Oncofusion

start / graph / gene-ebna1bp2

Geneweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q75-dot1l-partners

EBNA1BP2 (EBP2) - the DOT1L spoke with the HIGHEST stoichiometry and the WEAKEST evidence, and the case that shows why abundance is the third dimension

also called EBP2 · EBNA1-binding protein 2 · Nucleolar protein p40 · p40

EBP2 is a very common protein that lives in the nucleolus, the part of the nucleus where ribosomes are built, and there is about half as much of it as DOT1L in a DOT1L pulldown - which sounds like a real partnership. It is not: there is 466 times more EBP2 than DOT1L in the cell to begin with, so the amount recovered is what you would get by accident.

The same thing, for a scientist

EBNA1BP2/EBP2 (Q99848, 306 aa, nucleolar rRNA-processing factor) is the most instructive of the six unwritten DOT1L partners, because **on interaction stoichiometry alone it looks like a complex member and on the paper's own grading it is the weakest hit in the pulldown.** Its interaction stoichiometry is **0.442 (1 EBP2 per 2.3 DOT1L)** - higher than AF17 and level with AF10, the only prey flagged `core.interactor` - while its confidence class is **B-**, its enrichment is **1.056 SD** (the lowest of all ten rows, self included) and its prey-bait profile correlation is **0.057** [E2]. The resolving number is the third dimension: **abundance stoichiometry 466**, i.e. HeLa contains 1,732,944 copies of EBP2 against 3,779 of DOT1L [E2]. A protein present at 459-fold excess needs only 0.1% non-specific carry-over to reach a 1:2 apparent stoichiometry. EBP2 is nucleolar (Q99848, ECO:0000269 on two papers) [E3] while DOT1L's measured location is nucleoplasmic [E4], and `DOT1L[tiab] AND (EBNA1BP2 OR EBP2 OR "nucleolar protein p40")[tiab]` returns one record which is a *C. elegans* paper matching the worm gene EBP-2 [E5].

Go deeper
The number that would have fooled a stoichiometry filter

This beat was dispatched on the premise that interaction stoichiometry is what separates a complex member from a trace. EBNA1BP2 is the counter-example, and it is the sharpest one in the dataset.

interaction stoichiometryclassenrichmentcorrelationabundance stoichiometry
MLLT10 / AF10 (core.interactor)0.416A+6.920.4200.12
EBNA1BP20.442B-1.060.057466
MLLT3 / AF9 (co-structures exist)0.0058A+3.740.2750.39

Ranked on stoichiometry, EBP2 comes first among all non-self preys and AF9 comes last. Ranked on any other column it is last. That is the paper's whole argument in one row: a stoichiometry is meaningless without the abundance it is measured against, which is exactly why Hein et al. reported three dimensions and titled the paper for them 1.

The arithmetic: at 1,732,944 copies against DOT1L's 3,779, recovering 0.1% of the cell's EBP2 non-specifically gives 1,733 molecules against 3,779 baits - an apparent 1:2.2 stoichiometry, from pure carry-over 2.

The nucleolar pattern, checked rather than assumed

The dispatch asked whether a nucleolar-contaminant pattern is emerging across the DOT1L and KMT2A pulldowns, since another beat is separately looking at a nucleolar protein in an MLL complex. What I can say from my own data, and no more:

  • EBNA1BP2 is the only nucleolar protein among the six DOT1L spokes. It is not a pattern of six; it is one of six.
  • It is, however, the single most abundant protein in the whole DOT1L pulldown by a factor of 44 over the next prey, and nucleolar rRNA-processing factors are the best-documented AP-MS background class there is.
  • The reciprocal arm in the same dataset is suggestive. GFP-IBTK - another of the six, and a cytoplasmic protein - recovered a prey list dominated by ribosomal proteins and nucleolar factors: DDX21, NAT10, TSR1, WDR74, POP7, RPP25L and seven ribosomal subunits 6. The same background class appears under a completely unrelated bait in the same cells, which is what a contaminant looks like and is not what a specific partner looks like.

