Oncofusion

start / graph / ana-menin-neoppi-trap-ten-intact-partners-from-one-mutation-screen

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q36-menin-interactome

Ten of menin's 33 IntAct partners come from one BRET screen for MUTATION-created interactions — and five are gains caused by a named allele

also called menin neoPPI trap · 35512704 menin rows · IntAct MEN1 mutation-directed rows

A database says menin binds BRAF, PTEN, TP53, EGFR and AKT1, and it does not: those rows all come from one screen that deliberately looked for interactions created by cancer mutations. Writing them in as menin partners would have put five false entries into the graph at once.

The same thing, for a scientist

A full MITAB 2.7 pull of IntAct on MEN1 (O00255) returns 59 rows covering **33 distinct partners**. **Thirty-one of the 59 rows — and 10 of the 33 partners — come from a single publication, PMID 35512704**, *Systematic discovery of mutation-directed neo-protein-protein interactions in cancer* (Mo, Niu, Ivanov et al., Cell 2022): a BRET screen of 17,792 interaction pairs designed to find associations that a cancer **mutation creates** and the wild-type protein does not make [E1]. IntAct curates the allele on every one of those rows, and the annotation splits them cleanly [E2]. **Five are gain-of-interaction, annotated `mutation causing an interaction`**, and the alleles resolve to **BRAF V600E, AKT1 E17K, TP53 R282W, EGFR L858R and PTEN R173** [E3]. BRAF V600E is the only one with orthogonal validation in that paper — BRET plus a GST pull-down annotated `mutation increasing interaction strength` plus an anti-tag co-immunoprecipitation. **Five more are annotated `mutation with no effect`** — SPOP, SMAD4, SMARCA4, FBXW7 and PTPN11 — meaning the variant tested did not change a signal detected in a tagged over-expression BRET assay in live cells; no orthogonal method, no co-IP, no mapped region [E2]. **None of the ten should be written as a menin partner.** The five neo-interactions are properties of a mutant allele, and the five null-effect rows are single-method over-expression proximity measurements. **This corrects the dispatch brief in three places** (see the correction section).

Go deeper
The full split of menin's IntAct rows
partnersource PMIDmethod(s)what it is
KMT2A15199122, 18598942, 20153263anti-bait co-IP x3the real one
PSIP1 (LEDGF)18598942anti-bait co-IPthe real one
FANCD212874027anti-tag co-IP x2, anti-bait co-IP, MSinterface-menin-fancd2
NFKB111526476anti-bait co-IP, pull-downgene-nfkb1
NFKB2, RELA11526476pull-downgene-nfkb2, gene-rela
MYH914508515co-IP, cosedimentation, colocalisationgene-myh9 - three methods, one paper
HCFC215199122molecular sievinggene-hcfc2 - gel filtration co-elution, not a binding assay
EBF3, RPRM17353931anti-bait co-IPlarge-scale MS map
CSNK2A2, WDR527705803tandem affinity purificationPolycomb complexome AP-MS
ESR131527615tandem affinity purificationalready written
CD226, ZC3HC133961781anti-tag co-IPBioPlex
AFG2A38554706tandem affinity purificationreplisome proteostasis study
BUD13, NPM1, RPS1139251607BioIDproximity, not binding
H3C129568061BioIDMAC-tag proximity
BRAF, AKT1, TP53, EGFR, PTEN35512704BRET, allele-directedneo-interactions - a mutant allele's property
SPOP, SMAD4, SMARCA4, FBXW7, PTPN1135512704BRET, mutation with no effectsingle-method over-expression proximity
HOXA7, HOXA9, MEIS118598942chromatin immunoprecipitationthese are genes, curated as ChIP targets, not protein partners

The last row is a second, quieter trap in the same dataset: a PSICQUIC pull returns HOXA7, HOXA9 and MEIS1 as MEN1 'interactors' because the curation includes ChIP associations. They are menin's target loci, not its binding partners, and a script that reads the interactor column without reading the method column would enumerate them as proteins menin binds.

