CEP112 is a 955-residue centrosomal coiled-coil protein on 7 spectra with zero leukaemia literature — its only nuclear annotation is a single 2024 citation, and a long coiled-coil is the classic non-specific AP-MS passenger — declined
also called MLL-AF4-CEP112 · KMT2A-CEP112 · CCDC46 · Centrosomal protein of 112 kDa · Cep112
CEP112 is a long, rope-like protein normally found at the centrosome, the structure that organises cell division. Long rope-like proteins stick to things non-specifically in this kind of experiment. Seven detections, no leukaemia literature at all, and nothing that connects it to our fusion.
The same thing, for a scientist
**Verdict: DECLINED.** CEP112 (Q8N8E3, 955 aa, formerly CCDC46) is annotated *"Cytoplasm, cytoskeleton, microtubule organizing center, centrosome"* ECO:0000269|PubMed:21399614, with **Nucleus** and **Cytoplasm** added from a single later citation, ECO:0000269|PubMed:39349455 [E1]. Its MLL-AF4 evidence is p = 0.0436 — the weakest of the six — on spectral counts 4|3 = 7, neg_ctrl 0|0, rank 95 of 291 [E2]. `CEP112[tiab] AND leukemia[tiab]` returns **0 records** [E3], and before this beat CEP112 was named in exactly one node file in this corpus, the cycle-158 gap report that listed it. **Two things decide it:** a long predominantly coiled-coil protein is a recognised non-specific binder in affinity purification, and the nuclear annotation that would make it compartment-compatible rests on one paper that was not read here. There is nothing to weigh on the other side.
Why this is the easiest decline of the six
Every axis this beat has is unfavourable and none is favourable:
| axis | CEP112 |
|---|---|
| p-value | 0.0436 — weakest of the six, rank 95 of 291 |
| spectra | 7 (4|3) |
| compartment | centrosome, experimentally coded 2011; nucleus from one 2024 citation |
| structure | long coiled-coil, a recognised non-specific AP-MS binder |
| independent literature | CEP112[tiab] AND leukemia[tiab] = 0 records |
| corpus footprint before this beat | 1 file, the gap report that listed it |
| second method | none |
| mechanism | none proposed by anyone |
The one point that deserves care is the nuclear annotation. UniProt added Nucleus and Cytoplasm to Q8N8E3 from ECO:0000269|PubMed:39349455, a 2024 citation. Rule 91 says an evidence code means a curator read an experiment, not that the paper is about the fact, and that paper was not opened for this entry. So CEP112 is not a clean compartment mismatch in the way ACADM is — it is a protein with a recent and unexamined nuclear claim. That is why this entry declines on the totality rather than on compartment alone, and why the nuclear citation is named here so a later reader can open it rather than inheriting the verdict.
Bounded rejection: what was considered and dropped
Two lines were considered and not pursued, recorded here so a later cycle repeats the decision rather than only the search.
CEP112 as a fusion partner in its own right. Centrosomal coiled-coil genes are common 3' partners in oncogenic fusions, and a CEP112 fusion would make it interesting for reasons unrelated to this row. CEP112[tiab] AND leukemia[tiab] returns 0 and no fusion query was run; a beat interested in the recombinome could check CEP112 against this project's fusion census, which this beat did not.
Mitosis as a mechanism. KMT2A has documented mitotic-bookmarking behaviour and a centrosomal partner is not absurd a priori. But the pulldown is from a nuclear extract of an asynchronous population, no mitotic enrichment was performed, and nothing in the source paper touches mitosis. There is no route from seven spectra to a mitotic mechanism that does not begin with a new experiment.
