Oncofusion

start / graph / gene-ebf3

Proteinweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q36-menin-interactome

EBF3 — a single large-scale MS hit against menin, in a B-lineage transcription factor family, which is the only reason it is worth recording

also called EBF3 · COE3 · O/E-2 · transcription factor COE3 · Q9H4W6

EBF3 is a relative of the master switch that makes a cell become a B cell. It showed up once, in a very large fishing expedition for protein partners, as something menin might stick to. One hit in one big screen is thin evidence, and this entry says so — but the family it belongs to is the family that decides B-cell identity, which is the one thing that makes it worth writing down at all.

The same thing, for a scientist

EBF3 (UniProt Q9H4W6, 596 aa, primary gene name read back, recommended name Transcription factor COE3) is a paralogue of EBF1, the transcription factor that specifies B-lymphoid identity. Its menin evidence is **one row, from one publication**: the Ewing/Figeys large-scale mass-spectrometry map of human protein-protein interactions, in which 338 FLAG-tagged bait proteins were immunoprecipitated and analysed by LC-ESI-MS/MS, yielding 24,540 potential interactions filtered to 6,463 by empirical criteria and a calculated interaction confidence score [E1]. IntAct curates the MEN1-EBF3 row from it as anti-bait co-immunoprecipitation [E2]. **That is the whole evidence base.** A single hit in a screen that reported thousands is, on its own, at the level of a hypothesis-generating observation, and this entry is graded `weak` accordingly. **The reason to record it rather than drop it** is the one thing that distinguishes it from the other single-paper rows: KMT2A::AFF1 is a disease of B-lymphoid identity, this graph's Fanconi signal is stronger in the lymphoid half, and menin is the protein the therapeutic hypothesis is about. A menin contact with an EBF-family factor is the kind of thing that would matter if it were true. **It has never been tested.** `EBF3` or `COE3` against `KMT2A`, `MLL` or `menin` returns 1 record and title-checking removes it [E3].

Go deeper
What a single large-scale hit is and is not worth

The 2007 Ewing map is a real and useful dataset and was the first large-scale human protein interaction study by mass spectrometry 1. It is also a screen: 338 baits, 24,540 raw candidate interactions, filtered to 6,463 by an empirical confidence score. Any individual row in such a dataset carries the screen's aggregate false-discovery properties, not the confidence of a directed experiment.

Three specific reasons to discount this row:

  • FLAG-tagged over-expressed bait. The standard AP-MS caveat, and transcription factors are common background in nuclear IPs.
  • No orthogonal confirmation in nineteen years. No pull-down, no reciprocal IP, no region mapping, no functional experiment.
  • Not recovered by later maps. BioPlex's dual proteome-scale network reports CD226 and ZC3HC1 for MEN1 and does not report EBF3 4.

The correct status is weak, and the correct entry is a short one that says where the row came from so that nobody upgrades it by repetition.

Why record it anyway: the lymphoid coincidence, stated as a coincidence

Three things in this graph point at B-lymphoid biology and none of them currently connects to menin's partner list:

  • the index fusion, KMT2A::AFF1, is overwhelmingly a B-lineage disease and the infant B-ALL case is the project's central one;
  • the unexplained Fanconi selectivity signal is stronger in the lymphoid half of the DepMap panel in two independent analyses (−0.308 against −0.334 in one, −0.70 against −0.48 in the other);
  • EBF-family factors are the determinants of B-cell specification and commitment.

That is a coincidence, not an argument, and it is written here as a coincidence. The honest statement is: the single cheapest experiment that could make EBF3 interesting has not been done — a menin co-IP in a KMT2A::AFF1 B-lineage line, blotted for EBF3 alongside EBF1. If it is negative the row can be retired; if positive, menin acquires a contact with the lineage machinery of the disease.

This entry makes no claim that menin binds EBF3 in a leukaemic B cell, and no claim that EBF1 is a menin partner — EBF1 does not appear in menin's IntAct record at all.

