Withdrew the unsupported nuclear-exclusion argument, retained unresolved actin IBA decisions with a participation-aware rationale, and incorporated the existing OpenScientist epithelial-function assessment. Recorded old GOA versus current PAINT snapshot discrepancy.
Evidence inspected: PMID:21743456(https://pubmed.ncbi.nlm.nih.gov/21743456/), PMID:21401524(https://pubmed.ncbi.nlm.nih.gov/21401524/).
Existing OpenScientist findings incorporated from openscientist.md; no duplicate question was submitted for that covered hypothesis.
Remaining questions:
The project audit records all changed row indices and decisions in projects/IBA_REVIEW/rereview-2026-09-20/localization.yaml.
Incorporated the generated nucleoplasm hypothesis report. It independently confirms absence of the historical nuclear call in the current GO snapshot, no HPA subcellular call, and positive epithelial membrane/vesicle and cochlear stereocilia localization. It provides a classical-NLS heuristic, whose negative result cannot exclude nonclassical or partner-mediated import. These are useful limits on support, but do not establish the reported strong refutation. The old IBA source is known as PTN001022094; absence from current sources is a version discrepancy, not evidence that the original record never existed. Nucleoplasm remains UNDECIDED.
The report also suggests that the clathrin-complex IDA from PMID:21401524 is a Dictyostelium projection. The cached abstract explicitly describes human BAIAP2L2 at clathrin-coated plaques, so this suggested removal is not adopted. No generated report text was changed and no duplicate run was launched.