The current assessment supersedes earlier blanket interaction removals while preserving the original source assertions and notebook. All 159 imported assertions remain exact: 75 ACCEPT, 54 KEEP_AS_NON_CORE, 28 MODIFY and two UNDECIDED; no NEW rows or source retirement were introduced. The 56 generic-binding sources were individually reassessed: 20 have supported partner-class refinements and 36 retain positive non-core associations. The project-specific generic-binding policy in CLINGEN_MENDELIAN retains supported GO:0005515 when no informative refinement is established.
Three repeated motor-core entries are consolidated into one minus-end motor mechanism with transport, nuclear movement and mitotic roles. The four process terms were checked for typed ancestor redundancy. Nuclear movement includes mouse dynein-complex perturbation, and checkpoint silencing includes LIC1 perturbation; neither is described as a directly measured human heavy-chain-specific experiment. All 185 inherited annotation quotation occurrences are preserved. The 15 excerpts attached to the replaced redundant core entries are removed with explicit scientific approval; their references and original source caches remain. No new excerpts are introduced.
The two unresolved claims concern a source-specific filopodium observation and granule-lumen attribution. Positive RNA-capture, extracellular-vesicle, membrane-associated, meiotic nuclear-envelope and stress-response observations are retained with assay and organism limits. A validated partner class does not transfer its catalytic activity or physiological role to DYNC1H1. The PSEN2 co-IP band-size discrepancy, recorded interaction fragments and inferred ATP23 class remain explicit.
The official record contains one 4,646-residue Q14204 product and no alternative-product block. Historical experimental inserts, host species and reference-sequence checksums do not independently authenticate a natural product. Current PANTHER data place the exact target in PTHR10676; the older family pointer is preserved as historical input. Literal IBD placement and target descent support all nine inherited assertions.
Existing machine inputs, seven histories, provider artifacts and all preexisting main publication caches remain byte-identical. Genuine normal-fetch variants are retained privately and are not substituted for protected caches. The configured provider attempt and fallback failed; no provider text was fabricated. New publication additions are restricted to genuinely fetched references needed by the reviewed evidence.
See the source evidence guide and portable source-record package for public locators, original-access limits, measurements and reproducible extraction. These authored files organize evidence and are not cited as independent primary experiments.
The PSEN2 row is now UNDECIDED because the identity of the reported 140-kDa species remains unresolved; the positive experiment is preserved without alleging a wrong-gene result. The 15 source-specific PMID:32814053 refinements remain bounded partner-class binding claims, with independently cited functional classifications and explicit isoform/fragment and ATP23 inference limits. The original screen included pairwise retesting; its orthogonal validation of a subset is not assigned to every DYNC1H1 pair.
The single motor core now represents retrograde axonal transport rather than its broad ancestor, and includes spindle positioning alongside spindle organization. Current typed ancestry finds no redundant pair among the five processes. Human heavy-chain siRNA/recruitment evidence supports positioning; the checkpoint mechanism remains explicitly LIC1-dependent dynein-complex evidence. These changes do not add annotation rows.
Local reference-only links expose the already derived source records and their independent classification evidence. The displayed 4,646-residue product is now explicit. Citation verification for three incompletely recovered studies is scoped to identity and the inspected portions. All 159 imported source objects and all 185 annotation quotation occurrences remain exact; no new quotation is introduced. The current action totals are 75 ACCEPT, 54 KEEP_AS_NON_CORE, 27 MODIFY and three UNDECIDED.