What I cannot say is whether the KMT2A-side nucleolar observations are the same phenomenon. That is a different bait, possibly a different study, and it belongs to another beat. Two nucleolar hits under two baits is not a pattern until somebody counts nucleolar preys against their abundance across all six baits, and the memo records that as the test.

What would make this real

EBP2 is not an absurd DOT1L partner a priori. DOT1L has documented roles at rDNA-adjacent chromatin in other systems, EBP2 binds EBNA1 and chromosome-associated structures in mitosis, and both are chromatin-proximal. The evidence that would convert this from background to biology is not more AP-MS. It is a reciprocal quantitative pulldown from an EBP2 bait, or an imaging experiment showing DOT1L in the nucleolus. Neither exists: EBNA1BP2 was not one of Hein et al.'s 1,330 bait cell lines 6, and no published DOT1L localisation study reports a nucleolar pool 4.

Facts and figures
uniprotQ99848
length AA306
gene symbolEBNA1BP2
protein nameProbable rRNA-processing protein EBP2 (EBNA1-binding protein 2, nucleolar protein p40)
uniprot resolutionaccession:Q99848 AND organism_id:9606 AND reviewed:true; one reviewed hit; primary gene name read back as EBNA1BP2; length 306 aa checked
shown for fusionNO - nobody has looked in any fusion. Zero PubMed records pair MLL or KMT2A with EBNA1BP2 or EBP2.
interaction with dot1l
{
  "source_pmid": "26496610",
  "source_class": "proteome-scale AP-MS atlas",
  "intact_interaction_ac": "EBI-11060394",
  "imex_id": "IM-24272-464",
  "detection_method": "psi-mi:MI:0007 anti tag coimmunoprecipitation",
  "interaction_type": "psi-mi:MI:0914 association",
  "expansion": "psi-mi:MI:1060 spoke expansion - binary pair inferred, not observed",
  "intact_miscore": 0.35,
  "author_confidence_class": "B-",
  "core_interactor": false,
  "interaction_stoichiometry_prey_per_bait": 0.4422,
  "interaction_stoichiometry_as_ratio": "1 EBP2 per 2.3 DOT1L - the highest of any non-self prey in the pulldown",
  "log10_prey_bait_ratio": -0.354,
  "abundance_stoichiometry_prey_per_bait": 466.4,
  "log10_prey_bait_expression_ratio": 2.668787,
  "enrichment_sd": 1.056,
  "neg_log10_p": null,
  "neg_log10_p_note": "reported as Inf in the source table and therefore unusable",
  "prey_bait_correlation": 0.057,
  "valid_values": 3,
  "hela_copies_per_cell": 1732943.74,
  "independent_sources": 0,
  "reciprocal_tested": false,
  "reciprocal_note": "EBNA1BP2 is not one of the study's 1,330 bait cell lines, so no reciprocal arm exists"
}
localization
{
  "subcellular": "Nucleus, nucleolus (UniProt ECO:0000269 on PMID 19170763 and PMID 2879624), RNA-dependent nucleolar association. DOT1L's measured location is nucleoplasmic with no annotated nucleolar pool, so the two are annotated to different subnuclear compartments.",
  "source": "UniProt Q99848 and Q8TEK3, fetched 2026-09-21; coded but not traced to the cited experiments (brief rule 91)"
}
absence claim
{
  "queries_run": [
    "DOT1L[tiab] AND (EBNA1BP2 OR EBP2 OR \"nucleolar protein p40\")[tiab]",
    "(MLL OR KMT2A)[tiab] AND (EBNA1BP2 OR EBP2)[tiab]"
  ],
  "records": [
    1,
    0
  ],
  "on_topic": [
    0,
    0
  ],
  "rejected_hits": [
    "PMID 32781660 - C. elegans DOT-1.1 / H3K9me2 paper, matches the nematode gene EBP-2"
  ],
  "operand_controls": {
    "DOT1L[tiab]": 710
  },
  "date": "2026-09-21",
  "tool": "PubMed E-utilities esearch",
  "bounded_by": "titles and abstracts only; supporting figures not searched"
}
How this could bite us back
  • Written down carelessly, this becomes 'DOT1L is in the nucleolus', which nothing supports, and which would then be used to explain rDNA phenotypes that have other causes.
  • The opposite failure is the real risk for a graph that prunes: EBP2 has the highest interaction stoichiometry in the pulldown, and if the interaction is genuine then discarding it on an abundance prior removes the one DOT1L contact that looked stoichiometrically like a complex.
Still open
  1. Across all 1,330 baits in this atlas, how does a prey's recovery scale with its cellular abundance? That regression would turn 'EBP2 is abundant background' from a prior into a measurement, and the data to run it are the two workbooks already cited here.
  2. Is there any DOT1L in the nucleolus? No published DOT1L imaging study reports a nucleolar pool, and a single immunofluorescence with a nucleolar marker would settle whether this spoke can be spatially possible.
  3. Does the KMT2A side of this graph show the same nucleolar preys under its own baits? Two nucleolar hits under two baits is not a pattern until nucleolar preys are counted against abundance across all six core baits.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    PMID 26496610 (Hein et al., Cell 2015) is a proteome-scale GFP-tagged AP-MS atlas of 1,125 HeLa cell lines whose central claim is that interaction stoichiometry and cellular abundance together, not enrichment alone, define which associations are stable complexes - the minority of stable complexes being identifiable by their stoichiometry signature against a network dominated by weak substoichiometric interactions.

