## EMC7
- **UniProt:** Q9NPA0 · **batch:** proteostasis-batch-2026-06-11 · **review status:** COMPLETE
- **PN placement:** `ER proteostasis|Protein transport|Transmembrane protein import|EMC complex component` ; **PN-node mapping:** type=mapped/ok_for_propagation → GO:0072546 (EMC complex); group → GO:0044743 (protein transmembrane import into intracellular organelle); class → GO:0015031 (protein transport); branch=no_mapping.
- **Consistency:** Deep research (notes + falcon), review YAML, PN annotation, and the type-level mapping are mutually consistent. EMC7 is a single-pass type I, non-catalytic architectural EMC subunit (lumenal beta-sandwich; catalytic vestibule = EMC3/EMC6). No contradictions. The review additionally documents EMC7-specific roles (Rab7/syntaxin-18 ER–endosome tethering, PMID:32111841; VDAC1 interface, PMID:38517390; SQS-client selectivity filter, PMID:37199759) absent from the dossier, but these elaborate rather than conflict.
- **PN story / NEW pressure:** PN asserts only EMC membership + ER membrane-protein import/insertion, all already captured (GO:0072546 part_of; GO:0045050, GO:0071816, GO:0032977 contributes_to). No NEW GO term needed. The carbohydrate-binding IEA (GO:0030246) is correctly MARK_AS_OVER_ANNOTATED. EMC7's tethering/VDAC moonlighting is not pushed as a PN role.
- **Mapping strategy:** EMC7 does not change the shared EMC-node mapping. Type→GO:0072546 is exact and correct. The **group→GO:0044743 is semantically off**: GO:0044743 (def. "directed movement of proteins into an intracellular organelle, across a membrane") is lumenal import (TOM/TIM-type), whereas the EMC performs membrane-protein *insertion*; GO:0071816/GO:0045050 (via GO:0090150 establishment of protein localization to membrane) are NOT subclasses of GO:0044743. GO:0044743 over-reaches/diverges.
- **Evidence alignment:** Strong overlap — both cite the EMC insertase/structure literature; the review's PMIDs (22119785, 29242231, 32439656, 30415835, etc.) fully cover the PN process claim.
- **Verdict:** Consistent; well-reviewed. Sole concern is the group-level GO:0044743 import mapping (shared across EMC7-10), which mismatches EMC "insertion" semantics.