## CYB5R4
- **UniProt:** Q7L1T6 · **batch:** proteostasis-batch-2026-06-07 · **review status:** COMPLETE
- **PN placement:** `Cytonuclear proteostasis | Chaperone | HSP90 system | HSP90 cochaperone | CS domain containing` ; **PN-node mapping:** type (HSP90 cochaperone)=mapped, scope=ok_for_propagation_to_go, GO:0051879 Hsp90 protein binding (subtype/group/class/branch = no_mapping)
- **Consistency:** CONTRADICTION. The review's entire biology is an NAD(P)H-cytochrome b5 reductase / soluble ER flavohemoprotein (EC 1.6.2.2; heme + FAD binding; ROS/ER-stress protection). It contains an N-terminal CS/Hsp20-like (p23_NCB5OR) domain — the basis for the PN "HSP90 cochaperone, CS domain containing" placement — but the review, GOA, and notes contain ZERO Hsp90-binding or cochaperone evidence; the notes explicitly raise "how does the CS domain contribute to function" as an open question. The PN classification is a domain-architecture heuristic, not evidence.
- **PN story / NEW pressure:** PN asserts an HSP90-cochaperone role absent from existing GO. GO:0051879 Hsp90 protein binding is a real term (verified OLS), but there is no experimental support that CYB5R4 binds Hsp90; the InterPro "HSP20-like chaperone" CS-domain signature is structural homology only. Conclusion: **over-reaches** — propagating GO:0051879 onto CYB5R4 would be a speculative domain-based annotation contradicting the curated redox-enzyme function. Not a defensible ADD.
- **Mapping strategy:** The HSP90-cochaperone type→GO:0051879 mapping may be reasonable for canonical CS-domain HSP90 cochaperones, but CYB5R4 is a poor exemplar (a redox enzyme that happens to carry a p23-like domain). The node mapping itself is not necessarily wrong family-wide, but this gene should NOT inherit GO:0051879. Recommend de-coupling CYB5R4 from the HSP90-cochaperone propagation (treat the CS domain as a non-canonical/standalone module).
- **Evidence alignment:** No overlap. PN dossier lists no reference titles; review evidence (PMID:10611283, PMID:15131110) is entirely redox/localization. No paper supports the Hsp90 link.
- **Verdict:** Inconsistent — PN HSP90-cochaperone placement / GO:0051879 over-reaches and is unsupported by the review. **Recommended edits:** [MAP] exclude CYB5R4 from GO:0051879 Hsp90 protein binding propagation (CS/p23 domain is structural-homology only; no Hsp90-binding evidence); flag the PN "HSP90 cochaperone" classification as domain-based, not functional. [REF] none.