| partner a | gene-kmt2a |
|---|---|
| partner a uniprot | Q03164 |
| partner a length AA | 3969 |
| partner b | CEP112 |
| partner b uniprot | Q8N8E3 |
| partner b length AA | 955 |
| verdict | DECLINED |
| verdict scope | Declined on the totality - weakest p of the six, seven spectra, a coiled-coil architecture that binds non-specifically, and no literature or mechanism on any side. NOT declined on compartment alone, because UniProt carries a 2024 nuclear annotation that was not examined. |
| adjudicated by | Q180-the-af4-specific-interactors, cycle 159, 2026-09-24 |
| source publication pmid | 29777171 |
| source dataset | PXD009338, Skucha 2018 seven-bait MLL-fusion AP-MS |
| detection method | single-step StrepTactin affinity purification from Jurkat nuclear extract, LC-MS/MS spectral counting with a GLM background model |
| experimental role partner | prey |
| experimental role kmt2a | bait (as the MLL-AF4 fusion) |
| bait exists in dataset | true |
| n publications | 1 |
| n methods | 1 |
| orthogonal method exists | false |
| intact curated | false |
| af4 p | 0.0436 |
| af4 spectral | 4|3 |
| af4 spectral total | 7 |
| neg ctrl | 0|0 |
| rank in af4 by p | 95 |
| n fusions | 1 |
| druggability | Not assessable and not pursued. A coiled-coil scaffold with no pocket, no interface, no affinity and no disease rationale. |
| absence class | measured-not-searched |
| absence class note | Two cheap reads would settle this: the peptide evidence for the seven spectra in PXD009338, and PubMed 39349455, the sole source of CEP112's nuclear annotation. |
| shown for fusion | MLL-AF4 only, as one prey row in one AP-MS experiment in a Jurkat T-ALL host with a tagged transgene. |
| epigenetic readout | none - no histone mark, occupancy or chromatin measurement exists for CEP112 in any context. |
host cell lines
[ "Jurkat (human T-ALL)" ]
compartment
{
"partner_compartments": [
"Cytoplasm, cytoskeleton, microtubule organizing center, centrosome [ECO:0000269|PubMed:21399614]",
"Nucleus [ECO:0000269|PubMed:39349455]",
"Cytoplasm [ECO:0000269|PubMed:39349455]"
],
"note": "NOT a clean mismatch. The nuclear annotation is recent (2024 citation) and was not examined; opening PubMed 39349455 is the cheapest thing that could change this verdict.",
"architecture": "predominantly coiled-coil, 955 aa - a recognised class of non-specific affinity-purification binder"
}measurements
[
{
"quantity": "spectral_counts",
"subject": "CEP112|MLL-AF4",
"value": 7,
"units": "spectra (4|3 over two replicates)",
"n": 2,
"universe": "Skucha 2018 Suppl Table 5, MLL-AF4 column",
"null": 0,
"note": "null column is the neg_ctrl (0|0); p = 0.0436, rank 95 of 291 - last of the six by p-value, fourth of six by spectral count"
},
{
"quantity": "pubmed_records",
"subject": "CEP112 AND leukemia",
"value": 0,
"units": "records",
"n": null,
"universe": "PubMed titles and abstracts, 2026-09-24, one [tiab] per term; CCDC46 not searched",
"null": null,
"note": "querytranslation non-empty, so a genuine zero; bounded by a one-name list and therefore a floor"
}
]interface residues a
[]
structures
[]
ligands
[]
- Nothing here touches an intervention. The procedural risk is that a protein with zero literature is easy to write up as a 'novel' partner precisely because nothing contradicts it - novelty and absence of evidence look identical in a graph, and this entry exists so that they do not here.
- What does PubMed 39349455 actually show about nuclear CEP112? It is the sole basis of the nuclear annotation and the only cheap thing that could reopen this entry.
- Is CEP112 a documented 3' fusion partner in any leukaemia? Centrosomal coiled-coil genes recur in oncogenic fusions, and this beat did not check CEP112 against this project's fusion census. That question is independent of the AP-MS row and would be interesting whatever this entry concludes.
- How many of the 110 MLL-AF4-specific preys are long coiled-coil proteins? If the class is enriched for them, that is a second measurable contamination signature alongside the mitochondrial one and would let several members be declined in one pass.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
CEP112 is UniProt Q8N8E3 (CE112_HUMAN), primary gene symbol CEP112, synonym CCDC46, 955 aa, 'Centrosomal protein of 112 kDa', annotated to centrosome/microtubule organizing center with ECO:0000269|PubMed:21399614 and to Nucleus and Cytoplasm with ECO:0000269|PubMed:39349455.
