Wild-type KMT2A is one of the most promiscuously labelled proteins in the proximity record - 25 of 50 baits - and menin is one of the least, at 3
also called KMT2A BioID promiscuity · MEN1 proximity recovery
Counting how many tagging experiments find each protein nearby, KMT2A turns up in half of them and menin in almost none. That gap is either a real difference in how the two proteins sit in the nucleus, or just a difference in how much of each there is.
The same thing, for a scientist
In the K562 50-bait BioID2 map, **KMT2A is recovered by 25 of 50 baits (97.6th percentile of the baits-per-prey distribution)** while **MEN1 is recovered by 3 (34.4th percentile)** [E1]. Across the whole IntAct proximity record KMT2A carries **46 rows in five publications** and MEN1 four [E2]. Two structural readings compete and the data do not separate them: KMT2A is an abundant, very large chromatin protein that any nuclear bait's labelling radius will reach, or KMT2A genuinely occupies a shared compartment with RNA processing while menin does not. What the numbers **do** settle is that the complex-membership picture this graph inherits from affinity purification does **not** reproduce as a block: **WDR5 11 and ASH2L 9 but RBBP5 0 and DPY30 0**; **CDC73 25 and RTF1 15 but PAF1 6, CTR9 3 and LEO1 0** [E1].
The counts
KMT2A's 25 baits: AGGF1, BUD13, DDX52, DDX55, EFTUD2, EXOSC5, GPKOW, HNRNPA1, HNRNPC, ILF3, LIN28B, NPM1, PPIL4, QKI, RBM15, RPS11, SBDS, SF3B4, SMNDC1, U2AF2, YWHAG, ZC3H11A, ZC3H8, ZNF800, ZRANB2 1.
MEN1's three: BUD13, NPM1, RPS11 1. All three also reach KMT2A. There is no bait that reaches menin and not KMT2A.
For scale in the same dataset: HCFC1 34, CHD8 29, CDC73 25, KMT2A 25, SETD1A 17, RTF1 15, ARID1A 13, WDR5 11, KAT7 10, ASH2L 9, PSIP1 9, ICE1 9, ICE2 8, KMT2D 8, PAF1 6, CBX8 5, BRD4 5, AFF1 4, DOT1L 4, CTR9 3, MEN1 3 - and zero for RBBP5, DPY30, LEO1, SUPT4H1, EP300, CREBBP, MYB, MYC, RUNX1's partners SPI1 and CEBPA, HOXA9, MEIS1, PCGF1 and KMT2B 1.
Complexes do not come back as complexes
This is the most transferable thing in the entry, and it is a warning about using proximity data for composition.
WRAD: WDR5 11, ASH2L 9, RBBP5 0, DPY30 0. A four-subunit obligate module splits 11/9/0/0 across the same fifty baits in the same cells 1.
PAF1 complex: CDC73 25, RTF1 15, PAF1 6, CTR9 3, LEO1 0. A 25-to-0 spread across one complex 1.
SEC: CCNT1 16, CDK9 11, ELL 10, MLLT1 7, AFF1 4, AFF4 4, ELL2 0, EAF1 0, EAF2 0.
If proximity recovery tracked complex membership, these numbers would cluster. They do not, and the spread is far larger than any plausible stoichiometry difference. Recovery tracks size, abundance and lysine accessibility of the prey, and the correct use of a proximity count is as a detectability measure, not as an inclusion criterion.
Which way does the KMT2A/menin gap cut for the menin-interface hypothesis?
For it, weakly. Menin's narrow profile is consistent with a protein whose nuclear residence is restricted rather than diffuse, and all three of its baits also reach KMT2A - i.e. every place menin was seen, KMT2A was too. Nothing here shows menin somewhere KMT2A is not.
