT3 2/612; 275 cases in total.

    PMID 37019990DOI
    What this evidence cannot show

    **This replaces a claim that was wrong by roughly a factor of 600 and that this hypothesis was built on.** The earlier text read *'About half of KMT2A::AFF1 cases break upstream of AFF1 exon 1, producing no reciprocal allele'* and cited this same paper, which says the opposite for AFF1. Verified 2026-09-19 against the full text (PMC10169636, 98,196 characters), verbatim: *"Spliced fusions are rare events except for MLLT1 gene fusions (n = 460), where about 50% of the breakpoints (n = 241) localize upstream of MLLT1 exon 1... AFF1 (1 out of 1262 cases)"*. The counts are breakpoint-level from one diagnostic archive (Frankfurt DCAL), so they carry that archive's referral bias, and a spliced fusion is only one of several routes to reciprocal-negative status - an out-of-frame reciprocal allele or loss of the derivative chromosome gets there too.

    Evidence that argues against this (1)
    • **This project had already found this exact error once.** `mech-breakpoint-position-consequences` carries a correction section with this same table, and the correction did not propagate to this entry. That is the second time a corrected fact has failed to reach a sibling entry in this graph, and it is an argument for checking siblings by citation rather than by topic. Leukemia 2023
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Contract law treats a severed remainder as enforceable only when what is left is itself a complete and coherent agreement; severance that leaves an incomplete bargain fails.

    Severability clause (Wex legal encyclopedia) · Legal Information Institute, Cornell Law School · Cornell Law School Legal Information Institute 2024
    What this evidence cannot show

    Source-domain reference only; it motivates the hypothesis and supports none of its biological content.

    Evidence that argues against this (1)
    • A legal doctrine is a normative rule about what courts do, not a fact about molecules. The analogy is generative — it suggests what to look for — and it cannot supply evidence, because the doctrine was constructed to serve purposes that have nothing to do with this system, and its edge cases were settled by argument rather than by measurement. Every prediction it produces still has to be tested against biology. U.S. National Library of Medicine 2026
Connected to