    PMID 26496610DOI
    What this evidence cannot show

    Identifiers copied character-by-character from a live NCBI efetch XML record fetched 2026-09-21 reading only PubmedData/ArticleIdList/ArticleId (brief rule 81). Abstract read in full; main text unreadable - no PMCID, Europe PMC isOpenAccess N, DOI resolves to a 13-character Elsevier redirect stub. The paraphrase of the paper's thesis is taken from the abstract, not from the discussion.

    Evidence that argues against this (1)
    • The paper's framework is designed to classify a network, and it explicitly accepts that weak substoichiometric interactions are the majority and are real. Using its own logic to argue a single interaction away goes beyond what the framework claims, because 'substoichiometric' is the paper's description of most of the true interactome. Hein MY et al. 2015
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    In the source study's interactions table the DOT1L-EBNA1BP2 spoke is class B- with interaction stoichiometry 0.4422 (1 EBP2 per 2.3 DOT1L, log10 prey:bait -0.354), abundance stoichiometry 466.4 (log10 2.669), enrichment 1.056 SD - the lowest of all ten rows in the pulldown - prey-bait profile correlation 0.057, and core.interactor not set; in the same paper's HeLa proteome EBNA1BP2 is at 1,732,944 copies per cell against DOT1L's 3,779.

    What this evidence cannot show

    Downloaded 2026-09-21 from the Elsevier static CDN with magic bytes checked (PK). The -log10 p value for this row is reported as Inf in the source table and is therefore not usable; it is recorded as null in data rather than silently dropped. Copy numbers are proteomic-ruler estimates without dispersion. The prey group is Q99848;H7C2Q8, so the row does not distinguish EBNA1BP2 isoforms.

    Evidence that argues against this (1)
    • High abundance makes a genuine encounter more likely as well as a contaminant more likely, so the 459-fold excess lowers the posterior without settling it. The step from 'abundant' to 'artefact' is a prior, and the only thing that would settle it is a reciprocal or imaging experiment, neither of which exists. Cell 163(3), supplementary material 2015
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q99848 (EBP2_HUMAN) is human EBNA1BP2, 306 aa, 'Probable rRNA-processing protein EBP2 / EBNA1-binding protein 2 / Nucleolar protein p40', gene synonym EBP2, with subcellular location Nucleus/nucleolus carrying experimental evidence codes (ECO:0000269 on PMID 19170763 and PMID 2879624) and the note that association with the nucleolus is RNA-dependent.