Facts and figures
uniprotQ9H4W6
length AA596
uniprot resolutiongene_exact:EBF3 AND organism_id:9606 AND reviewed:true; exactly one reviewed record; primary gene name read back as EBF3; length 596 aa checked against the entry on 2026-09-20
familyEBF/COE family; paralogue of EBF1, the B-lymphoid specification factor
ebf1 in menin intact recordfalse
menin contact
{
  "method": "anti-bait co-immunoprecipitation within a 338-bait FLAG-tag AP-MS screen",
  "source_pmid": "17353931",
  "n_source_publications_in_intact": 1,
  "orthogonal_confirmation": "none in nineteen years",
  "recovered_by_later_maps": false
}
literature
{
  "kmt2a_menin_records_raw": 1,
  "kmt2a_menin_records_title_checked": 0,
  "aliases_searched": [
    "EBF3",
    "COE3"
  ],
  "aliases_not_searched": [
    "O/E-2"
  ],
  "query": "(EBF3[tiab] OR COE3[tiab]) AND (KMT2A[tiab] OR MLL[tiab] OR menin[tiab])"
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.0887,
  "chronos_mean_other_leukaemia": -0.0984,
  "difference": 0.0097,
  "pan_essentiality_all_lines": -0.0556,
  "selectivity_rank": 10424,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
How this could bite us back
  • Nothing here supports an intervention. The entry's function is to stop a one-row screen hit becoming a menin partner through citation drift.
  • If menin did contact EBF-family factors, a menin inhibitor would be perturbing B-lineage identity machinery in a B-lineage leukaemia, which is the context in which this graph has already recorded lineage switch as a resistance mechanism.
Still open
  1. Does menin co-immunoprecipitate with EBF3, or with EBF1, in a KMT2A::AFF1 B-lineage line? One blot retires or promotes this entry.
  2. Is EBF3 expressed at all in KMT2A-rearranged infant B-ALL blasts? If not, the row is irrelevant regardless of whether the interaction is real.
  3. Should this graph adopt a standing rule for single-hit large-scale rows - record with the screen's parameters attached and grade weak, as here - so they neither vanish nor get upgraded by repetition?
Every source behind this entry (4)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The first large-scale study of human protein-protein interactions by mass spectrometry mapped interactions for 338 bait proteins selected for known or suspected disease and functional associations, using large-scale immunoprecipitation of FLAG-tagged baits followed by LC-ESI-MS/MS; 24,540 potential interactions were identified and filtered by empirical criteria and a calculated interaction confidence score to 6,463 interactions among 2,235 distinct proteins.

    Large-scale mapping of human protein-protein interactions by mass spectrometry. · Ewing RM, Chu P, Elisma F, Li H, Taylor P, Climie S, McBroom-Cerajewski L, Robinson MD, O'Connor L, Li M, Taylor R, Dharsee M, Ho Y, Heilbut A, Moore L, Zhang S, Ornatsky O, Bukhman YV, Ethier M, Sheng Y, Vasilescu J, Abu-Farha M, Lambert JP, Duewel HS, Stewart II, et al. · Mol Syst Biol 2007
    PMID 17353931DOI
    What this evidence cannot show

    A screen, not a directed study. FLAG-tagged over-expressed baits; the confidence score is empirical and calibrated on the dataset as a whole, so it bounds the aggregate false-discovery rate and not the reliability of any single row. The abstract's author list is truncated here with 'et al.' because the paper has more than twenty-five authors.

    Evidence that argues against this (1)
    • Later proteome-scale interaction maps using different tags, cell lines and statistical frameworks have not recovered MEN1-EBF3, which is what a false-positive row in a 2007 screen would look like. Huttlin EL et al. 2021
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    IntAct curates the MEN1-EBF3 interaction from PMID 17353931 alone, with the method anti-bait co-immunoprecipitation and interaction type 'physical association'; RPRM is the only other MEN1 partner sourced to that publication. No binding region is curated on either side.

    What this evidence cannot show

    'Anti-bait co-immunoprecipitation' here describes the screen's own method - a FLAG IP with MS read-out - and should not be read as a targeted co-IP with a menin antibody and an EBF3 blot, which is what that phrase usually implies in a directed paper.

    Evidence that argues against this (1)
    • A curated interaction inherits every limitation of the single experiment behind it. Anti-bait co-immunoprecipitation and affinity-capture mass spectrometry cannot distinguish a direct contact from a shared-complex association, tagged-bait overexpression can create contacts that do not occur at endogenous stoichiometry, and a partner curated from one publication has by definition never been independently replicated. The enumeration is accurate about the database; the database is not a statement about the blast. Heppner C et al. 2001
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    There is no EBF3 literature in KMT2A-rearranged leukaemia or in menin biology. A title/abstract search over EBF3 and COE3 against KMT2A, MLL or menin returns 1 record, a whole-genome sequencing study of mutation-negative Cornelia de Lange syndrome, which is removed by title-checking. Checked count: 0.

    What this evidence cannot show

    EBF3 has its own substantial literature in neurodevelopmental disease (HADDS) and as a candidate tumour suppressor; it is specifically unstudied in this intersection. The alias 'O/E-2' was not searched and is old enough that a 1990s paper using it could be missed.

    Evidence that argues against this (1)
    • A title-checked search count is two judgements stacked: the query decides what is reachable, and the reader decides what is 'on-beat'. Both are recorded here, which is the right practice, but neither is a measurement. This project has since established that field-scoped queries can return zero regardless of the literature, and that a symbol-level count misses work indexed under the pathway or the complex rather than the gene - so a small count bounds what was found, not what exists. Shan X et al. 2017
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    EBF3 is UniProt Q9H4W6, 596 aa, primary gene name read back as EBF3, recommended name Transcription factor COE3.

    What this evidence cannot show

    A database record. Resolved with gene_exact, which matches synonyms; exactly one reviewed human record returned, with the primary gene name and 596 aa length read back off the entry.

    Evidence that argues against this (1)
    • A reviewed database record fixes the canonical sequence and primary symbol on the date it was read, and nothing more. It does not establish that the protein is expressed in KMT2A-rearranged blasts, which isoform is present, or that the subunit acts in this disease. Accession and length are asserted so that a swap fails verify-accessions.py - a defence this graph needs because IKK-family and NF-kB subunit names collide heavily in the older literature - not because the record carries biological weight. Xiu Y et al. 2018
Connected to