Oncofusion

start / graph / absence-af4-mll-affinity-purification-unreadable-by-all-four-routes

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q27b-fusion-attribution

Bounded unreadable: the only affinity purification of a SET-retaining KMT2A-derived fusion complex fails all four routes, and the molecule is the reciprocal product

also called PMID 21030982 unreadable · AF4-MLL complex unreadable · 16-component complex provenance

One 2011 paper is the only source for the claim that the mirror-image fusion protein sits in a sixteen-piece machine, and it cannot be read by any free route - only its summary is public. Its summary also shows the protein in question is the mirror-image product, not the main fusion this graph is about.

The same thing, for a scientist

This entry exists so that a future cycle does not repeat the search. **PMID 21030982** (Benedikt et al., *Leukemia* 2011) is the source for two numbers this graph carries - an 11-component wild-type AF4 complex and an AF4-MLL complex of at least 16 constituents - and for the claim that the AF4-MLL complex contains P-TEFb, H3K4me3 and H3K79me3 methyltransferases, a protein arginine N-methyltransferase and a histone acetyltransferase. All of that is from the **abstract**, which is public. Nothing beyond the abstract has been read by this project. **All four routes were re-run on 2026-09-20 and all four failed.** | route | result | |---|---| | 1. EPMC `search?query=EXT_ID:21030982&resultType=core` | `pmcid: null`, `isOpenAccess: N`, `inPMC: N`, `hasPDF: N`; `fullTextUrlList` offers only a subscription DOI link [E1] | | 2. `{PMCID}/fullTextXML` | no PMCID exists; the speculative `PMC21030982` path returns HTTP 500 | | 3. `efetch db=pmc` | `error id="21030982": The following PMCID is not available`; `elink dbfrom=pubmed db=pmc` returns only `pubmed_pmc_refs`, 48 citing articles, and **no** `pubmed_pmc` link [E2] | | 4. PMC HTML article page | cannot be constructed - there is no PMCID. The publisher page was fetched instead: HTTP 200, 368,813 bytes, **18,543 characters** of stripped text comprising abstract, paywall notice, a bare figure list and the reference list, with no Results and no Methods [E3] | This is a different failure mode from the one the brief's four-route rule was written for. That rule exists because a publisher can forbid XML while serving HTML; here **there is no PMC record at all**, so routes 2, 3 and 4 are structurally unavailable rather than blocked. Route 4's substitute - the publisher's own page - was tried and returns abstract only. **A second finding, which is a correction to the dispatch brief.** The brief describes this as 'the only affinity purification of a SET-retaining KMT2A fusion complex'. The abstract makes clear the molecule is **AF4-MLL, the reciprocal product** of t(4;11) - AFF1 N-terminus fused to the KMT2A C-terminus. It is SET-retaining precisely because it is the reciprocal. It is not KMT2A::AFF1, and the composition audit must not treat it as such.

Go deeper
What may and may not be asserted from this paper

May be asserted, with weak confidence and the source named as the abstract:

  • that the authors report an AF4 complex of 11 binding partners and an AF4-MLL complex of at least 16 constituents;
  • that they report P-TEFb kinase, H3K4me3 and H3K79me3 methyltransferase, protein arginine N-methyltransferase and histone acetyltransferase activities in the AF4-MLL preparation;
  • that they propose AF4-MLL disturbs the activation cycle of promoter-arrested RNA Pol II.

May not be asserted: the purification reagent, the epitope tag if any, the cell type, whether the protein was endogenous or ectopic, the gene symbols behind the activity classes, whether the H3K4me3 activity is the fusion's own SET domain or a co-purified SET1-family enzyme, or any quantity from a figure. This graph's entry complex-af4-mll-assembles-a-sixteen-component-elongation-machine currently states that the purification was done 'on over-expressed protein'; that is a reasonable inference from the model system the abstract describes, but it is an inference and should be marked as one.

