Oncofusion

start / graph / gene-brinp1

Geneweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.

BRINP1 (DBC1/DBCCR1/FAM5A) - the last bait-supported partner of the six core proteins without an entry, a neural-specific BMP/retinoic-acid-inducible protein pulled down with DOT1L as bait in one quantitative AP-MS experiment and nowhere else

also called BRINP1 · DBC1 (deleted in bladder cancer 1) · DBCCR1 · FAM5A · O60477 · the last unwritten directed interactor

A survey of which proteins are known to physically associate with the six core proteins of this project found one that had never been written up. It is a brain-specific protein, and the only evidence is a single large-scale experiment in which many proteins were fished out at once. It is recorded here so the survey is complete, with the honest note that one such result is weak evidence of a real partnership.

The same thing, for a scientist

**Why this entry exists.** The PI's first standing priority is *map EVERY interactor of KMT2A::AFF1 with evidence*. Re-running the directed-interactome coverage check this cycle gave **50 of 53** bait-supported partners of the six core proteins with an entry. Of the three without, **two are not partners at all**: `Q2Q440` is submitted to UniProt as *"MLL/ENL fusion protein"* and is a **20-residue** junction fragment, and `Q8TD98` is submitted as *"MLL/AF4 fusion protein"* — **the fusion proteins themselves**, appearing as accessions and therefore as apparent partners of their own components [E2]. **So the real figure is 50 of 51, and BRINP1 is the one.** **What it rests on, precisely.** IntAct holds exactly one DOT1L–BRINP1 record: **DOT1L (Q8TEK3) as bait, BRINP1 isoform 3 (O60477-3) as prey**, detection *anti-tag co-immunoprecipitation*, interaction type *association*, from a single publication — **Hein MY et al., "A human interactome in three quantitative dimensions organized by stoichiometries and abundances", *Cell* 163:712-23 (2015), PMID 26496610** [E1]. One experiment, one direction, one isoform. **BRINP1 is neural-specific.** It is a BMP/retinoic-acid-inducible protein that functions in neurogenesis and brain development and suppresses G1/S transition in postmitotic neurons; its aliases include **DBC1/DBCCR1**, from deletion in bladder cancer, which is a different gene from the CCAR2 protein also called DBC1 and is a live naming hazard in this corpus. 🔴 **The honest reading is that this is probably not a functional partner of the fusion complex.** A neural-specific secreted/membrane-associated protein recovered once, as prey, in a HeLa-based tagged-bait screen is the classic shape of a low-stoichiometry co-purifier. **It is written up because the survey's denominator should be complete and because an unwritten accession is indistinguishable from an unexamined one** — not because there is reason to think DOT1L and BRINP1 do anything together.

Go deeper
Two of the three 'missing partners' were the fusion proteins themselves

The coverage check's denominator is partners with at least one curated interaction where one of our six proteins was the bait. It already excluded two accessions with stated reasons — Q9V3G3 as not human and Q9Y5J1 as an IntAct curation error. Two more should join them and are recorded here so the figure stops needing rediscovery:

accessionUniProt submitted namelengthwhy it is not a partner
Q8TD98MLL/AF4 fusion protein624 aait is the index fusion, curated as an interactor of its own components
Q2Q440MLL/ENL fusion protein20 aaa junction fragment from an exon-combination survey, not a protein

Both are unreviewed TrEMBL entries with no gene name and no recommended protein name, which is why they surfaced as anonymous accessions rather than as recognisable genes.

🔵 This is a denominator artefact of a real and known kind — this project already keeps analyses/intact-fusion-construct-census/ for fusion constructs appearing in IntAct — and it means the PI's directed-interactome dimension reads 50/51 = 98.0%, not 94.3%.

What would make this a real partner, and what would retire it

It would become interesting if BRINP1 recurred with a second bait among the six, or appeared in a KMT2A-rearranged leukaemia AP-MS experiment, or carried a stoichiometry in Hein 2015 above the bystander range. That last is checkable from the paper's own supplementary tables and was not done here — the whole point of that study is that it reports stoichiometries, so the question is this a bystander? has a published answer that this entry does not yet quote.