    What this evidence cannot show

    Resolved 2026-09-21 by accession with the primary gene name read back as EBNA1BP2 and the length checked. Per brief rule 91 the ECO:0000269 codes mean a curator read an experiment in those papers, not that either is the founding or best localisation study for EBP2; I did not open PMID 19170763 or 2879624, so the nucleolar call is recorded as coded-but-untraced.

    Evidence that argues against this (1)
    • A nucleolar protein is not thereby excluded from nucleoplasmic complexes - nucleolar residence is dynamic and RNA-dependent by UniProt's own note, and several rRNA-processing factors have documented chromatin functions outside the nucleolus. Compartment mismatch lowers the prior; it does not close the question. UniProt Consortium 2026
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    UniProtKB Q8TEK3 gives DOT1L the subcellular location Nucleus on an experimental code (ECO:0000269, PMID 15851025) plus Cytoplasm by similarity only, with a note that cytoplasmic localisation is dependent on and mediated by MLLT6/AF17; no nucleolar pool is annotated.

    What this evidence cannot show

    Fetched 2026-09-21. The cytoplasmic call is ECO:0000250 (by similarity to the mouse orthologue Q6XZL8), not a human measurement - this graph already records that in localization-dot1l-nucleus-measured-cytoplasm-by-similarity. An absent nucleolar annotation is an absence in a database, not a measured negative, and UniProt does not systematically annotate nucleolar sub-pools.

    Evidence that argues against this (1)
    • The AF17-dependent cytoplasmic pool is itself a reminder that DOT1L's compartment is partner-determined, so an argument of the form 'DOT1L is not in the nucleolus, therefore the EBP2 spoke is background' is weaker than it looks - a partner could in principle take DOT1L somewhere UniProt does not list. UniProt Consortium 2026
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Counted absence, PubMed E-utilities esearch run 2026-09-21: DOT1L[tiab] AND (EBNA1BP2 OR EBP2 OR "nucleolar protein p40")[tiab] returns 1 record, PMID 32781660, 'Caenorhabditis elegans Deficient in DOT-1.1 Exhibit Increases in H3K9me2 at Enhancer and Certain RNAi-Regulated Regions', which matches on the nematode gene EBP-2 and is off topic; (MLL OR KMT2A)[tiab] AND (EBNA1BP2 OR EBP2)[tiab] returns 0; operand control DOT1L[tiab] returns 710.

    What this evidence cannot show

    The querytranslation showed that 'nucleolar protein p40' expanded loosely into a MeSH nuclear-proteins set ANDed with p40, so that limb of the query is imprecise rather than tight; the single returned record was title-checked and rejected. Bounded by titles and abstracts (brief rule 80). EBP2 is a short alias that also names an unrelated nematode gene and a yeast protein, which is why every hit needed reading rather than counting.

    Evidence that argues against this (1)
    • The literature would be silent whether or not the interaction is real, since this pair has only ever been reported in a supplementary table. The absence is recorded for completeness and carries little weight on its own. NCBI PubMed (queries run 2026-09-21) 2026
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    EBNA1BP2 is not among Hein et al.'s 1,330 bait cell lines, so no reciprocal arm exists for this pair; IBTK, which is a bait in that study, recovered a prey list dominated by ribosomal proteins and nucleolar factors (DDX21, NAT10, TSR1, WDR74, POP7, RPP25L and seven ribosomal subunits) despite being a cytoplasmic protein, showing that this background class appears under unrelated baits in the same cells.

    What this evidence cannot show

    One comparison bait is an anecdote, not a background model. A proper test would rank nucleolar and ribosomal preys against their cellular abundance across all 1,330 baits and ask whether the DOT1L pulldown is unremarkable - that test was not run here and the memo records it as the next step. IBTK's prey list may also be genuinely ribosome-related, since IBTK has reported translational functions, which would weaken the background reading.

    Evidence that argues against this (1)
    • IBTK has documented roles in translation and ribosome biology, so its ribosomal prey list may be specific rather than background - in which case it is a poor negative control and the nucleolar-contaminant reading loses its strongest support in this entry. Cell 163(3), supplementary material 2015
Connected to