UniProtKB Q8N8E3 (CE112_HUMAN), Centrosomal protein of 112 kDa — accession, primary gene name, synonym, length and subcellular-location comment retrieved 2026-09-24 · UniProt Consortium, EMBL-EBI/SIB/PIR 2026What this evidence cannot show
Resolved by accession with the primary gene name read back (rules 62, 65); no row in data/symbol-collisions.json for CEP112 or CCDC46. Rule 91 in full force: PubMed 39349455 was NOT opened, so whether it establishes a nuclear CEP112 pool as a finding or mentions it incidentally is unknown, and this entry does not claim to know. The retired symbol CCDC46 was not used in any literature query, which under the retired-symbol rule means the literature counts here are floors.
Evidence that argues against this (1)
- The 2024 nuclear annotation is the one thing that could make this row worth a second look, and it arrived AFTER the 2018 experiment. If PubMed 39349455 reports a substantive nuclear CEP112 function, the compartment argument against this row weakens considerably and the decline would rest only on the seven spectra. Anyone revisiting this entry should open that paper first; it is the cheapest thing that could change the verdict. UniProt Consortium 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
CEP112's entire evidence as an MLL-AF4 partner is one prey row in Skucha 2018 Supplementary Table 5: p = 0.0436 — the weakest of the six AF4-specific candidates below p 0.05 — on spectral counts 4|3 (7 summed), neg_ctrl 0|0, rank 95 of 291, recovered under MLL-AF4 and none of the other six baits.
MLL-fusion-driven leukemia requires SETD2 to safeguard genomic integrity. · Skucha A, Ebner J, Schmöllerl J, Roth M, et al. (20 authors; Grebien F senior author) · Nature Communications 2018What this evidence cannot show
Read from the cached Supplementary Table 5 workbook with the column map asserted against the sheet's header row. CEP112 is not discussed anywhere in the paper. At rank 95 of 291 it sits well into the region where the MLL-AF4 list is majority non-significant; the largest p retained in that column is 0.175.
Evidence that argues against this (1)
- Seven spectra split 4 and 3 across replicates with a zero control is reproducible detection, and on the raw counts CEP112 is better supported than RGPD8 (3 spectra) or RPA2 (6). Ranking the six by p-value, as the dispatch brief did, puts CEP112 last; ranking by spectral count puts it fourth of six. Which ranking is right is not settled, and this entry's 'weakest of the six' is true only on the p-value axis. Nature Communications 2018
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
CEP112 has no leukaemia literature at all: CEP112[tiab] AND leukemia[tiab] returns 0 records on 2026-09-24, and before this beat CEP112 appeared in exactly one node file of this corpus — the cycle-158 entry that listed it as an unwritten gap.
PubMed esearch via scripts/pubmed_safe.py plus scripts/beat-inventory.py word-boundary naming counts, both run 2026-09-24 by Q180. One [tiab] tag per term (rule 94); querytranslation read back as "CEP112"[Title/Abstract] AND "leukemia"[Title/Abstract], non-empty, so this is a genuine zero and not an instrument failure (rule 97). · NCBI PubMed E-utilities and this repository 2026What this evidence cannot show
Bounded by titles and abstracts (rule 80). Name list tried: CEP112 only. NOT tried: CCDC46 (the retired symbol, and the one under which any pre-2011 literature would be indexed), Cep112, 'centrosomal protein 112', or the British spelling leukaemia. Under the retired-symbol rule this is a substantial gap - that rule's worked example found a tenfold undercount for exactly this reason - so the zero is bounded by a weak name list and is a floor.
Evidence that argues against this (1)
- A zero literature is an absence of anyone having looked, not evidence against an interaction, and this project's standing rule keeps those apart. It is cited here only to establish that there is no external corroboration to weigh, and the weak name list means even that is provisional: CCDC46 was not searched and is where older work would be indexed. UniProt Consortium 2026
- related to →KMT2A / MLL1 — the gene that breaks
declined KMT2A::AFF1-CEP112 pair: weakest p of the six, no literature on any side
sits at rank 95 of 291, well inside the region the calibration shows is non-significant
CEP112 at rank 95 of 291 sits inside the region this decomposition shows is majority non-significant