It should be retired to a bounded absence if the stoichiometry is at the bottom of the distribution and no second observation exists. Either way the survey figure is unaffected, because coverage measures whether the partner has been examined, not whether it survived examination.

Facts and figures
symbolBRINP1
uniprotO60477
uniprot IDBRNP1_HUMAN
length AA761
alias collision warningDBC1 is also a common alias for CCAR2, a different and better-studied chromatin-associated protein; searches for DBC1 will conflate them
function uniprotBMP/retinoic acid-inducible neural-specific protein 1; roles in neurogenesis and brain development; reported to suppress G1/S transition in postmitotic neurons
why this entry existsit was the last bait-supported partner of the six core proteins without an entry; the PI's first priority is to map EVERY interactor
recurrence1 bait; this project has measured that held partners recur across a mean of 5.6 baits and missing ones across 2.4, so BRINP1 sits at the extreme weak end
absence classmeasured-not-curated
bite backThis entry exists to complete a denominator and says so. If a future reader treats the presence of a gene- entry as evidence that DOT1L-BRINP1 is a credible interaction, the coverage statistic will have manufactured the belief it was meant to measure.
aliases
[
  "DBC1",
  "DBCCR1",
  "FAM5A"
]
evidence for association
{
  "bait": "DOT1L (Q8TEK3)",
  "prey": "BRINP1 isoform 3 (O60477-3)",
  "method": "anti tag coip",
  "interaction_type": "association",
  "publications": 1,
  "pmid": "26496610",
  "pride": [
    "PXD002829",
    "PXD002815"
  ],
  "intact_records_for_brinp1_total": 35,
  "of_which_involve_a_core_protein": 1
}
denominator correction
{
  "before": "50 of 53 = 94.3%",
  "after": "50 of 51 = 98.0%",
  "excluded_Q8TD98": "UniProt submitted name 'MLL/AF4 fusion protein', 624 aa, unreviewed",
  "excluded_Q2Q440": "UniProt submitted name 'MLL/ENL fusion protein', 20 aa, unreviewed",
  "reason": "these are the fusion proteins themselves, not partners of their own components"
}
Still open
  1. Retrieve the DOT1L-BRINP1 stoichiometry from Hein 2015's supplementary tables. The paper's central contribution is stoichiometry and abundance, so whether this is a bystander has a published answer, and quoting it would either promote or retire this entry in one step.
  2. Check whether BRINP1 appears in any KMT2A-rearranged leukaemia AP-MS dataset, including the Skucha MLL-AF4 interactome table this project already holds. A second observation in a relevant system would change the reading entirely.
  3. Decide whether the two fusion-construct accessions belong in the denominator. They are excluded here with reasons, but a curated interaction between AFF1 and an MLL/AF4 construct may be real evidence about self-association rather than noise.
  4. Add Q8TD98 and Q2Q440 to the coverage script's exclusion list with these reasons, so the 50/51 figure stops needing rediscovery - the same number has appeared in this project's ledger as 100%, 68.5% and 94.3% simultaneously.
Every source behind this entry (2)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Querying IntAct for BRINP1 returns 35 interaction records, of which exactly one involves any of this project's six core proteins: DOT1L, UniProt Q8TEK3, in the experimental role of bait, against BRINP1 isoform O60477-3 as prey. The detection method is anti-tag co-immunoprecipitation and the interaction type is association. The supporting publication is PMID 26496610, DOI 10.1016/j.cell.2015.09.053, with associated PRIDE deposits PXD002829 and PXD002815. UniProt records O60477 as BRNP1_HUMAN, BMP/retinoic acid-inducible neural-specific protein 1, 761 residues, gene BRINP1 with synonyms DBC1, DBCCR1 and FAM5A, functioning in neurogenesis and brain development and reported to suppress G1/S transition in postmitotic neurons.