What would make it readable
  • Institutional access. Leukemia is a Springer Nature subscription journal and Northeastern or the Broad may hold it. This is the cheapest fix and it is not a research problem.
  • Author request. The corresponding laboratory (Marschalek, Frankfurt) is the same group behind the recombinome and the gene-internal promoter work this graph already cites, and is listed in this graph's collaborator entries.
  • The 48 citing articles returned by elink pubmed_pmc_refs are in PMC and free. Several will describe the purification in their introductions in more detail than the abstract does. That is second-hand and should be labelled as such, but it is a real route to the gene symbols.
Facts and figures
pmid21030982
statusunreadable by all four routes, abstract only
date checked2026-09-20
moleculeAF4::KMT2A (reciprocal product), NOT KMT2A::AFF1
titleThe leukemogenic AF4-MLL fusion protein causes P-TEFb kinase activation and altered epigenetic signatures.
authorsBenedikt A, Baltruschat S, Scholz B, et al.
venueLeukemia
year2011
title provenanceTitle/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.
absence classunclassified-not-checked
absence class basisNo archive query has been run for this absence. Assigned as the honest default rather than guessed; a single query in the matching repository would settle it.
absence class assigned2026-09-21 (cycle 47, king)
routes
[
  {
    "route": "1. EPMC search EXT_ID core",
    "result": "pmcid null, isOpenAccess N, inPMC N, hasPDF N"
  },
  {
    "route": "2. {PMCID}/fullTextXML",
    "result": "no PMCID exists; speculative PMC21030982 path returns HTTP 500"
  },
  {
    "route": "3. efetch db=pmc (test for <body>)",
    "result": "error - PMCID not available; elink gives only pubmed_pmc_refs (48 citing articles)"
  },
  {
    "route": "4. PMC HTML article page",
    "result": "not constructible - no PMCID; publisher page substituted: HTTP 200, 18,543 chars, abstract + paywall only"
  }
]
what the abstract supports
[
  "AF4 complex of 11 binding partners",
  "AF4-MLL complex of at least 16 constituents",
  "activity classes: P-TEFb kinase, H3K4me3 and H3K79me3 HMTs, protein arginine N-methyltransferase, histone acetyltransferase",
  "AF4-MLL expression in murine HSPC gives proB ALL"
]
what it does not support
[
  "the purification reagent or tag",
  "the cell type purified from",
  "endogenous versus ectopic expression",
  "gene symbols behind the activity classes",
  "whether H3K4me3 activity is the fusion's own SET domain or a co-purified SET1-family enzyme"
]
routes to readability
[
  "institutional subscription to Leukemia",
  "author request to the Marschalek laboratory",
  "the 48 citing articles in PMC, as clearly-labelled second-hand description"
]
How this could bite us back
  • A therapeutic argument that the reciprocal product is a catalytically active H3K4 methyltransferase - and therefore a target in its own right - currently rests on a single unreadable paper's abstract, so a programme built on it would be building on an inference nobody in this project has been able to check.
  • About half of KMT2A::MLLT1 patients express no reciprocal product at all, so any intervention aimed at the reciprocal product is addressed to a subset defined by an RT-PCR call, and the size of that subset for t(4;11) specifically is not established here.
Still open
  1. What are the gene symbols of the 16 constituents? The abstract gives activity classes only, and a complex inventory described by enzymatic activity rather than by subunit cannot be joined to anything else in this graph.
  2. Was the H3K4me3 activity in the AF4-MLL preparation the fusion's own SET domain or a co-purified SET1-family enzyme? That single question decides whether the reciprocal product is a catalytically active KMT2A species, and it cannot be answered from the abstract.
Every source behind this entry (4)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    The Europe PMC core record for PMID 21030982 returns pmcid null, isOpenAccess N, inPMC N and hasPDF N, with a fullTextUrlList offering only a subscription-required publisher link.

    What this evidence cannot show

    A live API response recorded on 2026-09-20. Europe PMC coverage changes: a paper with no PMCID today can acquire one, for example through a funder deposit, so this record should be re-checked rather than treated as permanent. The brief's own rule is that an empty abstract field carries no information about readability; here the abstract field is populated and the pmcid field is the one that is null, which is the field that matters.

    Evidence that argues against this (1)
    • Europe PMC occasionally lags NCBI, so a null pmcid there is not by itself decisive - which is why the NCBI ID converter and elink were run independently and agree. NCBI E-utilities 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    The NCBI PMC ID converter returns status error with 'Identifier not found in PMC' for PMID 21030982, efetch db=pmc returns 'The following PMCID is not available', and elink dbfrom=pubmed db=pmc returns only a pubmed_pmc_refs linkset of 48 citing articles with no pubmed_pmc link to the article itself.

    What this evidence cannot show

    Live API responses recorded on 2026-09-20. The elink result is the one worth noting for method: a pubmed_pmc_refs linkset is well-formed and populated and looks like a successful lookup, but it lists articles that CITE the paper, not the paper. Reading it as a PMCID hit would give 48 wrong identifiers.

    Evidence that argues against this (1)
    • A pubmed_pmc_refs linkset of 48 free full texts is not nothing: those 48 papers are readable and several will restate the purification in their introductions. The absence is of the primary source, not of all information about it. Benedikt A et al. 2011
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    The publisher page for the article returns HTTP 200 and 368,813 bytes, which strip to 18,543 characters consisting of abstract, subscription notice, a bare figure list and the reference list, with no Results or Methods text.

    What this evidence cannot show

    Fetched without authentication from this project's network on 2026-09-20. An institutional subscription would very likely return the full text from the same URL, so this is a statement about free access rather than about the article's existence online. The stripped character count is given for comparison with the brief's benchmark cases, where a readable PMC HTML page gave 69,414 and 82,300 characters.

    Evidence that argues against this (1)
    • 18,543 characters is large enough to look like a successful full-text fetch to any check that tests only byte count or HTTP status - the same family of failure as the body-less PMC XML and the zero-byte supplementary file this project has already logged. The test has to be for Methods and Results content, not for size. Europe PMC 2026
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The molecule purified is AF4-MLL, the reciprocal product of t(4;11), whose expression in murine haematopoietic progenitor/stem cells results in proB acute lymphoblastic leukaemia; it retains the KMT2A SET domain because it carries the KMT2A C-terminus, and it is therefore not KMT2A::AFF1.

    PMID 21030982DOI
    What this evidence cannot show

    Read from the abstract only, which states the model system and names both complexes. The abstract does not state the cell type the purification was performed from, which is the detail that would settle whether the composition claim is endogenous or ectopic.

    Evidence that argues against this (1)
    • The two products of the same translocation are not independent: a preparation of AF4-MLL from a cell that also expresses KMT2A::AFF1 could contain both, and no reagent described in the abstract would separate them. Without the methods this cannot be assessed in either direction. Yokoyama A et al. 2010
Connected to