The five alleles, resolved

Each IntAct feature annotation carries either a dbSNP rs number or a ClinVar variation id. Both were resolved against the source database on 2026-09-20 3.

partnerIntAct featureidentifierresolved allele
BRAFmutation causing an interaction: 600-600ClinVar 13961NM_004333.6(BRAF):c.1799T>A (p.Val600Glu)
AKT1mutation causing an interaction: 17-17ClinVar 13983NM_001382430.1(AKT1):c.49G>A (p.Glu17Lys)
TP53mutation causing an interaction: 282-282ClinVar 12364NM_000546.6(TP53):c.844C>T (p.Arg282Trp)
EGFRmutation causing an interaction: 858-858dbSNP rs121434568EGFR L858, substitutions to R/Q reported
PTENmutation causing an interaction: 173-173dbSNP rs121913293PTEN R173, substitutions to C/G/S reported

EGFR and PTEN are given at residue rather than substitution because the rs record covers several substitutions at the codon and the screen's specific allele could not be read off the IntAct row. Naming the residue and stopping there is the honest resolution; asserting L858R and R173C would be recall dressed as a lookup.

BRAF V600E is different from the other four and deserves its own note. In the same paper it carries a GST pull-down annotated mutation increasing interaction strength and an anti-tag co-immunoprecipitation with the FLAG-tagged V600E construct. So the BRAF V600E-menin association is the only one of the ten with more than one method behind it. It is still a mutant-allele property, still in an over-expression system, and still irrelevant to KMT2A-rearranged leukaemia, where BRAF V600E does not occur.

Corrections to the dispatch brief

Recorded because the brief asked for premises to be checked against the literature, and three of its statements about this dataset are wrong in the direction of under-warning.

  1. "Six of the menin 'partners' above come from it alone." It is ten: SMAD4, SMARCA4, SPOP, FBXW7, PTEN and BRAF as the brief lists, plus AKT1, EGFR, TP53 and PTPN11.
  2. AKT1 and EGFR are filed in the brief under "MYH9, CD226, RPRM, ZC3HC1, AKT1, EGFR — various, single-paper; low-evidence". They are not various and not merely low-evidence: both are 35512704 rows and both are annotated mutation causing an interaction, i.e. they are exactly the class the brief's own trap warning exists to catch. Filing them outside the warning is the most dangerous of the three errors.
  3. PTPN11 and TP53 are listed under "partially covered under a different filing", pointing at comut-ptpn11-kmt2a-r and resistance-tp53-apoptotic-priming-menin. Those entries concern comutation frequency and apoptotic priming; neither has anything to do with a physical menin interaction, and the IntAct rows that prompted the listing are 35512704 BRET rows. Reading the existing entries as "partial coverage" of a menin-PTPN11 or menin-TP53 interaction would import a false premise.

None of this is a criticism of the brief's instinct, which was right and which is the reason no false partner was written. It is the enumeration that was incomplete.

