## MMGT1
- **UniProt:** Q8N4V1 (EMC5; legacy "membrane magnesium transporter 1" / TMEM32) · **batch:** proteostasis-batch-2026-06-11 · **review status:** COMPLETE (very thorough; ~40 annotations including a large block of legacy transporter terms all adjudicated)
- **PN placement:** `ER proteostasis | Protein transport | Transmembrane protein import | EMC complex component`; **PN-node mapping:** type=mapped/ok_for_propagation → GO:0072546 EMC complex (already_in_goa_exact); group→GO:0044743 (goa_status=more_specific_than_existing_goa); class→GO:0015031 (new_to_goa); branch=no_mapping.
- **Consistency:** Mostly consistent, but MMGT1 is the gene where the legacy-name trap is live. Review, deep research, and PN annotation all treat MMGT1 as EMC5 (an EMC insertase subunit), and the review systematically MARK_AS_OVER_ANNOTATED the legacy metal-transport terms (GO:0015095 Mg, GO:0006824 Co, GO:0034755 Fe, GO:0022857, GO:0055085, GO:1903830, etc.) derived from rodent-ortholog By-similarity. GO:0072546 + ER membrane + insertion BPs captured. No contradiction with PN, which places it firmly as an EMC component.
- **PN story / NEW pressure:** PN asserts only EMC-component/import role — no new pressure; fully captured by GO:0072546 and the insertion BPs. Conclusion: already captured.
- **Mapping strategy:** Subtle. The dossier flags group→GO:0044743 as `more_specific_than_existing_goa`. MMGT1's existing GOA "transport" terms are GO:0055085 (transmembrane transport, IEA) and metal-transport terms — which the review is actively flagging as OVER-ANNOTATED (legacy magnesium-transporter artifact, not protein import). So the PN projection (GO:0044743 protein transmembrane *import*) is conceptually correct (insertion) but the `more_specific_than_existing_goa` comparison is being made against a transport term the review rejects, not against a genuine import annotation — the framing is misleading. type→GO:0072546 is exact/correct. Projected GO:0044743/GO:0015031 are broader than the review's specific insertion terms (broader-ancestor pattern, cf. TOMM20/HSPA8/RAB7A).
- **Evidence alignment:** Good overlap on EMC papers (PMID:22119785, PMID:29242231, PMID:38517390 — EMC5 used as Strep handle). Review adds the MMGT1-specific MtB persistence/lipid-droplet axis PMID:37269834 (mechanism unresolved, kept as context). PN cites no row-1 titles; no divergence.
- **Verdict:** Consistent on EMC identity; legacy Mg/metal-transporter terms correctly over-annotated; PN adds no NEW pressure; note the misleading `more_specific_than_existing_goa` framing (compared against a rejected transport term).