All 34 fetched GOA assertions were reviewed individually, retaining every source field and all five official product identities and sequence notes. The single core integrates cytosolic 3′–5′ RNA exonucleolysis, selected mRNA turnover, RNA recognition and divalent-metal dependence. It does not split metal binding or structured-RNA processing into unsupported separate activities.
The source guide records public original locators, assay organisms and constructs, precursor versus mature-miRNA substrates, and the actual PAINT nodes. Human soluble cytoplasmic localization is distinguished from fission-yeast P-body context. Positive RNA-dependent XRN1 association is retained as non-core under the project's instruction. Direct human precursor-RNA evidence is distinguished from fly mature-miRNA grounding of the valid phylogenetic assertion.
The specific PMID:22306653 sister-chromatid-separation assay remains unverified because its full original was inaccessible; the explicit abstract and curator-supported broader effects are retained with appropriate scope. Normal publication-cache access flags remain unchanged even where an original was separately recovered. The human 885-residue construct is distinguished from the four shorter natural products, for which no isoform-resolved functional equivalence or inactivity is asserted.
Normal gene fetch and publication caching succeeded. Falcon authentication failed and its configured perplexity-lite fallback returned an insufficient-quota error. No provider report was fabricated; the evidence review was performed manually. No new annotations or functional-isoform classes were introduced.
The retained XRN1 protein-binding row follows the explicit project instruction; the RNA-dependent association does not justify an invented direct-binding mechanism. The mature-miRNA rows intentionally retain different actions because the fly-grounded phylogenetic assertion and pooled gene-level claim have evidence that the human precursor experiment and mouse-only transfer do not supply. The source guide identifies those assays separately. These are deliberate biological qualifications, not unresolved formatting defects.
PMID:27974621 Figure 3B compares mature miR-1010 with precursor RNA using wild-type Drosophila Dis3L2 immunoprecipitates. The catalytic-dead comparison is the separate Figure 3A experiment with an artificial 77-nt substrate. The rationale and source guide distinguish these controls explicitly; the phylogenetic annotation decision is unchanged.
The four miRNA-catabolism annotations now share KEEP_AS_NON_CORE, judging the combined gene-level claim. The valid phylogenetic assertion and fly mature-miRNA assay support that claim; human and mouse precursor experiments retain their stated limits. This supersedes the earlier validation paragraph's defense of differing actions. The two broader RNA-catabolism replacements are removed because they are ancestors of the retained miRNA process.
Core representation now includes polyuridylation-dependent mRNA catabolism and explicitly represents oligo(U) recognition as a coupled substrate-selection component. Human U30 mobility-shift binding, performed without added magnesium, and catalytically inactive substrate trapping distinguish binding from cleavage. The two core entries describe coupled molecular activities, not independent pathways. Magnesium binding remains a supported non-core cofactor property. The description retains biology; product counts and assay-product scope remain in the product section.
Primary sources are now linked directly from the annotation support lists as well as through the prose guide. The separately read publisher original for PMID:23756462 remains distinct from its abstract-only normal cache; the access flag is not falsified. Existing quotations are unchanged and are not duplicated. No source annotation, reference identity, product identity or new annotation is introduced.