Against it, weakly. A protein recovered by 3 of 50 baits is a protein near the detection floor, and a target whose own neighbourhood is that sparse gives no proximity-based handle for pharmacodynamics. And the 25-to-3 gap means that if the fusion's oncogenic neighbourhood were mapped, menin would be a minor feature of it by count - which is exactly the shape of a mechanism where one small contact carries a large effect, and equally the shape of an artefact.
The honest statement is that this dataset cannot adjudicate a therapeutic hypothesis, and it is recorded so nobody quotes the 25-versus-3 in either direction without the caveat.
| script | none (database query) |
|---|---|
| status | run |
| kmt2a n baits | 25 |
| kmt2a percentile | 97.6 |
| men1 n baits | 3 |
| men1 percentile | 34.4 |
| result summary | KMT2A 25/50 baits, MEN1 3/50, and obligate complexes split 11/9/0/0 and 25/15/6/3/0. Proximity recovery is a detectability measure, not a membership criterion. |
inputs
[ "IntAct PSICQUIC pubid:39251607", "IntAct PSICQUIC detmethod:\"MI:1314\" per accession" ]
men1 baits
[ "BUD13", "NPM1", "RPS11" ]
kmt2a baits
[ "AGGF1", "BUD13", "DDX52", "DDX55", "EFTUD2", "EXOSC5", "GPKOW", "HNRNPA1", "HNRNPC", "ILF3", "LIN28B", "NPM1", "PPIL4", "QKI", "RBM15", "RPS11", "SBDS", "SF3B4", "SMNDC1", "U2AF2", "YWHAG", "ZC3H11A", "ZC3H8", "ZNF800", "ZRANB2" ]
complex split wrad — the numbers behind the chart
{
"WDR5": 11,
"ASH2L": 9,
"RBBP5": 0,
"DPY30": 0
}complex split paf1c — the numbers behind the chart
{
"CDC73": 25,
"RTF1": 15,
"PAF1": 6,
"CTR9": 3,
"LEO1": 0
}kmt2a rows by publication — the numbers behind the chart
{
"26638075": 2,
"35140242": 11,
"35384245": 1,
"36931259": 7,
"39251607": 25
}- A 25-versus-3 ratio is the kind of number that gets quoted as if menin were peripheral. It is a detectability ratio and nothing more.
- If a future proximity map of the fusion shows menin as a minor feature by count, that will be misread as evidence against the menin interface. State in advance that count is not importance.
- Normalised for protein copy number in K562, is KMT2A still unusually promiscuous and menin still unusually sparse?
- RBBP5, DPY30 and LEO1 are recovered by zero baits while their obligate partners are recovered by 9-25. Is that lysine accessibility, size, or genuinely different neighbourhoods?
- Every bait reaching menin also reaches KMT2A. Is there any proximity experiment anywhere in which menin appears without KMT2A?
- Does a menin bait exist in any unpublished dataset? It is the single most valuable missing construct for this project.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
In the K562 50-bait BioID2 deposition, KMT2A is recovered as prey by 25 baits and MEN1 by 3 (BUD13, NPM1, RPS11, all of which also reach KMT2A); within nominally obligate complexes recovery does not cluster - WDR5 11, ASH2L 9, RBBP5 0, DPY30 0; CDC73 25, RTF1 15, PAF1 6, CTR9 3, LEO1 0; CCNT1 16, CDK9 11, ELL 10, MLLT1 7, AFF1 4, AFF4 4, ELL2 0, EAF1 0, EAF2 0.
IntAct PSICQUIC query: pubid:39251607, full 19,403-row deposition (IMEx IM-30059), retrieved 2026-09-19 · IntAct molecular interaction database (EMBL-EBI) 2026What this evidence cannot show
Retrieved from IntAct 2026-09-19 over the authors' curated bona fide pair list. Recovery is abundance- and size-weighted and every zero is a zero after filtering, so a small count is not a claim about a protein's real neighbourhood. K562 has no KMT2A rearrangement and the labelled KMT2A is wild type. No expression normalisation was applied, which is the analysis's main weakness.
