Bounded failure: a PMCID that exists and a body that does not - the MLL amino-terminus transformation paper is unreadable by both PMC routes
also called PMID 17671196 bounded failure · PMC7566887 empty body · MLL N-terminal menin transformation
A paper this graph cites for menin binding through the front end of MLL does not say which fusion partners it used. It has a PMC record, so it looks readable, but both ways of asking PMC for the text return only the title page - and it deposited no data.
The same thing, for a scientist
`ana-menin-dependence-untested-in-septin-and-cytoskeletal-kmt2a-fusions` hedges PMID 17671196 with *"The fusion partners used are not specified in the abstract and the full text was not consulted for this entry"* [E1]. This beat consulted it and could not read it either, in a failure mode worth recording because **it looks like success at every checkpoint until the last one**. PubMed esummary returns `pmc: PMC7566887`. Europe PMC returns `pmcid: PMC7566887` with `inEPMC: Y`. By the cycle-50 rule's first clause, a PMCID exists, so the paper is not unreadable. But `fullTextXML` returns **HTTP 500** on repeated attempts across two sessions, and `efetch db=pmc` returns **7,476 characters of front matter with no `<body>` element** - author list, affiliations, custom-meta tags including `pmc-prop-in-epmc: yes`, and nothing else [E2]. Four papers in this beat share that exact signature (PMC7566887, PMC3593849, PMC6234363, PMC3795288); five others with the same Europe PMC 500 **were** recoverable through `efetch db=pmc` [E3]. So `inEPMC: Y` and a PMCID together do not imply readable text. The operational rule this beat adds: **a PMCID is readable only if a fetch returns a `<body>` element** - test the content, not the identifier.
The signature to recognise
esummary -> articleids contains pmc: PMC7566887 # looks fine
EPMC core -> pmcid PMC7566887, inEPMC Y # looks fine
EPMC -> /PMC7566887/fullTextXML -> HTTP 500 # ambiguous
NCBI -> efetch db=pmc id=7566887 -> 7476 chars,
ends '</article-meta></front></article>' # NO <body>
The last response is valid XML that parses cleanly and would pass any check testing for well-formedness, non-emptiness or HTTP 200. An agent that flattened it with itertext() would get a title, an author list and a journal name - enough text to look like a paper and none of the content. This is the same family as the empty-ArticleTitle and dropped-italics traps this project already documents: the output remains well-formed with the content removed.
The check that works is a single substring test for <body before parsing 2.
What is lost, and it is less than it looks
Only one graph entry depends on this citation, and it depends on it for a supporting rather than a load-bearing claim. The larger loss is that this paper is the natural corroboration for the founding menin paper recorded in ana-unstated-system-yokoyama-menin-unresolved - the two together would settle whether menin dependence is an N-terminal property shared across partners. Both are unreadable, by different mechanisms, and neither deposited anything. GEO records under the corresponding author's name exist but belong to breast-cancer studies from a later period and are not this work 3.
That is the shape of the beat's residual: the menin-interface question is not blocked by missing experiments, it is blocked by two papers behind paywalls.
| script | none - Europe PMC, NCBI efetch db=pmc, elink and GEO author search |
|---|---|
| status | run |
| result summary | Unresolved. A PMCID exists and inEPMC is Y, but Europe PMC fullTextXML returns HTTP 500 and NCBI efetch db=pmc returns front matter with no body element. No deposition. Adds an operational rule: test for a <body> element, not for a PMCID. |
| what would resolve it | Institutional subscription access to Cancer Res 2007, or a later publisher deposit of full text to PMC. Retry the fullTextXML endpoint in a future cycle. |
| resolution class | bounded failure |
| pmc body recoverable count | 5 |
| pmc body unrecoverable count | 4 |
inputs
[ "PMID 17671196", "PMC7566887" ]
routes tried
[ "PubMed esummary (pmc: PMC7566887)", "Europe PMC core (pmcid PMC7566887, inEPMC Y, isOpenAccess N)", "Europe PMC fullTextXML (HTTP 500, twice, cache cleared between)", "NCBI efetch db=pmc (7,476 chars, no <body>)", "elink pubmed->gds (0)", "esearch db=gds Caslini C[Author] (3 unrelated breast-cancer series)" ]
entries affected
[ "ana-menin-dependence-untested-in-septin-and-cytoskeletal-kmt2a-fusions" ]
pmc unrecoverable IDS
[ "PMC7566887", "PMC3593849", "PMC6234363", "PMC3795288" ]
- A citation that passes every identifier check and yields no content is the most dangerous kind in this toolchain, because verify-citations.py confirms the title matches while nobody has read a word of the paper.
