## SERP2
- **UniProt:** Q8N6R1 · **batch:** proteostasis-batch-2026-06-11 · **review status:** COMPLETE
- **PN placement:** `ER proteostasis|Protein transport|SEC61 channel accessory protein` ; **PN-node mapping:** group "SEC61 channel accessory protein"=no_mapping (broad category); parent class "Protein transport"=mapped/ok GO:0015031 protein transport; branch=no_mapping.
- **Consistency:** Deep research ↔ review YAML ↔ PN annotation consistent. All describe SERP2/RAMP4-2 as a single-pass ER membrane RAMP4-family paralog of SERP1, translocon-associated, stabilizing nascent membrane/secretory substrates during ER stress; function established largely by orthology/paralogy (less experimentally characterized than SERP1). No contradictions. The review adds a SERP2-specific experimental fact the PN row omits: RAMP4-2 is a tail-anchored substrate of the SPPL2c/SPP intramembrane proteases (PMID:30733280, PMID:37261541).
- **PN story / NEW pressure:** PN asserts only the unmapped accessory-protein grouping. SERP2's core functions — GO:0030968 ER unfolded protein response (IBA) and ER-membrane localization — are already in GOA/review. No defensible NEW GO term; review proposes none. Already captured. (The SPPL2c-substrate property is interesting but the review correctly added it as references only, no new GO assertion.)
- **Mapping strategy:** Concern — parent **class** node projects **GO:0015031 protein transport** as `new_to_goa`. As for SERP1, SERP2 is a translocon-associated stabilizer, not a transport carrier; GO:0015031 over-reaches (TOMM20/HSPA8/RAB7A precedent). The group-level no_mapping is correct; do not propagate class-level GO:0015031 to SERP2.
- **Evidence alignment:** PN row is terse; review references go well beyond it (SPPL2c substrate papers, translocon Insight PMID:38787756, Gemmer & Forster review PMID:32019826). Overlap is essentially at the family/orthology level since both rely on RAMP4-family inference rather than SERP2-specific GOA experimental annotations (GOA has only IBA + 2 bare protein-binding IPI).
- **Verdict:** Consistent and well-curated. Recommend NOT propagating class-level GO:0015031 protein transport to SERP2 (over-reach); UPR + ER-membrane already captured.