The five verdicts in one place, plus what this beat read and did not use
also called attribution audit summary · Q27b verdict table · fusion attribution bounded rejections
A single table saying which kinds of claim about the KMT2A::AFF1 protein hold up once you ask what reagent was actually used, and which do not. It also records the papers that looked relevant and were not used, so a future reader does not repeat the search.
The same thing, for a scientist
**The verdicts.** | claim class | reagent that carries it | survives? | why | |---|---|---|---| | **half-life and turnover** | anti-MLL Western on elutriated SEM fractions | **YES** | fusion and wild-type resolved by size in the same lane, at endogenous level [E1] | | **chromatin occupancy** | MLL-N minus MLL-C subtraction; anti-HA ChIP on a tagged fusion; degron subtraction | **PARTLY** | qualitative map survives; quantitative occupancy does not; the intersect-with-partner heuristic fails [E2][E3] | | **complex composition** | anti-AF4 AP-MS from K562; shared-epitope co-IP western; FLAG-tagged overexpression | **PARTLY** | single-partner blots survive by size resolution; the inventory does not [E3] | | **subnuclear localisation** | anti-AF4 IF in t(4;11) lines; transfected chimeras | **NO at endogenous level** | an image has no size dimension, and the fusion is the minor species [E4] | | **post-translational modifications** | shotgun phosphoproteomics; phospho-specific antibodies | **NO** | every tryptic peptide but the junction peptide is shared with a wild-type parent; no partial survivor exists [E5] | **The pattern.** Survival tracks the readout, not the reagent. Methods that keep a size or mass dimension attribute; methods that discard it do not, unless a tag or a degron restores uniqueness (`concept-the-size-dimension-is-the-attribution-reagent-this-literature-actually-has`). **Two surprises worth flagging.** First, **the surviving classes are the ones nobody advertises**: the field's headline claims are occupancy and composition, and its best-attributed endogenous measurement is a single cell-cycle blot from 2007 that is almost never cited for what it actually shows [E1]. Second, the class the brief expected to find survivors in - a partner antibody in a line expressing that partner only as the fusion - **is empty in this disease**, because every recurrent KMT2A partner is broadly expressed.
Bounded rejections - read, judged off-beat, recorded so the search is not repeated
Each is a paper a reader could reasonably expect this beat to have cited.
- PMID 29796335, TALEN-Mediated FLAG-Tagging of Endogenous Histone Methyltransferase DOT1L 6. The nearest published thing to the knock-in this beat calls for, in the same pathway, and it tags DOT1L rather than a fusion. It is evidence the technique is applicable here, not evidence it has been applied to a fusion. Would change my mind: a follow-up applying the same approach to a rearranged allele.
- PMID 40389585, Highly specific immunoproteasome inhibitor M3258 induces proteotoxic stress and apoptosis in KMT2A::AFF1 driven ALL 7. Title is squarely on the turnover beat; the content is proteotoxic stress and apoptosis in the fusion-driven disease, not a measurement of fusion protein turnover. Would change my mind: a fusion-band blot under M3258.
- PMID 26137584, Unraveling the Activation Mechanism of Taspase1 which Controls the Oncogenic AF4-MLL Fusion Protein 8. Taspase1 enzymology and inhibitor design, relevant to the reciprocal product's processing rather than to attribution of any measurement. Would change my mind: data separating taspase1 action on the fusion from action on wild-type KMT2A in the same cells.
- PMID 41274245, Development of a CRISPR/Cas9-degron system that enables in vivo specific gene depletion in leukemia models 9. Comes up on every degron-plus-leukaemia query; the degron is on Cas9, for temporal control of editing. Nothing to do with tagging a fusion.
Two query traps found, recorded for whoever searches next. (SEM OR RS4;11 OR ...) is unusable: PubMed translates SEM to Structural Equation Modeling and Scanning Electron Microscopy journal names and RS4;11 to rs4 11, returning seven records of pure noise. And the parser collapses MLL::AF4 to MLL-AF4 and KMT2A::AFF1 to KMT2A-AFF1, so a query that appears to search four fusion spellings searches two.
What this audit does not say
It does not say the field's conclusions are wrong. Most of them are probably right, for the ordinary reason that the fusion and its wild-type parents are expected to behave similarly in many respects. It says something narrower and more actionable: the confidence attached to those conclusions was earned by experiments that could not have distinguished the two, and the cost of distinguishing them is lower than the field has assumed - in several cases a change of readout rather than a new reagent.