Facts and figures
scriptmanual MITAB 2.7 pull and parse; no committed script
statusrun
result summary59 IntAct rows, 33 distinct MEN1 partners. 31 rows and 10 partners come from one mutation-directed BRET screen (PMID 35512704). Five are gain-of-interaction alleles: BRAF V600E, AKT1 E17K, TP53 R282W, EGFR L858, PTEN R173. Five are 'mutation with no effect' single-method over-expression rows: SPOP, SMAD4, SMARCA4, FBXW7, PTPN11. None was written as a menin partner. Three transcription factors (HOXA7, HOXA9, MEIS1) appear as interactors via ChIP curation and are target genes, not partners.
neoppi alleles
AKT117.0AKT1: 17PTEN173.0PTEN: 173TP53282.0TP53: 282BRAF600.0BRAF: 600EGFR858.0EGFR: 858
inputs
[
  "https://www.ebi.ac.uk/Tools/webservices/psicquic/intact/webservices/current/search/interactor/O00255?format=tab27",
  "NCBI ClinVar esummary",
  "NCBI dbSNP Variation Services refsnp"
]
neoppi alleles — the numbers behind the chart
[
  {
    "partner": "BRAF",
    "residue": 600,
    "identifier": "ClinVar 13961",
    "allele": "p.Val600Glu",
    "orthogonal_validation": "GST pull-down (mutation increasing interaction strength) and anti-tag co-immunoprecipitation, both in the same paper"
  },
  {
    "partner": "AKT1",
    "residue": 17,
    "identifier": "ClinVar 13983",
    "allele": "p.Glu17Lys",
    "orthogonal_validation": "none"
  },
  {
    "partner": "TP53",
    "residue": 282,
    "identifier": "ClinVar 12364",
    "allele": "p.Arg282Trp",
    "orthogonal_validation": "none"
  },
  {
    "partner": "EGFR",
    "residue": 858,
    "identifier": "rs121434568",
    "allele": "residue 858 only; substitution not resolvable from the IntAct row",
    "orthogonal_validation": "none"
  },
  {
    "partner": "PTEN",
    "residue": 173,
    "identifier": "rs121913293",
    "allele": "residue 173 only; substitution not resolvable from the IntAct row",
    "orthogonal_validation": "none"
  }
]
mutation with no effect partners
[
  "SPOP",
  "SMAD4",
  "SMARCA4",
  "FBXW7",
  "PTPN11"
]
not partners at all
[
  "HOXA7",
  "HOXA9",
  "MEIS1"
]
brief corrections
[
  "the 35512704 group is ten partners, not six",
  "AKT1 and EGFR are 35512704 neo-interaction rows, not 'various, single-paper' rows",
  "the PTPN11 and TP53 rows are 35512704 BRET rows and are unrelated to comut-ptpn11-kmt2a-r and resistance-tp53-apoptotic-priming-menin"
]
How this could bite us back
  • Writing the five neo-interactions as menin partners would have created a false rationale for combining menin inhibitors with BRAF, EGFR, AKT or PI3K inhibitors, in each case built on a mutation the patient population does not carry.
  • The 'mutation with no effect' rows are the more insidious half: they look like ordinary wild-type interactions and carry no warning label, and five of menin's 33 partners are of that kind.
Still open
  1. Does the 35512704 supplementary table state which specific EGFR and PTEN substitutions were screened against menin? Parsing it would close the two residue-level entries here.
  2. Is BRAF V600E-menin a real complex worth anything to anyone? It does not occur in KMT2A-rearranged leukaemia, but it does occur in hairy cell leukaemia and in Langerhans cell histiocytosis, where menin inhibitors are not used and perhaps should be asked about.
  3. How many other baits in this graph have IntAct partner lists contaminated by 35512704 rows? The screen covered 17,792 pairs, so menin will not be the only one.
  4. Should the project's interactome tooling filter IntAct rows on the method column by default, so ChIP-curated target genes never enter a partner enumeration again?
Every source behind this entry (4)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    PMID 35512704 is a quantitative high-throughput differential screening platform coupling BRET biosensors with miniaturised co-expression in live cells, designed to identify variant-enabled neomorph protein-protein interactions - interactions the mutant protein makes and the parental molecule does not. It screened 17,792 interactions with 2,172,864 data points, and its headline result is that BRAF V600E mediates a KEAP1 neoPPI.

