## DNAJC22
- **UniProt:** Q8N4W6 · **batch:** proteostasis-batch-2026-06-07b · **review status:** COMPLETE
- **PN placement:** `ER proteostasis | Chaperone | HSP70 system | J-domain containing HSP70 cochaperone` (branch ER) ; **PN-node mapping:** type=mapped, scope=ok_for_propagation_to_go, GO:0030544 Hsp70 protein binding (goa_status=more_specific_than_existing_goa)
- **Consistency:** Consistent at the inference level. Notes and review agree DNAJC22 is a poorly characterized polytopic membrane J-domain protein (liver-enriched) with NO direct human functional data; the co-chaperone role is inferred from the C-terminal J domain and the Drosophila ortholog Wurst (recruits Hsc70/clathrin for endocytosis). The PN node and review both treat it as a predicted HSP70 co-chaperone — no contradiction, but both rest on family/orthology inference, not experiment.
- **PN story / NEW pressure:** PN proposes GO:0030544 Hsp70 protein binding (verified real, OLS). The review's core MF is the parent GO:0031072 heat shock protein binding (also inference-only). GO:0030544 is a verified more-specific child of GO:0031072, so the projection is internally defensible and consistent in direction with the review — but it remains a domain/orthology inference (candidate, no experimental HSP70-binding data). Existing GOA has only membrane (IBA/IEA) and bare protein binding (MEOX2 Y2H), so an HSP70-binding term would be new — but should be ISS/family-level, not a confident annotation.
- **Mapping strategy:** Node heuristic applies cleanly here (genuine J domain). PN-projected GO:0030544 is appropriately specific (more specific than the review's GO:0031072) and the "more_specific_than_existing_goa" direction is at least internally consistent. Keep as inference-level propagation given the absence of any functional data; do not elevate to a confident new annotation.
- **Evidence alignment:** Minimal and aligned. Only literature reference is PMID:32296183 (HuRI Y2H, MEOX2 — relevance LOW, supplies the uninformative protein-binding term). Both review and PN lean on UniProt/Wurst orthology; no functional paper supports the Hsp70-binding claim directly.
- **Verdict:** Consistent family-level inference; GO:0030544 is a defensible more-specific candidate but unverified experimentally — flag as ISS/family inference rather than confident new annotation. **Recommended edits:** [MAP] retain GO:0030544 for DNAJC22 but mark as family/domain-level inference (no experimental HSP70 binding; Wurst-orthology basis only).