The one place where it does bite is quantitative. Any number describing the fusion - how much of it is at a locus, what fraction of a pulldown it represents, how long it lives - is currently unmeasured, because a number requires attribution and the qualitative claims do not.
| empty class | a partner antibody in a line expressing that partner only as the fusion - no such case exists in KMT2A leukaemia, because every recurrent partner is broadly expressed |
|---|
verdict table
[
{
"claim_class": "protein half-life and turnover",
"survives": "yes",
"carrier": "anti-MLL Western on elutriated SEM fractions",
"reason": "size resolution, endogenous"
},
{
"claim_class": "chromatin occupancy",
"survives": "partly",
"carrier": "MLL-N minus MLL-C subtraction; anti-HA ChIP; degron subtraction",
"reason": "qualitative map survives, quantitative occupancy does not; intersect-with-partner heuristic fails"
},
{
"claim_class": "complex composition",
"survives": "partly",
"carrier": "size-resolved co-IP western",
"reason": "single-partner claims survive, the inventory does not"
},
{
"claim_class": "subnuclear localisation",
"survives": "no at endogenous level",
"carrier": "anti-AF4 IF; transfected chimeras",
"reason": "no size dimension in an image; overexpression images attribute but distort"
},
{
"claim_class": "post-translational modifications",
"survives": "no",
"carrier": "shotgun phosphoproteomics; phospho antibodies",
"reason": "every tryptic peptide but the junction peptide is shared"
}
]bounded rejections — the numbers behind the chart
[
{
"pmid": "29796335",
"reason": "endogenous FLAG-tagging, but of DOT1L rather than a fusion allele",
"would_change_my_mind": "the same approach applied to a rearranged allele",
"title": "TALEN-Mediated FLAG-Tagging of Endogenous Histone Methyltransferase DOT1L.",
"authors": "An C, Zhu G, Martos SN, et al.",
"venue": "Adv Biosci Biotechnol",
"year": 2017,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pmid": "40389585",
"reason": "proteotoxic stress and apoptosis in the right disease, not a measurement of fusion protein turnover",
"would_change_my_mind": "a fusion-band blot under drug",
"title": "Highly specific Immunoproteasome inhibitor M3258 induces proteotoxic stress and apoptosis in KMT2A::AFF1 driven acute lymphoblastic leukemia.",
"authors": "Jenkins TW, Fitzgerald JE, Park J, et al.",
"venue": "Sci Rep",
"year": 2025,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pmid": "26137584",
"reason": "taspase1 enzymology and inhibitor design, not attribution of a measurement",
"would_change_my_mind": "data separating taspase1 action on fusion from wild-type in the same cells",
"title": "Unraveling the Activation Mechanism of Taspase1 which Controls the Oncogenic AF4-MLL Fusion Protein.",
"authors": "Sabiani S, Geppert T, Engelbrecht C, et al.",
"venue": "EBioMedicine",
"year": 2015,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pmid": "41274245",
"reason": "the degron is on Cas9, for temporal control of editing",
"would_change_my_mind": "nothing - it is off-beat as evidence but useful as method",
"title": "Development of a CRISPR/Cas9-degron system that enables in vivo specific gene depletion in leukemia models.",
"authors": "Shimura R, Yamamoto K, Chang YH, et al.",
"venue": "Biochem Biophys Res Commun",
"year": 2025,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
]query traps found
[
{
"trap": "SEM as a search token",
"detail": "PubMed translates SEM to Structural Equation Modeling and Scanning Electron Microscopy journal names, and RS4;11 to 'rs4 11'; the query returns pure noise",
"date": "2026-09-20"
},
{
"trap": "double-colon fusion nomenclature",
"detail": "the parser collapses MLL::AF4 to MLL-AF4 and KMT2A::AFF1 to KMT2A-AFF1, so a four-spelling query searches two",
"date": "2026-09-20"
}
]- Target selection in this disease draws on composition and occupancy claims, the two classes that only partly survive, so a target list assembled from them carries an unquantified fraction of wild-type-derived entries.
- Biomarker development draws on localisation and PTM claims, the two classes that do not survive at all, so any pharmacodynamic assay built on them is measuring a mixture whose fusion fraction nobody has ever determined.
- The one class that does survive - turnover - is the one with no published number, so the quantity most relevant to dosing schedule is simultaneously the best-attributed and the least measured.
- Which of the field's qualitative conclusions actually change when attribution is enforced? The honest prior is 'few', and the way to find out is the knock-in line, whose anti-HA CUT&RUN can be overlapped with the existing MLL-N maps directly.
