FCA (At4g16280, UniProt O04425) curation notes

2026-10 review session (autonomous pathway module)

• Accession verified: UniProt O04425 FCA_ARATH, At4g16280 (gene_exact search returned a single reviewed entry).
• Deep research (falcon) failed (HTTP 429); review built from cached primary literature.
• FCA: two RRMs + WW domain; nuclear RNA-binding protein PMID:12805228
• WW domain binds FY (Pfs2p/WDR33) PMID:12809608
• Autoregulation by proximal polyadenylation in intron 3 PMID:12805228
• FLC repression needs FLD (LSD1 homolog) PMID:17996704
• Proximal antisense (COOLAIR) polyadenylation triggers silencing PMID:19965720
• FCA binds nascent COOLAIR and interacts with CLF PMID:31032401
• Liquid-like nuclear bodies, FLL2 PMID:31043738
• FCA does NOT bind ABA PMID:19078995.

Decisions

• NEW: lncRNA binding (GO:0106222, COOLAIR) and mRNA alternative polyadenylation (GO:0110104). FCA does the step (binds RNA and recruits FY/CPSF to the proximal site), so the participation test is met.
• Cytoplasm IBA/ISM and RNP complex IBA (PABP-seeded node) flagged as over-annotation/removal.
• Bare protein-binding rows removed (interactions themselves not disputed).

2026-10-26 autonomous-flowering RNA-factor audit

All rows were reviewed, preserving original terms, evidence, references, qualifiers, NOT flags and supporting entities. Falcon/fallback research and gene-publication retrieval ran concurrently under the parent workflow; provider attempts failed and logs remain under .cache/module-audit-logs/autonomous-flowering-fetch/. No provider report was fabricated. Primary source caches and curated UniProt evidence were inspected independently. Annotation-reviewer consultation with the chromatin-panel reviewer agreed that nuclear predominance is not organelle exclusion, broad RNP membership does not require PABP identity, and FPA termination reinterpretations are locus-specific.

The authoritative PAINT inspection is reproducible in genes/ARATH/FCA/FCA-bioinformatics/. It uses exact existing PTNs and is not a reconstructed family phylogeny. The failed full-family download was not used; only its incomplete .part file was removed after disk exhaustion. Current source retrieval is recorded honestly: PMID:18223036 refreshed complete; PMID:19439664 cache abstract-only but full web text inspected; 16820865 and PMID:19561057 recovered normally; optional PMID:16753815 fetch returned 429 and was not used to support a new claim.

• Cytoplasm IBA and AtSubP rows are now UNDECIDED: predominantly nuclear FCA:GUS (PMID:12805228) and native nuclear bodies do not prove absence of a cytoplasmic population. PTN000609546 was recovered from authoritative IBD.gaf; no node-placement defect was demonstrated.
• Broad RNP-complex claim now ACCEPT: GO:1990904 requires RNA plus protein, not PABP activity or permanent CPSF membership. PMID:31032401 directly demonstrates FCA–COOLAIR association. PTN002768817 was inspected; prior unsupported PROPAGATION_BAD metadata was replaced.
• PMID:31043738 shows FLL1/FLL3 binding by yeast two-hybrid (Extended Data5d), while FLL2 interaction uses crosslinking/phase separation. The source-specific protein-binding row now cites the actual partners.
• NOT GO:0010427 ABA binding was explicitly audited and retained ACCEPT based on PMID:19078995. No sign inversion or positive ABA-receptor assertion.
• Existing two NEW annotations are retained; no new assertions added. The mRNA-binding IBA remains an inherited biochemical inference, not a claim that autoregulation alone proves direct binding to each mRNA.

PR #5574 prose and justification follow-up

Corrected citation spacing in reviewer-authored prose without altering original annotation payloads, source quotations, raw source files or prior histories. Biological decisions, annotation actions and source-access limits are unchanged.