Notes for DANRE coq8a

2026-05-09 review notes

• Core function is mitochondrial membrane COQ8A/ADCK kinase activity associated with ubiquinone biosynthesis [file:DANRE/coq8a/coq8a-uniprot.txt "Atypical kinase involved in the biosynthesis of coenzyme Q"].
• Protein kinase activity was modified to generic kinase activity because UniProt explicitly says substrate specificity is unclear and may involve small-molecule/lipid kinase activity [file:DANRE/coq8a/coq8a-uniprot.txt "substrate specificity is still unclear"].
• Protein phosphorylation/process annotations are marked over-annotated because no physiological protein substrate is established; ubiquinone biosynthesis is the directly supported process [file:DANRE/coq8a/coq8a-uniprot.txt "PATHWAY: Cofactor biosynthesis; ubiquinone biosynthesis."].

Re-review 2026-09-29

• The previous review cited zero PMIDs. Cached and cited the three primary COQ8 papers, all with full text: PMID:25498144 (ADCK3 atypical PKL fold), PMID:27499294 (COQ8A knockout mice; the source of the human NOT annotations), PMID:29198567 (conserved lipid-stimulated ATPase). A PubMed search found no zebrafish coq8a study, so every row is an ortholog-based inference and each reason says so.
• Resolved the PENDING IBA GO:0006744 row (node PTN000059692) to ACCEPT, arguing node placement rather than donor count: the UbiB family was defined by the ubiquinone phenotype and it holds across the clade PMID:29198567.
• Resolved the internal contradiction in the GOA record. The positive GO:0004672 protein kinase activity ISS row comes from the COQ8B paralog (Q96D53); the negated row comes from the COQ8A ortholog (Q8NI60) and traces to a direct assay PMID:27499294. Changed the positive row KEEP_AS_NON_CORE -> MODIFY to GO:0016301 kinase activity, with a propagation_review naming COQ8B as SOURCE_BAD for this claim, and changed both negated rows (GO:0004672, GO:0006468) KEEP_AS_NON_CORE -> ACCEPT, since the negation is what the experiments established. Validation still warns that GO:0004672 has two actions; that is the intended positive/negated split and is explained in both reasons.
• GO:0016310 phosphorylation ISS MARK_AS_OVER_ANNOTATED -> KEEP_AS_NON_CORE. Human COQ8A does carry an IDA for this term, so the row is not baseless, but the phosphotransfer is autophosphorylation that appears only when the regulatory loop is broken, and that same change blocks the pathway PMID:25498144.
• GO:0043531 ADP binding kept non-core, now grounded in the primary measurement PMID:25498144 rather than the UniProt paraphrase alone. GO:0016301 kinase activity ACCEPTed as the core MF at the level human curators chose from the same data.
• Localization rows now cite the ortholog evidence for the matrix face of the inner mitochondrial membrane [PMID:27499294 "Coq8p and COQ8A localize to the matrix face of the inner mitochondrial membrane"; PMID:29198567 "Endogenous COQ8 contains a TM domain that anchors it to the inner mitochondrial membrane (IMM) facing the matrix"].
• Comparator check on a possible NEW GO:0016887 ATP hydrolysis activity: not proposed. Human COQ8A, which is where the ATPase activity was demonstrated, carries no ATP hydrolysis activity annotation at all (its MF rows are GO:0016301 kinase activity IDA and GO:0043531 ADP binding IDA). A systematic absence on the best-characterized member is a convention, not a gap; the point is raised as a suggested question instead.
• Rewrote description as standalone biology (topology, autoinhibition, CoQ complex, disease); core_functions supporting quotes replaced with primary evidence. Deep research retained only for its one interpretive statement, attached to the kinase activity row. Validation: zero errors, one intrinsic warning.