Enumerating the WRAD-versus-KMT2A-fusion literature: what was searched, found, rejected and could not be read

Every count below is a live PubMed esearch total, not an estimate.

Subunit versus fusion, alias-expanded. Fusion term set in all four: (MLL-AF4 OR MLL-AF9 OR MLL-ENL OR MLL-AF10 OR MLL-AF6 OR MLL-ELL OR "MLL fusion" OR "MLL fusion protein" OR "KMT2A fusion" OR "MLL-rearranged" OR "KMT2A-rearranged").

Bare-symbol and retired-symbol counts, to bound the alias question in both directions: WDR5 611, ASH2L 216, RBBP5 165, DPY30 119, RBQ3 10, ASH2L1 2, DPY30L 1, HDCMD6P 0. RBQ3 AND (MLL OR KMT2A OR leukemia) returns 1, off topic. ("BIG-3" OR BIG3) AND (MLL OR KMT2A) returns 0.

Complex-level and method-level queries.

Europe PMC full-text-inclusive. "AF4-MLL" 292; "AF4-MLL" AND "WDR5" 29; "AF4-MLL" AND ("RBBP5" OR "ASH2L" OR "DPY30") 28; "MLL-AF9" AND ("RBBP5" OR "DPY30") 118; ("MLL-AF4" OR "MLL-AF9" OR "MLL-ENL" OR "MLL-AF10") AND ("co-immunoprecipitation" OR "coimmunoprecipitated") AND ("WDR5" OR "RBBP5" OR "ASH2L" OR "DPY30") 40. Titles checked on all five; the primary papers in them are already cited in this beat and none reports a WRAD-fusion association.

Read in full text: PMID 15199122 (via PMC HTML), 27462455, 23352661, 34489550 and the relevant sections of 39922194's abstract-level record. Abstracts read for a further 45.

A parenthesised karyotype string silently zeroed a query. (AF4-MLL OR "AF4-MLL" OR "AFF1-KMT2A" OR der(4)) AND leukemia returned 0 records with a clean No items found., while AF4-MLL AND leukemia alone returns 49. PubMed parsed der(4) as syntax. This graph's brief already warns about t(7;11)(p15;p15); der(4) is the same trap in a shorter string, and it is more dangerous because a reader scanning the query does not see a karyotype at all. Any query containing a bare der(n) should be assumed broken.

This beat's own counted absence was falsified by a paper already in the graph. The tiab proximity query returns 0 records for all four subunits against five proximity terms. The graph already holds PMID 39922194, a BioID study of the NUP98-PHF23 oncofusion whose data recover WDR5 and RBBP5 - and whose abstract says 'the KMT2/MLL H3K4 methyltransferases'. The absence claim is true about titles and abstracts and false about the literature. Recording it is the point: a counted absence in this field measures indexing, and complex-level framing is the systematic leak.

Recorded because a reader would reasonably expect them cited.