- Does the same audit applied to other fusion families give the same verdict table? NUP98 and RUNX1::RUNX1T1 have different partner-expression patterns, and a family where the partner is lineage-restricted would populate the empty class this beat could not fill.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In SEM cells separated by centrifugal elutriation, MLL-AF4 protein did not fluctuate through the cell cycle whereas wild-type MLL showed biphasic expression - an endogenous, fusion-attributable protein measurement made possible by size resolution on a blot.
Bimodal degradation of MLL by SCFSkp2 and APCCdc20 assures cell cycle execution: a critical regulatory circuit lost in leukemogenic MLL fusions. · Liu H, Cheng EH, Hsieh JJ. · Genes & development 2007What this evidence cannot show
Read verbatim from the PMC HTML page for PMC1993870 after fullTextXML returned HTTP 500 and efetch returned no <body>. One figure panel in one line, qualitative rather than a half-life, not replicated in a second t(4;11) line in that paper.
Evidence that argues against this (1)
- Impure elutriation fractions would flatten a real oscillation and look identical to a genuinely constant protein, and the same paper reports that three chemical synchronisation methods failed in these lines - which is a reason to hold the result at moderate rather than high confidence. Olsen SN et al. 2022
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Attribution in chromatin occupancy is restored either by an epitope unique to the fusion (anti-HA ChIP-seq on HA-FKBP12-tagged MLL::AF9) or by degron subtraction, where MLL-N signal lost on degradation with no matching MLL-C or H3K4me3 change is fusion signal.
MLL::AF9 degradation induces rapid changes in transcriptional elongation and subsequent loss of an active chromatin landscape. · Olsen SN, Godfrey L, Healy JP, et al. · Molecular cell 2022What this evidence cannot show
Read verbatim from the PMC body (PMC9044330). Retroviral overexpression, MLL::AF9 rather than KMT2A::AFF1, and a three-hour window that is not validated as an attribution window.
Evidence that argues against this (1)
- Degron subtraction cannot separate direct binding from loss of a co-factor that required the fusion, so a site losing MLL-N signal after three hours may be losing something the fusion recruited rather than the fusion itself. Kerry J et al. 2017
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The composition and occupancy literatures rest on shared-epitope reagents: the AEP inventory was purified with an anti-AF4 monoclonal from K562, which carries no KMT2A rearrangement, and AEP is physiologically recruited to MLL-target chromatin, which invalidates the heuristic of intersecting MLL-N peaks with partner peaks.
A higher-order complex containing AF4 and ENL family proteins with P-TEFb facilitates oncogenic and physiologic MLL-dependent transcription. · Yokoyama A, Lin M, Naresh A, et al. · Cancer cell 2010What this evidence cannot show
Read verbatim from the PMC body (PMC2824033). The co-occupancy evidence in that paper is ChIP-qPCR at selected loci rather than genome-wide, so the extent of physiological MLL/AEP overlap is not quantified.
Evidence that argues against this (1)
- If physiological MLL/AEP co-occupancy is confined to few loci, the intersection heuristic could be nearly valid in practice despite being invalid in principle. The control that would settle it - genome-wide wild-type MLL and wild-type AF4 occupancy in a non-rearranged haematopoietic cell - has not been run. Kerry J et al. 2017
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The only endogenous imaging of a KMT2A fusion used anti-AF4 polyclonals in t(4;11) lines, found the same punctate nuclear pattern in rearranged and non-rearranged cells, and reported the fusion as the less abundant species - so the image could not have detected a fusion-specific localisation either way.
AF4 encodes a ubiquitous protein that in both native and MLL-AF4 fusion types localizes to subnuclear compartments. · Li Q, Frestedt JL, Kersey JH. · Blood 1998What this evidence cannot show
Abstract read in full via the Europe PMC core record; no PMCID and the full text was not reached.
Evidence that argues against this (1)
- An overexpression comparison in a different family reaches the opposite conclusion: HRX/ALL1-eps15 localised to smaller, more numerous nuclear bodies than wild-type HRX/ALL1. That result is attributable and says fusion does change compartmentalisation, so the endogenous non-result is not a licence to conclude that it does not. Yokoyama A et al. 2010
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the canonical KMT2A::AFF1 chimera the KMT2A contribution ends at residue 1406 and the AFF1 contribution begins at residue 347, so every tryptic peptide except the junction-spanning one is sequence-identical to a peptide of a wild-type parent, and no peptide-level PTM assignment can distinguish fusion from parent.
