DNAAF6 (HGNC:28570; UniProt Q9NQM4), formerly PIH1D3/CXorf41, was reviewed from a genuine 40-row normal bootstrap. Every source dictionary is preserved. The official product is 214 residues with no recorded alternative products; untranslated-exon variants and disease truncations were not converted into natural protein isoforms.
The final scientific draft has 9 ACCEPT, 13 KEEP_AS_NON_CORE and 18 MODIFY decisions, one integrated core and no NEW assertions. The core combines DNAI2 association with outer/inner dynein-arm assembly, supported by human Paff et al. and Olcese et al. and the actual mouse Dong et al. experiments. Reciprocal human co-IP does not establish a purified binding interface; residual arms and movement vary across affected individuals. The CS-like fold is predicted, and a full R2TP-like assembly architecture remains a model.
The precise human binding and location anchors are Figure 4 and its associated Methods in Paff, and Figure 5a,c,d with the accompanying Results in Olcese. Mouse Results, Figure 5 and co-IP Methods establish native Hsp70/Hsp90/Dnaic2 association and germ-cell localization. The mouse intermediate-chain complex still forms after knockout. These observations support positive chaperone association without a separate autonomous folding core. TGN refinements use the independent human TGN46 experiment while preserving the original HPA/SL-source resolution and additional cellular pools.
Sixteen screening-derived partner rows were checked against 49 exact positive interaction records and 98 participant reference-sequence checksums. The three literal partner isoforms were reconstructed independently. Eleven pairs receive supported binding-class refinements; five retain generic positive binding. Human DNAAF2/DNAAF4 association also remains non-core generic binding, while the DNAI2 pair is refined to the already supported specific intermediate-chain term. The explicit project instruction governs these retained positive annotations. No interaction was removed merely for genericity, unexpected cell context or lack of demonstrated endogenous physiology.
The genuine Falcon attempt failed authentication; the genuine fallback failed with an insufficient-quota response. No provider-named research text was fabricated. Manual primary-source work, official database records and the independent core consultation supplied the evidence. Protected main publication caches remain byte-identical; genuine additive caches retain their fetched metadata and content. The evidence guide gives public source links, assay boundaries, propagation grounding and access limits. Short exact anchors appear once in the YAML; this journal and the guide do not repeat them.
This entry supersedes the earlier POLDIP2 comparator and count summary without changing the 40 imported source assertions. The current decisions are 9 ACCEPT, 19 MODIFY and 12 KEEP_AS_NON_CORE, with one integrated core and no NEW annotations. Four non-core propagation diagnoses now use NO_FAILURE_NON_CORE; the experimentally grounded transfer itself remains supported.
The eleven challenged screen-derived binding refinements remain supported by the positive, sequence-resolved target–partner assays and independently established partner classes. A binding-class term does not transfer the partner's catalytic or regulatory activity. Native coexpression, channel topology and physiological interaction remain unestablished where the screen does not resolve them. The interaction audit now reproduces the 16 source groups, 49 selected records and 98 reference-sequence checks from unchanged raw inputs, with stage and fusion-feature limits.
The earlier polymerase-only account of POLDIP2 omitted its experimentally established cytoskeletal-regulatory role. PMID:30354218, Figure 6 directly tests endogenous POLDIP2 perturbation in human aortic smooth-muscle cells. Its actin–vinculin and focal-adhesion effects justify GO:0005519 for the existing positive DNAAF6–POLDIP2 binding row, without asserting that DNAAF6 performs those processes. The current official term is nonobsolete. Q8IY57-5 is officially named YAF2 isoform 2; its source accession is unchanged. The altered ARID4B-3 C terminus leaves its more specific partner role unresolved, without negating the measured interaction.
The core now mentions observed human HSP90-fragment binding and native mouse chaperone association. Chaperone recruitment and complete maturation-complex architecture remain models; autonomous DNAAF6 folding or ATPase activity is not asserted. Two new short primary anchors occur once in the YAML, alongside the two unchanged inherited excerpts. The evidence guide retains public source and assay locators. Existing publication and provider bytes, product identity and prior history remain unchanged; the new POLDIP2 publication cache was genuinely fetched.
The human-cell attribution for the POLDIP2 experiments is retained on the separately recovered original Figure 6 caption, together with the Methods definition of HASMs. The normal publication cache omits those captions despite its unchanged full-text metadata. The evidence guide now gives the precise anchors and two short excerpts; no figure-pixel analysis is claimed.
POLDIP2 acts upstream through NOX4 in these actin–vinculin and focal-adhesion experiments; direct actin crosslinking by POLDIP2 was not demonstrated. That distinction qualifies the partner class without changing the positive DNAAF6 binding evidence or transferring the pathway to DNAAF6. All 40 imported source objects and actions, the product, core and prior quotations remain unchanged. The interaction audit retains its raw inputs, script and executed results; its README corrects the literal partner name to PID1-3.