- Two of the menin interface's foundational citations are unreadable by every free route, so the graph's confidence in partner-independent menin dependence is inherited rather than verified.
- An agent that flattens a front-matter-only PMC response would obtain plausible text and could write a summary of a paper it never read, with no checker in this toolchain able to catch it.
- Which fusion partners were used in the amino-terminal transformation experiments? Together with the founding menin paper this would settle whether menin dependence is partner-independent.
- Does Europe PMC's HTTP 500 for these four PMCIDs reflect a publisher metadata-only deposit, and would a later cycle's retry succeed?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The paper reports that interaction of MLL amino-terminal sequences with menin is required for transformation, and its abstract does not specify the fusion partners used.
Interaction of MLL amino terminal sequences with menin is required for transformation. · Caslini C, Yang Z, El-Osta M, Milne TA, Slany RK, Hess JL · Cancer Res 2007What this evidence cannot show
Identifiers taken from the NCBI esummary record fetched on 2026-09-19. No claim is made about the paper's content beyond its abstract; this entry deliberately does not guess the fusion roster.
Evidence that argues against this (1)
- If the requirement really is carried by MLL amino-terminal sequences, the partner roster is secondary by construction and this unreadable paper may matter less than its companion - but that reading cannot be checked without the text. Yokoyama A et al. 2005
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PMC7566887 returns HTTP 500 from Europe PMC fullTextXML on repeated attempts and 7,476 characters of front matter with no body element from NCBI efetch db=pmc, despite PubMed esummary listing the PMC identifier and Europe PMC reporting inEPMC Y.
Europe PMC fullTextXML and NCBI efetch db=pmc responses for PMC7566887 · Europe PMC and NCBI Entrez Programming Utilities 2026What this evidence cannot show
Run on 2026-09-19 from a single network location, with the Europe PMC attempt repeated after clearing a cached error response to ensure the retry was real. A 500 can be transient; this one was not across two attempts, but a later cycle should retry rather than treat the failure as permanent. Front-matter-only delivery usually reflects a publisher deposit of metadata without text, which can change if the publisher later deposits.
Evidence that argues against this (1)
- Five other papers in this beat returned the same Europe PMC HTTP 500 and were fully recoverable through NCBI efetch db=pmc, so this failure mode is not a property of the error code and must be tested per paper rather than assumed. Rogic S et al. 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
No deposition corresponds to this paper: elink from the PMID to GEO returns zero records, and a GEO author search on the corresponding author returns three human breast-cancer series from a later period that belong to different studies.
NCBI elink pubmed-to-gds for PMID 17671196 and esearch db=gds author query · NCBI Entrez Programming Utilities 2026What this evidence cannot show
Run on 2026-09-19. The author-name search is a weak instrument: it would miss a deposition submitted under a co-author or a laboratory-head name, and it produced only false positives here. A 2007 transformation-assay paper would not be expected to deposit sequencing in any case, so this negative carries little information.
Evidence that argues against this (1)
- A GEO author search returning unrelated studies by the same name is a reminder that contributor-name matching is as unreliable as gene-symbol matching, which this project has documented at length; neither a hit nor a miss from such a query should be trusted without checking the study. Souren NY et al. 2022
the second unreadable menin citation; PMCID exists but no body element is served
adds the rule that a PMCID is readable only if a fetch returns a body element
both menin foundational citations are unreadable, by different mechanisms
same PMC front-matter-only failure signature
bounded failure in the unstated-system worklist