    A human interactome in three quantitative dimensions organized by stoichiometries and abundances · Hein MY, Hubner NC, Poser I, Cox J, Nagaraj N, Toyoda Y, Gak IA, Weisswange I, Mansfeld J, Buchholz F, Hyman AA, Mann M · Cell 163(3):712-723 2015
    PMID 26496610DOI
    What this evidence cannot show

    THIS IS ONE OBSERVATION FROM ONE EXPERIMENT AND IT IS NOT IN A LEUKAEMIA SYSTEM. The interaction type is 'association', which in IntAct means co-purification and not a demonstrated direct contact, and a single anti-tag co-IP in a tagged-bait screen is the canonical shape of a low-stoichiometry co-purifier. The partner is recorded against ISOFORM 3 specifically, and nothing here establishes that the canonical isoform behaves the same way. BRINP1 IS NEURAL-SPECIFIC while the bait screen is not a neural system, which makes biological relevance to a leukaemia fusion complex doubtful on its face. THE STOICHIOMETRY IS PUBLISHED IN THIS VERY PAPER AND WAS NOT RETRIEVED - the study's central contribution is stoichiometry and abundance estimates, so the question of whether this is a bystander has a published answer this entry does not quote, and the confidence here would move in either direction once it is read. The IntAct query was run against the gene symbol BRINP1; a query by accession or by an alias such as DBCCR1 might return a different record set.

    Evidence that argues against this (3)
    • The strongest evidence against this being a partner is the absence of a second observation anywhere. This project's own interactome work found that held partners recur across a mean of 5.6 baits while missing ones recur across 2.4; BRINP1 recurs across ONE, which places it at the extreme weak end of the distribution the project has already characterised as a long private tail whose members may not be reliable partners at all.
    • Writing an entry for a partner in order to complete a coverage statistic is the behaviour the coverage statistic was meant to motivate, and it is worth being explicit that the causality has run backwards here. Nothing about BRINP1 attracted attention; only its absence from a denominator did. A reader should not infer from this entry's existence that anyone considers DOT1L-BRINP1 biologically interesting.
    • The alias DBC1 is shared with CCAR2, an entirely different and far better-studied protein that does interact with chromatin regulators. Any future automated search of this corpus for 'DBC1' will conflate them, and this project has already logged symbol collisions of exactly this kind as a recurring source of wrong identifications.
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis

    Re-running the directed-interactome coverage check against IntAct gives a union of 56 bait-supported partner accessions across the six core baits, reduced to a clean human denominator of 53 after excluding an empty accession, the non-human Q9V3G3 and the miscurated Q9Y5J1. Of those 53, 50 have an entry. Resolving the three that do not against UniProt shows that two are unreviewed TrEMBL entries carrying no gene name and no recommended protein name, whose submitted names are 'MLL/AF4 fusion protein' (Q8TD98, 624 residues) and 'MLL/ENL fusion protein' (Q2Q440, 20 residues) - that is, KMT2A fusion proteins curated as interactors of their own constituent proteins. The third is O60477, BRINP1.

    What this evidence cannot show

    THE CLEANED FIGURE OF 50/51 IS A COVERAGE STATISTIC, NOT A BIOLOGICAL ONE: it says every bait-supported partner has been examined, not that the interactome is understood. The same check records that only the BAIT-SUPPORTED interactome is near-complete, while the prey-only population - 136 partners at the last count - remains almost entirely unwritten, and that population is far larger. EXCLUSIONS ARE JUDGEMENTS: removing the two fusion-construct accessions is defensible but it is a decision to shrink a denominator, and the same move made carelessly is how coverage figures are inflated. Both exclusions are recorded with their UniProt submitted names so the judgement can be reversed. THE FIGURE IS CACHED BETWEEN INTACT REFRESHES and drifts silently; this project has already had this exact number reported as 100%, 68.5% and 94.3% in three places in its own ledger simultaneously.

    Evidence that argues against this (1)
    • Excluding the fusion-construct accessions makes the coverage number look better, and it was decided by the same person who benefits from the better number. An equally defensible reading is that a curated interaction between AFF1 and an MLL/AF4 fusion construct is REAL EVIDENCE about the fusion's self-association or about construct behaviour in the assay, and deleting it from the denominator discards information rather than cleaning noise.
Connected to