This review covers the 31 imported GOA contexts without altering their source fields or order. It retains 25 reference records, two curated UniProt products and no supporting-text quotations. The direct evidence guide, database evidence guide and four interaction records record source access and assay limits.
The integrated decisions are 21 ACCEPT, 7 KEEP_AS_NON_CORE, 1 TOO_GENERAL and 2 UNDECIDED. The two nucleoid records describe the same partial localization and are both non-core. This preserves the positive observation from PMID:25683715; it does not infer DNA binding or absence from the nucleoid region. The precise matrix assignment in PMID:27974379 remains source-specifically uncertain, while independent curated matrix inferences are retained. The transferred RNA-modification claim remains unresolved because association and PTCD1/synthase perturbations do not isolate FASTKD2's contribution to the chemical-modification step.
One RNA-binding activity covers rRNA and mRNA substrates. Human experiments support mitochondrial RNA maturation and large-subunit assembly (PMID:25683715; PMID:26370583). The ancestral mRNA-stability assertion is retained with its actual PAINT placement and donor experiments; FASTKD2's ND6 contact and abundance phenotype do not by themselves resolve direct decay from secondary assembly defects (PMID:27789713; PMID:28335001; PMID:29617655). Mitochondrial translation is supported as a process annotation through RNA handling and assembly, without inventing a separate catalytic translation activity. No kinase, autonomous nuclease, pseudouridine-synthase or ribosomal structural activity is assigned.
The RNA-granule term is a cellular location rather than a protein-containing complex. The human module study describes largely monomeric migration and transient RNA-associated interactions (PMID:27974379); no stable obligate holocomplex is asserted. Organelle-level mitochondrial annotations remain informative alongside matrix and granule evidence.
Official UniProt and sequence reconstruction identify products Q9NYY8-1 (710 residues) and Q9NYY8-2 (648 residues). The second has the curated C-terminal replacement/deletion, not a measured alternative mitochondrial cleavage. Three captured RefSeq proteins match the first product. The captured Ensembl transcript models are distinguished from experimentally resolved protein products; sequence identity is not asserted for untested translations. UniProt leaves the transit-peptide endpoint and mature-chain start unresolved. Experimental tags and predicted targeting coordinates do not establish a mature product boundary.
The configured automated research providers failed authentication/quota checks, and no provider report was fabricated. The review instead uses genuine primary publications and database records with the access limits described in the evidence guides. A publication's committed normal cache may remain abstract-only even when the original article was recovered and read separately. Questions and proposed experiments are scoped to the inspected evidence rather than claims that no relevant work exists anywhere.
A bounded current search located PMID:41994142(https://pmc.ncbi.nlm.nih.gov/articles/PMC13082123/), a Research Square preprint (2026-04-07, version 1; DOI 10.21203/rs.3.rs-8605748/v1). Its original Results, relevant Discussion and RNA-assay Methods were inspected. The glioma-cell study reports FASTKD2/YTHDF2 association with 12S rRNA (MT-RNR1) and altered distribution of that RNA in tumour microtubes. This is distinct from the ND6-mRNA mechanism discussed above. The Results assign the MT-RNR1 stability perturbation to YTHDF2, while the RNA-stability Methods name FASTKD2 siRNA; this discrepancy remains unresolved in the inspected text. The preprint therefore does not establish isolated FASTKD2-dependent ND6 decay or settle the modification-module question. Its context-specific transport and drug-mechanism claims are not promoted to core functions or new GO assertions here. Figures, supplements and the drug-screen evidence were not independently adjudicated. The genuine source remains in the private research record; existing publication caches are unchanged.
The RNA-binding synthesis now distinguishes two substrate scopes: selected mitochondrial mRNAs and other transcripts, and 16S rRNA supporting large-ribosomal-subunit assembly. Human iCLIP supports the first (PMID:26370583); UV-crosslink immunoprecipitation/RNA-seq and sedimentation support the second (PMID:25683715, Figures 4–5). Direct ND6 turnover remains distinguishable from secondary RNA loss after ribosome-assembly failure. The two curated products and the standalone description are unchanged.
Matrix localization is accepted consistently, with the unresolved freely soluble versus matrix-facing membrane position retained as a precision limit. The committed UniProt CC SUBCELLULAR LOCATION and the original RNA-granule localization evidence support that compartment (PMID:25683715; PMID:27974379). Mitochondrial translation remains a positive, non-core role while the synthesis prioritizes resolved RNA handling and assembly. RNAi reduced translation without appreciable RNR2 abundance or processing change on the measured Northern blots, so the phenotype cannot be attributed solely to observed rRNA depletion. The proximal mechanism and a possible direct noncatalytic contribution remain unresolved (PMID:26370583, Figure 4D and pulse-labeling Methods).
The apoptosis annotation is marked over-annotated as a mechanistic process assignment. Its positive staurosporine and overexpression observations remain explicit. The original Results report normal respiratory-chain complex activities in the tested fibroblasts, with COX deficiency in skeletal muscle; a demonstrated fibroblast OXPHOS defect must not be invented to explain the phenotype (PMID:18771761, Figure 5 Results and final Discussion). A specific regulatory role versus indirect effects and overexpression toxicity remains unresolved.
The modification-module association is real, including RNase-resistant NGRN/WBSCR16 co-immunoprecipitation (PMID:27667664). The original paper leaves single versus overlapping complexes open, and the later 2D gels describe assay-specific largely monomeric migration (PMID:27974379). Ptcd1 perturbation and RPUSD4 chemistry do not independently isolate Fastkd2's possible structural or regulatory contribution to pseudouridylation (PMID:29617655). The corresponding transfer remains UNDECIDED; no absence claim or catalytic FASTKD2 activity is inferred.
The current totals are 21 ACCEPT, 7 KEEP_AS_NON_CORE, 1 TOO_GENERAL, 1 MARK_AS_OVER_ANNOTATED and 1 UNDECIDED. All 31 source objects, the 25 prior whole reference objects and both products are preserved; the curated UniProt record is appended as a reference with section locators. The four binding records remain exact. The feedback decisions, source guide and phylogenetic evidence provide publicly usable claim-level provenance without adding quotations or rewriting cached sources.