All 42 rows were reassessed. PRC2-associated CC assertions are retained with context-specific membership, while the description no longer calls FVE an invariant core subunit or implies that one fixed HDA5/HDA6/FLD assembly explains all activity. Both epigenetic-regulation rows now share the same directional refinement. The mitochondrial HDA row retains OVER only after inspecting all seven deposited spreadsheets: exact FVE entries say nucleus and the mixed-microsome experiment does not establish mitochondrial residence. This supersedes the old absence-of-targeting-signal rationale. Cytoplasmic/cytosolic and nucleolar experimental reports are retained separately with source-access limits; nuclear function is not used to reject them. DA photoperiodism remains unresolved, not dismissed by the label autonomous. See FVE-source-analysis for checksummed extraction. No new annotations were added.
All seeded rows were reviewed, with original term, evidence, source, supporting-entity and qualifier/NOT payloads preserved. The shared normal Falcon attempt and perplexity-lite fallback both failed; logs are retained under .cache/module-audit-logs/autonomous-flowering-fetch/ARATH-FVE-provider.log. No provider report was fabricated. Manual primary-literature assessment and normal publication fetching supplied this review. Source access limits are stated per reference/claim. Generic protein binding follows the current informative-function policy; its removal does not reject the physical interactions. Nuclear localization does not by itself exclude additional compartments.
The final source check supersedes the earlier unverified retention rationale for the cytoplasm/cytosol rows: both are now UNDECIDED because their exact target image/table could not be recovered. The nucleolar row remains non-core on independent direct target corroboration in PMID:15608277 Figure 1 (NU:no:NAS), not merely nuclear plausibility. Original nuclear assertions are independently supported and retained. No experimental localization is removed for incomplete access.
Corrected citation spacing in reviewer-authored prose without altering original annotation payloads, source quotations, raw source files or prior histories. Reconsidered the reviewer suggestion to retain cytoplasm/cytosol HDA claims. The original target-level evidence remains unavailable: PMID:15610358 is abstract-only, while PMID:25293756 supplementary retrieval failed as recorded in localization-25293756.json. Both source claims and payloads are preserved with separate UNDECIDED reasons; neither is removed, and no cytoplasmic exclusion follows from nuclear localization. No blanket action reversal was warranted without new evidence.