Evidence that argues against this (2)
- In the largest human BioID survey (192 baits, HEK293) a prey's bait profile is dominated by subcellular compartment rather than binding, so an abundant chromatin protein recovered by many nuclear baits is the expected background of the method. Go CD et al. 2021
- BioID and affinity purification of chromatin-associated complexes recover few shared partners and enrich for different processes, so proximity counts cannot be read as complex membership. Lambert JP et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
Across the whole IntAct proximity record KMT2A carries 46 rows drawn from five publications - the 50-RBP K562 map (25), a human transcription-factor BioID network (11), a 14-3-3 proximity study (7), a centrosome-cilium interaction landscape (2) and a receptor-tyrosine-kinase interactome atlas (1) - and in every one of them KMT2A is the prey.
IntAct PSICQUIC query: detmethod:"MI:1314" (proximity-dependent biotin identification) intersected with the UniProt accessions of the super elongation complex, KMT2A, MEN1 and DOT1L, retrieved 2026-09-19 · IntAct molecular interaction database (EMBL-EBI) 2026What this evidence cannot show
IntAct is one database with one curation policy and does not serve every proximity resource. The five publications have nothing in common methodologically except the detection method, so pooling their counts mixes cell lines, tag chemistries and filtering thresholds.
Evidence that argues against this (1)
- Four of the five publications had no interest in KMT2A; it appears in them as an incidental prey of baits chosen for centrosomes, receptor kinases, transcription factors and 14-3-3 proteins. A pooled count over such a set measures how often KMT2A shows up in other people's experiments, which is a statement about its detectability. Go CD et al. 2021
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Oncogenic mutant forms of MLL retain an ability to interact with menin but not with other identified complex components.
Leukemia proto-oncoprotein MLL forms a SET1-like histone methyltransferase complex with menin to regulate Hox gene expression. · Yokoyama A, et al. · Mol Cell Biol 2004What this evidence cannot show
2004 affinity purification. Cited to frame why a wild-type KMT2A proximity profile cannot be transferred to the fusion.
Evidence that argues against this (1)
- This graph's own reading of that paper is that it is equally a statement about everything except menin, and the same logic applies here: the 25 baits that reach wild-type KMT2A are reaching a protein the fusion largely is not. Yokoyama A et al. 2004
the census behind 7 of the 12 RNA-processing failures; its 25-of-50 promiscuity confound is why those partners should not be promoted individually
reproduces its 19,403 rows, 50 baits, KMT2A 25 and MEN1 3 exactly, then separates the two readings it declined to: length-matched, KMT2A falls from the 97th to the 90.7th percentile
- related to →KMT2A / MLL1 — the gene that breaks
wild-type KMT2A recovered by 25 of 50 baits
menin recovered by 3 of 50 baits, all of which also reach KMT2A
- related to →WDR5 — the partner that is not a partner: it binds the half of KMT2A the fusion throws away
WRAD does not recover as a block: WDR5 11, ASH2L 9, RBBP5 0, DPY30 0
PAF1 complex does not recover as a block: CDC73 25, RTF1 15, PAF1 6, CTR9 3, LEO1 0
principal data source
- ← supportsA BioID hit is a neighbour, not a complex member - and the two methods are known to disagree
counts measure detectability, not membership
same K562 BioID2 census; this analysis adds the PSIP1 side, 9 of 50 baits at the 68th percentile of preys, which that entry did not compute
the promiscuity finding is the standing deflationary reading of any single KMT2A prey row, including this one
- ← related toYY1 - the enhancer-promoter looping factor that recovers KMT2A in a 109-transcription-factor BioID screen
same deflationary control applies to the YY1 row
re-derives the 25 of 50 and adds the abundance term the source analysis declared missing
revives the shared-compartment reading that entry named and could not separate
The observed WRAD split across one 50-bait deposition - WDR5 11, ASH2L 9, RBBP5 0, DPY30 0 - is what abundance-weighted second-shell recovery of an obligate complex looks like.