Acute mixed-lineage leukemia t(4;11)(q21;q23) generates an MLL-AF4 fusion product. · Domer PH, Fakharzadeh SS, Chen CS, et al. · Proc Natl Acad Sci U S A 1993What this evidence cannot show
Junction anchor GenBank AAC37520.1 from this paper; residue arithmetic is this project's own computation, independently reproduced by the coordinator to four decimal places. Breakpoint variation moves the junction between patients; the structural argument holds for every in-frame breakpoint.
Evidence that argues against this (1)
- A protein-level rather than peptide-level PTM readout escapes the argument entirely - an IP followed by a phospho-blot with the fusion band resolved by size would attribute occupancy of a known site, though not discover a new one. That design has not been published for any KMT2A fusion. Li Q et al. 1998
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Endogenous FLAG-tagging by genome editing has been applied to DOT1L, a KMT2A-pathway histone methyltransferase, demonstrating the technique in this field - but not to a fusion allele.
TALEN-Mediated FLAG-Tagging of Endogenous Histone Methyltransferase DOT1L. · An C, Zhu G, Martos SN, et al. · Advances in bioscience and biotechnology (Print) 2017What this evidence cannot show
Cited from the PubMed record for existence and subject; the full text was not retrieved and no quantity is asserted. Published in a low-visibility journal, which is itself a reason it is easy to miss and worth recording. Recorded here as a bounded rejection, not as support for any claim about fusions.
Evidence that argues against this (1)
- TALEN-based tagging in 2017 is a different technical proposition from CRISPR knock-in in a suspension leukaemia line today, so its existence is weaker evidence of feasibility than it first appears. Olsen SN et al. 2022
- E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
An immunoproteasome inhibitor induces proteotoxic stress and apoptosis in KMT2A::AFF1-driven acute lymphoblastic leukaemia - a protein-degradation paper in the right disease that does not measure fusion protein turnover.
Highly specific Immunoproteasome inhibitor M3258 induces proteotoxic stress and apoptosis in KMT2A::AFF1 driven acute lymphoblastic leukemia. · Jenkins TW, Fitzgerald JE, Park J, et al. · Scientific reports 2025What this evidence cannot show
Cited from the PubMed record for subject only; the full text was not retrieved. Recorded as a bounded rejection: a reader searching protein degradation in this disease will find it and should know why this beat did not use it.
Evidence that argues against this (1)
- If that paper does contain a fusion-band blot under drug, it would belong in the turnover audit after all, and this rejection would be wrong. The full text was not read, so the rejection is on the abstract. Liu H et al. 2007
- E8moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Taspase1 activation mechanism work addresses the protease that processes the KMT2A C-terminus and hence the reciprocal product, rather than the attribution of any measurement to a fusion.
Unraveling the Activation Mechanism of Taspase1 which Controls the Oncogenic AF4-MLL Fusion Protein. · Sabiani S, Geppert T, Engelbrecht C, et al. · EBioMedicine 2015What this evidence cannot show
Cited from the PubMed record for subject only; the full text was not retrieved. Recorded as a bounded rejection because its title names AF4-MLL and would lead a reader to expect it here.
Evidence that argues against this (1)
- Taspase1 cleavage is genuinely attribution-relevant in one respect: it is what makes wild-type KMT2A a two-epitope species and the fusion a one-epitope species, which is the whole basis of the MLL-N minus MLL-C subtraction. The rejection is of the enzymology, not of the cleavage fact. Kerry J et al. 2017
- E9moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A CRISPR/Cas9-degron system for leukaemia models places the degron on Cas9 for temporal control of editing, not on a fusion oncoprotein.
Development of a CRISPR/Cas9-degron system that enables in vivo specific gene depletion in leukemia models. · Shimura R, Yamamoto K, Chang YH, et al. · Biochemical and biophysical research communications 2025What this evidence cannot show
Abstract read in full via the Europe PMC core record; no PMCID. Recorded as a bounded rejection because it is the top hit on several degron-plus-leukaemia queries and looks on-beat.
Evidence that argues against this (1)
- The system is nonetheless useful infrastructure for the knock-in proposed here, because temporal control of Cas9 reduces culture-induced pre-selection during clone generation - so it is off-beat as evidence and on-beat as a method. Olsen SN et al. 2022
verdict 1 of 5
verdict 2 of 5
verdict 3 of 5
verdict 4 of 5 - the one that survives
verdict 5 of 5
the single property that predicts the verdict table
the single experiment that would settle the four classes that do not fully survive