If the target promoter set is no more conserved within a patient than between patients, then nothing in the oncoprotein complex remembers the first disease
The cancer protein does not seem to carry a memory of where it used to sit when the leukaemia changes type. Where it lands looks like it is decided fresh by the cell it finds itself in.
The same thing, for a scientist
A same-patient pair shares a genome, a founding translocation, a breakpoint and a clonal history. If any of those imposed a durable addressing constraint on the oncoprotein - a sequence-level preference, a retained chromatin mark, an inherited accessibility landscape - the two states should overlap more than two unrelated children with the same fusion do. Measured on one universe with one rule and a matched null, they do not: 37th percentile of 27 unrelated KMT2A::AFF1 pairs, inside the central 80%. The mechanistic reading is that oncoprotein addressing is dominated by the CURRENT chromatin state of the cell, and that the genotype-level facts the two samples share - including the menin interface, which is identical in both - are permissive rather than instructive.
What is held constant between p279 and p318 and still does not predict the target set
- the germline genome
- the KMT2A::AFF1 fusion and its breakpoint
- the menin-binding motifs MBM1/MBM2 and therefore the menin interface itself
- the CXXC domain and its CpG addressing preference
- the patient's clonal ancestry
What changes: the lineage programme, the accessibility landscape, the cofactor availability, and the oncoprotein's own expression level (which the paper reports falls at lineage switch). The things that change are the things that predict the target set. That is a mechanistic statement, not a statistical one, and it is what makes this measurement relevant to the corpus's central question about the menin interface.
The asymmetry that survives
Only 54.6% of the B-ALL's target promoters are still called in the AML and 40.7% of the AML's calls are new. Both figures are above the matched chance floor, so there IS continuity - a little over half. But the continuity is the same size as the continuity between two unrelated children, which means it is the fusion's generic preference for active promoters rather than anything this patient's first disease established.
The competing reading this does not exclude
The oncoprotein LEVEL falls at lineage switch 1. At a fixed absolute threshold, a lower level produces fewer calls at the weaker sites, and losing the weak tail of a ranked list looks exactly like redistribution. This beat's depth matching removes library-depth differences but cannot remove a real difference in how much oncoprotein is present. A dilution model and a redistribution model both predict the observed Jaccard, and separating them needs a level-matched comparison that this data cannot supply - no spike-in exists in the deposit.
- If addressing is set by current chromatin state rather than by genotype, then a menin inhibitor's effect on WHERE the oncoprotein sits will be state-dependent and will differ between the same patient's two diseases - so pharmacodynamic markers validated at diagnosis may not read out at relapse.
- It also means resistance cannot be predicted from the fusion or the breakpoint, which are the two things a diagnostic assay actually measures.
- Does the continuous promoter-signal correlation between the two states (high, ~0.85) versus the thresholded-set overlap (at the unrelated base rate) mean the memory is real but sub-threshold? A quantitative rather than binary comparison of the two states at matched level would answer it.
- ENL persists at oncoprotein-target loci where the oncoprotein is undetectable in the revumenib case. Is ENL's target set more conserved within a patient than the oncoprotein's is? That would locate the memory in the cofactor rather than the fusion.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In lineage-switching samples the oncoprotein levels are reduced and the oncoproteins preferentially activate granulocyte-monocyte progenitor genes, i.e. both the amount and the destination change together.
MLL oncoprotein levels influence leukemia lineage identities · Janssens DH, Duran M, Otto DJ, Wu W, Xu Y, Kirkey D, Mullighan CG, Yi JS, Meshinchi S, Sarthy JF, Ahmad K, Henikoff S · Nature Communications 15:9341 2024What this evidence cannot show
The paper's level measurement and its destination measurement come from the same channels, so a level fall and a destination change are not independently established; that is the dilution/redistribution confound named above.
Evidence that argues against this (1)
- The paper also reports ENL persisting on numerous oncoprotein-target loci in the sample where the oncoprotein is undetectable - so at least some target loci retain a cofactor when the oncoprotein has gone, which is evidence for a memory the oncoprotein channel cannot see. Janssens DH et al. 2024
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
The overlap measurement itself: same-patient pair at the 37th percentile of 27 unrelated KMT2A::AFF1 sample pairs on excess over an activity-stratified null, inside the central 80%, in the primary cell.
GSE252378: AutoCUT&RUN of KMT2A-rearranged primary leukemia samples (KMT2A N- and C-terminal antibodies, ENL, DOT1L, H3K4me3, IgG) · Janssens DH, et al. · NCBI Gene Expression Omnibus 2024What this evidence cannot show
n = 1 same-patient pair. A single pair cannot establish a general mechanism, and the post-hoc institution-matched analysis places it lowest rather than average.
Evidence that argues against this (1)
- The high rank correlation between the two samples' promoter signal profiles (Spearman 0.85 on the KMT2A N channel, measured by an earlier beat on the same samples) says the underlying landscapes ARE similar - so 'nothing remembers' is a statement about thresholded sets and would be too strong if applied to the continuous signal. Janssens DH et al. 2024
the overlap measurement this mechanism reads
the 54.6% retained / 40.7% new asymmetry
the same argument, now with a matched null and 27 comparator pairs instead of 3 samples
- part of →Lineage switch in KMT2A-rearranged leukemia — the same clone changes what kind of blood cell it is
the chromatin-level account of the switch
the menin interface is identical in both states and does not predict the target set
the mechanism that makes the map non-durable