Reviewed all 140 seeded annotations and the three official products. The final draft contains 49 ACCEPT, 46 KEEP_AS_NON_CORE, 37 MODIFY and 8 UNDECIDED decisions, one integrated RNase III core and no NEW annotations. The original source terms, evidence codes, references, partner identifiers, qualifiers and negative flag are preserved. The 144-row raw QuickGO response contained four duplicate pairs collapsed by the normal bootstrap.
The evidence guide records actual original access, public source locators, construct/species limitations, interaction evidence, ontology scope and propagation paths. Current UniProt products reconstruct to 1922, 1829 and 820 amino acids. Historical 1912-aa clones and engineered fragments are not assigned to current isoforms from length alone. The t-Dicer assay uses pre-miR-549; partial asymmetric cleavage does not establish universal inactivity or dominant-negative behavior.
The core integrates precursor RNA cleavage and performed loading/assembly work. AGO2 performs target slicing. Cytosolic localization is nonexclusive; nuclear, perinuclear and extracellular pools are retained contextually. The NOT-RISC row reflects source-specific effector release and is not rewritten as a positive source row or an absolute exclusion from every dynamic complex. The separate contributes-to precursor-binding annotation is retained at its original weaker complex scope.
Supported protein associations are retained as non-core when a specific partner-class refinement cannot be justified, pursuant to the explicit project instruction. Enzyme, kinase, ATPase and cytokine partner classes are verified independently and do not transfer partner activities to Dicer. Missing full relevant evidence yields uncertainty rather than removal based on a title or abstract.
The expression of concern for PMID:25446899 is recorded as DISPUTED. Its exosomal Dicer–TRBP/AGO2 experiments remain unresolved despite independent ordinary loading-complex support; no retraction or pair-specific falsification is inferred. The separate pericardial-fluid vesicle study supports cargo localization without establishing extracellular catalysis. The SND1/TNRC6A/AGO2 multi-partner row remains unresolved because the relevant supplementary Dicer experiment was not recovered.
The initial draft treated the DNase and apoptotic-fragmentation IBA rows as unresolved because human activation experiments were lacking. Independent review corrected that interpretation: both functions are grounded by genuine worm fragment chemistry at PAINT node PTN001029377, the human target is a descendant, and no target-specific loss or faulty node placement was established. Both inherited assertions are retained non-core. Their human activation mechanism remains a research question; absence of a target experiment is not a phylogenetic refutation.
Other unresolved claims concern the specific mouse-to-human transcriptional-regulation scope, ATP binding, the complete PMID:17507929 partner set, the PMID:22105995 Dicer assay, and standalone siRNA binding versus TRBP-complex binding. They are not asserted false. The broad TNF-signaling row is refined to the experimentally measured TNF-production outcome; the original source fields remain unchanged.
Normal source and publication caches were not edited. Additional citation dependencies were fetched through the ordinary publication tool. Separate genuine originals supplement incomplete or abstract-only caches without changing their flags. The configured research provider returned authentication/quota failures; no substitute text was attributed to that provider. Short exact anchors occur once per source, with reference-only support elsewhere.
The final consistency pass aligns the MITF-expression-event cytosol row with the independently established gene-level cytosolic location, while retaining the output-versus-catalyst distinction. A structured consistency exception records why the negative mature-RISC release result and positive loading-associated RISC observation can coexist. The source-specific PMID:22105995 uncertainty remains explicit. Current-main source/runtime drift is separately qualified; no original evidence snapshot or source annotation is rewritten.
The broad GO:0051239 annotation remains MODIFY. Its replacements now include all four supported descendants already present in the review: GO:0010626 (negative regulation of Schwann cell proliferation), GO:0014040 (positive regulation of Schwann cell differentiation), GO:0031643 (positive regulation of myelination), and GO:0032720 (negative regulation of tumor necrosis factor production). Current QuickGO is_a/part_of paths were checked for each. GO:0032290, peripheral nervous system myelin formation, has no such path to GO:0051239 and remains a separate contextual annotation. Rat primary Schwann-cell experiments support proliferation and maturation effects (PMID:20648646(https://pmc.ncbi.nlm.nih.gov/articles/PMC3124220/), Figures 3–6 and Methods). The mouse studies retain their different ages and proliferation results (PMID:20463238(https://pmc.ncbi.nlm.nih.gov/articles/PMC6632556/), Results on conditional deletion and cell counts; PMID:20805985(https://pmc.ncbi.nlm.nih.gov/articles/PMC2929198/), Figures 1–3 and Discussion). Human Huh7 DICER knockdown, mature-miRNA measurements and miR-125b/miR-130a rescue support the TNF-regulatory context (PMID:26435691(https://pubmed.ncbi.nlm.nih.gov/26435691/), Figures 2–4 and cell-culture Methods). These are specificity refinements of the existing broad assertion, not new annotations or additional core functions.
The earlier TWEAK cytokine-binding refinement is superseded by KEEP_AS_NON_CORE for the original protein-association row. PMID:29657270(https://pubmed.ncbi.nlm.nih.gov/29657270/), Results on two-hybrid screening and coimmunoprecipitation, Discussion of the full-length negative result, and construct/IP Methods, establishes positive binding of the human Dicer C-terminal fragment and RNase-resistant recovery of cellular Dicer with expressed Flag-TWEAK. Full-length Dicer was negative in yeast two-hybrid; that result does not invalidate the positive observations. The cellular experiment does not exclude a protein bridge, and native access to transmembrane or soluble TWEAK remains unresolved. TWEAK's cytokine identity is sound, but the combined assay limits do not justify the more specific whole-product molecular-function assertion here. No receptor or signaling function is inferred. This qualification supersedes the earlier note's inclusion of TWEAK among established partner-class refinements.
The standalone description now uses “cytosolic Dicer” without duplicating cytoplasm/cytosol terminology; its nuclear and perinuclear context remains. All 140 imported annotation objects, three products, references and their cache flags, the single integrated core, and the four existing quotation anchors are unchanged. The action totals are 49 ACCEPT, 47 KEEP_AS_NON_CORE, 36 MODIFY and 8 UNDECIDED. No NEW or TOO_GENERAL action is introduced.