Oncofusion

start / graph / ana-tf-bioid-network-reaches-kmt2a-and-af10

Analysismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q20-proximity-interactome

A 109-transcription-factor BioID network puts KMT2A next to MYC, YY1 and the ETS and KLF families - and AF10 next to ERG and KLF5

also called Göös TF BioID · TF interaction network KMT2A

A 2022 study tagged 109 human transcription factors and listed everything each one touched. KMT2A turned up beside MYC, YY1 and several other well-known factors; AF10, a KMT2A fusion partner, turned up beside ERG and KLF5.

The same thing, for a scientist

Göös et al. mapped 6,703 BioID and 1,536 AP-MS interactions for **109 human transcription factors**, and state that the BioID arm yielded **more high-confidence interactions, highlighting the transient and dynamic nature of many TF interactions** [E1]. In the IntAct deposition, **KMT2A is recovered as prey by eleven TF baits: EN1, NHLH1, ELF1, ELF2, FEV, KLF12, KLF3, YY1, MYC, NFIX and NFYC** - the MYC row carrying an elevated MI-score of 0.48 - while **MLLT10/AF10 is recovered by KLF5, ERG and FEV**, **DOT1L by GATA3 and TBXT**, and **MLLT1/ENL by SP7** [E2]. This is a second independent proximity dataset containing this graph's subject matter, reached through the method code and invisible to every gene-symbol query.

Go deeper
Why this matters more than the counts suggest

AF10/MLLT10 is recovered by zero baits in the 50-RBP K562 map and by three transcription-factor baits here 2. A protein this graph treats as mechanistically central - the MLLT10-directed endogenous H3K79 unit is one of the project's load-bearing findings - has its entire proximity record in a study about transcription factors.

The pairing is also interesting on its own terms. ERG and KLF5 are haematopoietic transcription factors; ERG is a driver in several leukaemias. AF10 sitting in their labelling radius is the kind of observation that could nominate a lineage-specific recruiter, and nobody has followed it because nobody looking for AF10 biology would run this query.

MYC and YY1

The MYC-KMT2A row carries MI-score 0.48 rather than the 0.27 that a single-method single-publication proximity record scores, which means IntAct holds corroborating evidence from elsewhere for that pair 2. This entry does not claim to have read that corroborating evidence and the score is reported as a pointer, not as support.

YY1 and the ETS members ELF1/ELF2/FEV are all sequence-specific factors with established roles at active promoters. KMT2A being in their radius is unsurprising for an abundant promoter-associated protein, and is the compartment explanation again 3.

Retained region

Retained region: not determinable from this data. The SEC proteins here were preys, never baits, so what was labelled is the endogenous full-length wild-type protein. Biotinylation happens at accessible lysines anywhere on the prey and no site-level biotinylation was reported, so nothing distinguishes a contact with the AFF1 N-terminus that KMT2A::AFF1 discards from one with the C-terminal half it keeps.

Facts and figures
scriptnone (database query)
statusrun
pmid35140242
n tf baits109
n bioid interactions6703
n apms interactions1536
result summaryA second independent proximity dataset containing this graph's subject matter, reachable only by method code. AF10's entire proximity record lives here.
titleHuman transcription factor protein interaction networks.
authorsGöös H, Kinnunen M, Salokas K, et al.
venueNat Commun
year2022
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.
inputs
[
  "IntAct PSICQUIC detmethod:\"MI:1314\" per accession"
]
kmt2a tf baits
[
  "EN1",
  "NHLH1",
  "ELF1",
  "ELF2",
  "FEV",
  "KLF12",
  "KLF3",
  "YY1",
  "MYC",
  "NFIX",
  "NFYC"
]
mllt10 tf baits
[
  "KLF5",
  "ERG",
  "FEV"
]
dot1l tf baits
[
  "GATA3",
  "TBXT"
]
mllt1 tf baits
[
  "SP7"
]
How this could bite us back
  • Eleven transcription factors near KMT2A reads like a regulatory network and is more likely a promoter-occupancy census of an abundant protein.
  • AF10's entire proximity record is three rows in a study about something else. Any mechanism built on it is built on three rows.
Still open
  1. What cell line did the BioID arm of this study use? Nothing in the abstract or the MITAB rows examined says, and every conclusion depends on it.
  2. AF10/MLLT10 near ERG and KLF5 - is that a lineage-specific recruiter worth testing, or three incidental hits among 6,703?
  3. What is the corroborating evidence behind the MYC-KMT2A MI-score of 0.48, and does it predate or postdate this study?
  4. Does AF9/MLLT3 appear anywhere in this study's full deposition? IntAct serves zero proximity rows for it from any publication.
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Göös et al. identified 6,703 protein-protein interactions for 109 human transcription factors by proximity-dependent biotinylation and 1,536 by affinity purification mass spectrometry, and report that the BioID analysis identified more high-confidence interactions, which they attribute to the transient and dynamic nature of many TF interactions.

    Human transcription factor protein interaction networks. · Göös H, et al. · Nat Commun 2022
    PMID 35140242DOI
    What this evidence cannot show

    Abstract read; full text not read. The abstract does not name the cell line used for the BioID arm, and no attempt was made here to resolve it from the study's depositions. KMT2A, MLLT10, MLLT1 and DOT1L are not discussed in the abstract - they are reachable only through the IntAct deposition.

    Evidence that argues against this (2)
    • BioID and affinity purification of chromatin-associated complexes recover few shared partners and enrich for different processes, so a proximity hit list is not a complex roster. Lambert JP et al. 2015
    • More high-confidence interactions from BioID than from AP-MS can also reflect that the two methods have different false-positive structures rather than that BioID captured genuine transient contacts; the same laboratory's comparison of the two approaches on chromatin complexes found little overlap and distinct process enrichment in both directions. Lambert JP et al. 2015
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    In the IntAct deposition from this study, KMT2A is recovered as prey by eleven transcription-factor baits - EN1, NHLH1, ELF1, ELF2, FEV, KLF12, KLF3, YY1, MYC, NFIX and NFYC, the MYC row carrying MI-score 0.48 - while MLLT10 is recovered by KLF5, ERG and FEV, DOT1L by GATA3 and TBXT, and MLLT1 by SP7.

    What this evidence cannot show

    Retrieved 2026-09-19. Every row has the KMT2A-complex protein as prey; no such protein was a bait. The elevated MI-score on the MYC row indicates corroborating evidence in IntAct from another source, which this entry did not read. The cell line is not stated in the MITAB rows examined.

    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 by binding, so a splicing bait recovering co-transcriptional machinery is the expected background of the method. Go CD et al. 2021
    • KMT2A is recovered by 25 of 50 baits in an unrelated RBP-focused map, placing it at the 97.6th percentile of prey promiscuity. A protein that appears in half of one study's neighbourhoods appearing in eleven of another's is consistent with detectability rather than with eleven specific relationships. IntAct molecular interaction database (EMBL-EBI) 2026
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Proximity recovery in the largest published human BioID survey is organised by subcellular compartment, with a prey's bait profile dominated by localisation rather than by binding.

    PMID 34079125DOI
    What this evidence cannot show

    HEK293, 192 baits, and the survey's purpose was explicitly to define compartments, so it is the best available statement of the compartment effect and also the study most likely to emphasise it.

    Evidence that argues against this (1)
    • The same survey recovers genuine complex structure within compartments, so compartment dominance does not mean proximity data carry no interaction information - only that compartment must be subtracted first, which this beat could not do. Go CD et al. 2021
Connected to