1. The WDR5-MLL1 medicinal chemistry series (PMIDs 41270727, 41114172, 39804067, 34105966, 31858797, 31768400, 30013062, 29857126, 27598236, 27116709, 22989411, 35525905, 35252869, 31923859). Fourteen papers whose titles pair WDR5 with MLL1 or with MLL-rearranged leukaemia. All optimise binders of the wild-type WIN interface; none assays a fusion protein. Overturned by: any of them reporting engagement of a fusion-associated WDR5 pool, or a co-IP arm.
2. PMID 21445285, NF-Y recruits Ash2L to impart H3K4 trimethylation on CCAAT promoters. Title is squarely on-beat by symbol; content is CCAAT-box promoter biology with no KMT2A fusion. Overturned by: evidence that CCAAT-box recruitment operates at fusion target genes.
3. PMID 20581860, MOZ/MLL crosstalk in cord blood CD34+ cells. WDR5 appears; the biology is wild-type MLL and KAT6A. Overturned by: a fusion arm.
4. PMID 30488017 (WDR5 review) and PMID 41864982 (ASH2L-DPY30 review, 2026). Used as maps to primaries only, per this project's rule. The 2026 review is the most current synthesis of the ASH2L-DPY30 interface and contains no KMT2A-fusion data.
5. PMID 27192115, WDR5 high expression in leukaemia. Real leukaemia data, not stratified by KMT2A status, so it cannot speak to fusion specificity. Overturned by: a KMT2A-rearranged versus wild-type split of the same cohort.
6. PMID 30679799, CEBPA-mutated leukaemia and the MLL1 complex. A genuine MLL1-complex dependency paper, framed entirely at complex level, in a non-KMT2A-rearranged disease. Retained here as the worked example of the complex-name search trap rather than as evidence.
7. PMID 35036869 (SETBP1 recruits the MLL1 complex) and PMID 37535603 (SETD1A-BuGZ/BUB3). Both concern WRAD-containing complexes in leukaemia with no KMT2A fusion. Overturned by: the same recruitment being shown for a KMT2A fusion.
8. PMID 26990989, SON controls MLL complex assembly via menin. About complex assembly and menin, not WRAD-fusion contact.
9. PMID 33924850, nuclear FGFR2 interacts with MLL-AF4. A real post-2004 KMT2A::AFF1 interactor paper and a reminder that the t(4;11) interactome is not closed - but not a WRAD result.
10. PMID 39002676, MLL-AF9 regulates transcriptional initiation. No WRAD data, but it directly contradicts PMID 27462455 on whether MLL1 and MLL-AF9 occupy the same chromatin. Recorded as an unresolved dispute rather than a rejection, because it bears on how the genome-wide argument for the negative should be read.
11. PMID 32733069, restoring MLL and the H2Bub-ASH2L-MLL axis. ASH2L in an MLL-AF4 xenograft context; the mechanism concerns wild-type MLL chromatin accumulation, not fusion association. Overturned by: an arm showing the axis acts on the fusion.
12. Yeast and worm orthologue literature (Bre2, Swd1/2/3, Sdc1, dpy-30). Not searched against fusion terms because no KMT2A fusion exists in those organisms. Searched only where it bore on module architecture.
13. ASH2 alone as a search term. Returns predominantly Drosophila and yeast records; excluded on precision grounds after title-checking.
14. The permissive ASH2L query (ASH2L2 OR "ASH2 like" OR Ash2l) AND (MLL OR KMT2A) AND (leukemia OR leukaemia), 61 records. Recorded and NOT counted: MLL matches every mixed-lineage-leukaemia paper and Ash2l duplicates the current symbol. This is the precision half of the alias rule.
15. DepMap re-analysis. Not re-run. Four prior designs in this graph already address WRAD subunit dependencies and their disagreements are recorded; adding a fifth would not answer an association question.

PMID 21030982, Benedikt et al. 2011, Leukemia — 'The leukemogenic AF4-MLL fusion protein causes P-TEFb kinase activation and altered epigenetic signatures'. This is the only published affinity purification of a SET-domain-retaining KMT2A fusion complex, and its abstract states the complex has at least sixteen constituents including an H3K4(me3) methyltransferase. Whether any WRAD subunit is among them would answer this beat's question directly. Routes tried: Europe PMC core record returns no PMCID; fullTextXML is therefore unavailable; efetch db=pmc has no record; nature.com redirects to an authentication endpoint. Abstract only. Confidence on anything from it is weak and this beat does not assert what the sixteen constituents are. Resolving it needs institutional access to the supplementary tables, which is a one-hour task for anyone with a subscription and is the single highest-value unresolved item in this beat.

A route note, not an unreadable. Yokoyama 2004 (PMID 15199122, PMC480881) returns HTTP 500 from fullTextXML, and efetch db=pmc returns a well-formed record carrying the comment 'The publisher of this article does not allow downloading of the full text in XML form' with no <body> element. The PMC HTML page serves the complete text, 68,159 characters after tag-stripping. Under this project's rule the paper would have been logged unreadable after two failures; the HTML route is a third route and it worked, and the entire adjudication in this beat rests on it.