Oncofusion

start / graph / ana-sima6-is-junction-spanning-and-it-pins-sems-breakpoint

Analysisunrated confidence 0 of 5: unrated

siMA6 confirmed junction-spanning by reconstruction - 8 KMT2A bases plus 13 AFF1 bases - and it targets the 1362/1363 breakpoint, not the one this graph modelled as canonical

also called siMA6 alignment · SEM breakpoint from the siRNA

A short RNA used to switch off the fusion gene was said to sit exactly across the join between the two halves. Rebuilding the join from scratch shows it does - and shows which of two possible joins the cells in question actually carry.

The same thing, for a scientist

`ana-attribution-audit-chromatin-occupancy-kmt2a-aff1` rests on the MLL-AF4 siRNA being junction-spanning, taken on the paper's own description. Reconstructing the chimeric transcript from RefSeq NM_001197104.2 and NM_005935.4: **siMA6 is 8 KMT2A bases plus 13 AFF1 bases** and **matches neither parent transcript alone**. The claim is independently confirmed. But it matches only the chimera built at the **UniProt 1362/1363** breakpoint, not the **1406/1407** one that `analyses/junction-peptide-detectability` modelled as canonical - so SEM carries the former, and the junction peptide computed earlier is for a different fusion.

Go deeper
The arithmetic, and it confirms the reagent

siMA6 sense, from the supplemental methods of PMID 28076791: AAGAAAAGCAGACCUACUCCA (21 nt), sequences attributed there to Thomas et al. 2005.

Reconstructing the chimera with KMT2A ending at transcript nt 4107 and AFF1 resuming at its nt 1175:

...GTAAAACAAAAACCAAAAGAAAAG | CAGACCTACTCCAATGAAGTCCAT...
   ...V  K  Q  K  P  K  E K | Q  T  Y  S  N  E  V  H ...
                  AAGAAAAG   CAGACCTACTCCA          <- siMA6

8 bases from KMT2A, 13 from AFF1. The 21-mer occurs in neither wild-type transcript. The reagent is fusion-specific by sequence, not by assertion, and the audit's verdict - that this is the only fusion-attributable perturbation of an endogenous KMT2A::AFF1 protein in this literature - stands on a verified foundation.

Note how little of it is KMT2A. Eight bases is short, and an siRNA tolerates mismatches, so specificity rests on the 13-base AFF1 arm plus those eight. That is a design constraint worth knowing rather than a defect: it is why the paper pairs siMA6 with a mismatch control rather than a scrambled one.

And it corrects something this graph computed for itself

UniProt annotates three recurrent translocation breakpoints in KMT2A: 1362/1363, 1406/1407 and 1444/1445. analyses/junction-peptide-detectability modelled the 1406/1407 one as the canonical KMT2A::AFF1 architecture and computed its junction peptide - VDFKQTYSNEVHCVEEILK, MH+ 2281.1169 - which I verified to four decimal places.

siMA6 does not match that chimera. It matches only the 1362/1363 one:

donor breakpointKMT2A tailsiMA6 present?
~1406/1407 (modelled canonical)...HRIRVDFKno
~1362/1363...VKQKPKEKyes, 8 + 13

Since siMA6 was designed against "the MLL exon 9 and AF4 exon 4 MLL-AF4 fusion site present in SEM cells", SEM carries the 1362/1363 breakpoint - and the junction peptide this graph computed is for a different, also real, KMT2A::AFF1 fusion.

The AFF1 acceptor is constant and the KMT2A donor is not. Both chimeras resume at QTYSNEVHCVEEILK, which is why the AFF1 half of siMA6 matched immediately while the KMT2A half matched nothing - the signature of a shared acceptor with a different donor exon.

Consequence for the PRM experiment. exp-junction-peptide-prm-assay-kmt2a-fusions proposes targeting a junction peptide in patient material. Run in SEM, it must target the 1362/1363 peptide - the KMT2A side there ends ...PKEK, and as at the other breakpoint trypsin cuts at that terminal K, so a missed cleavage is again required. A PRM assay built on the 2281.1169 mass would return a clean negative in SEM and the negative would mean nothing.

Still open
  1. Read SEM's KMT2A::AFF1 junction directly from a published sequence rather than inferring it from an siRNA. It is a fact this graph asserts and has never sourced.
  2. Compute the junction peptide for all three UniProt-annotated breakpoints, not one, and give the PRM assay entry a mass per breakpoint. A single-mass assay returns an uninterpretable negative in any patient carrying a different one.
  3. Model siMA6's seed region against wild-type AFF1. Thirteen of its 21 bases are AFF1, so partial wild-type AFF1 knockdown is a live possibility that the mismatch control does not separately exclude.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis

    The siMA6 21-mer spans the KMT2A::AFF1 junction with 8 bases from KMT2A and 13 from AFF1, and occurs in neither wild-type parent transcript, confirming by reconstruction that the reagent is fusion-specific by sequence.

    What this evidence cannot show

    Computed 2026-09-20 against RefSeq NM_001197104.2 and NM_005935.4, with the junction located by anchoring known peptides in all three frames rather than by any coordinate written from recall. **An exact 21-mer match is a stronger test than an siRNA actually needs**: RISC tolerates mismatches, particularly outside the seed, so 'matches neither parent exactly' does not prove no off-target knockdown of either parent. The 8-base KMT2A arm is short enough that seed-region considerations matter and were not modelled here.

    Evidence that argues against this (1)
    • A junction-spanning **sequence** guarantees a fusion-specific **target site**, not a fusion-specific **effect**. The published knockdown data would still be consistent with partial depletion of wild-type AFF1 through the 13-base AFF1 arm, and the paper's mismatch control tests specificity of the whole duplex rather than of each arm. Cell Reports 2017
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis

    siMA6 matches a chimera built at the UniProt 1362/1363 KMT2A breakpoint and does not match one built at 1406/1407. Since it was designed against the fusion site present in SEM cells, SEM carries the 1362/1363 breakpoint, and the junction peptide this project computed for the 1406/1407 architecture is for a different fusion.

    What this evidence cannot show

    The inference from 'the siRNA matches breakpoint X' to 'SEM carries breakpoint X' depends on the paper's statement that siMA6 targets the fusion site present in SEM, which is an assertion in a methods section rather than a sequencing result. **SEM's breakpoint has not been read off a sequence here.** RefSeq and UniProt numbering differ by about seven residues for KMT2A, so the breakpoint labels are approximate; the transcript nucleotide positions (4107 against 4239) are exact.

    Evidence that argues against this (2)
    • A cleaner reading is available and would be better: the SEM junction has been sequenced in the primary literature more than once, and reading it directly would replace this inference with a fact. This analysis reached it sideways, from a reagent, because that is what the open question asked for. Oncofusion knowledge graph (internal analysis) 2026
    • This does not invalidate the earlier junction-peptide computation - the 1406/1407 architecture is a real recurrent breakpoint and its peptide is correctly computed. What is wrong is calling it *the* canonical KMT2A::AFF1 architecture, and building a patient-facing assay on one breakpoint when three are annotated. Oncofusion knowledge graph (internal analysis) 2026
Connected to