    Systematic discovery of mutation-directed neo-protein-protein interactions in cancer. · Mo X, Niu Q, Ivanov AA, Tsang YH, Tang C, Shu C, Li Q, Qian K, Wahafu A, Doyle SP, Cicka D, Yang X, Fan D, Reyna MA, Cooper LAD, Moreno CS, Zhou W, Owonikoko TK, Lonial S, Khuri FR, Du Y, Ramalingam SS, Mills GB, Fu H · Cell 2022
    PMID 35512704DOI
    What this evidence cannot show

    Full text read from PMC9597701 via efetch db=pmc (97,095 characters). MEN1 appears in the body exactly once, in the key resources table as a ThermoFisher anti-menin antibody (PA5-79663); every menin result is in the supplementary tables, which were not parsed here. So the allele assignments in this entry come from IntAct's curation of the paper, not from the paper's own narrative.

    Evidence that argues against this (1)
    • A BRET signal in a co-overexpression system is a proximity measurement in live cells, which is a real assay but is not a demonstration of a direct binary complex; the paper itself validates only a small subset orthogonally, and menin is not among the validated subset it discusses in text. Mo X et al. 2022
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    A MITAB 2.7 pull of IntAct on O00255 returns 59 rows and 33 distinct partners; 31 rows and 10 partners derive from PMID 35512704. On those rows IntAct annotates the partner-side feature as 'mutation causing an interaction' for BRAF 600, AKT1 17, TP53 282, EGFR 858 and PTEN 173, and as 'mutation with no effect' for SPOP 87/102/133, SMAD4 351/361/386, SMARCA4 885/910/1232/1243, FBXW7 465/479/505 and PTPN11 72. BRAF 600 additionally carries a pull-down annotated 'mutation increasing interaction strength' and an anti-tag co-immunoprecipitation.

    What this evidence cannot show

    PSICQUIC is one view of IntAct and may not match the web interface's aggregated evidence counts; the dispatch brief listed four PMIDs for CSNK2A2 where this pull returns one, which is the signature of that difference. The counts here describe the MITAB 2.7 response retrieved on the stated date and should be re-pulled rather than trusted if the numbers matter.

    Evidence that argues against this (1)
    • 'mutation with no effect' means the variant did not change a detected signal; it does not certify that the wild-type interaction is real. Treating those five as validated wild-type partners because the mutation did nothing would be the mirror error of treating the neo-interactions as wild-type partners. IntAct (EMBL-EBI) 2026
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    The ClinVar and dbSNP identifiers carried on the IntAct feature annotations resolve as follows: ClinVar 13961 = NM_004333.6(BRAF):c.1799T>A (p.Val600Glu); ClinVar 13983 = NM_001382430.1(AKT1):c.49G>A (p.Glu17Lys); ClinVar 12364 = NM_000546.6(TP53):c.844C>T (p.Arg282Trp); rs121434568 = EGFR residue 858; rs121913293 = PTEN residue 173; rs121918454 = PTPN11 residue 72; ClinVar 376414 = FBXW7 p.Arg465Cys; ClinVar 24832 and 8543 = SMAD4 p.Arg361His and p.Arg361Cys; ClinVar 30344 = SMARCA4 p.Arg885Cys; ClinVar 161486 and 694182 = SPOP p.Phe102Cys and p.Phe133Leu.

    What this evidence cannot show

    dbSNP rs records list every substitution observed at a codon, so an rs number identifies a residue and not an allele. EGFR 858 and PTEN 173 are therefore given at residue level only; the specific substitution used in the screen is in the paper's supplementary tables, which were not parsed. Protein positions from the Variation Services API are zero-based and were converted.

    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
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    IntAct's PSICQUIC response for MEN1 also returns HOXA7, HOXA9 and MEIS1 as interactors, sourced to PMID 18598942 with the method 'chromatin immunoprecipitation assay'. These are menin's target genes, not protein binding partners.

    PMID 18598942DOI
    What this evidence cannot show

    This is a property of how IntAct curates ChIP data, not an error in the source paper, which reports menin occupancy at HOXA loci exactly as everyone understands it. The trap is downstream: a script that enumerates the interactor column without filtering on the method column returns three transcription factors as menin binding partners.

    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
Connected to