## LMAN2
- **UniProt:** Q12907 (Vesicular integral-membrane protein VIP36) · **batch:** proteostasis-batch-2026-06-11 · **review status:** COMPLETE, thorough (detailed notes + Falcon).
- **PN placement:** `ER proteostasis|Glycoproteostasis|N-glycosylation system|Lectin chaperone` ; **PN-node mapping:** type "Lectin chaperone" no_mapping; group "N-glycosylation system" mapped→GO:0006487 (protein N-linked glycosylation, ok_for_propagation, new_to_goa); class/branch no_mapping.
- **Consistency:** Deep research ↔ review ↔ PN annotation consistent. LMAN2/VIP36 is correctly an L-type (ConA-like) lectin cargo/sorting receptor (D-mannose binding GO:0005537, carbohydrate binding GO:0030246) — NOT a mannosidase, matching the PN "Lectin chaperone" type. Distinctive feature well captured: retrograde Golgi→ER recycling (GO:0006890) in post-ER QC of alpha1-antitrypsin, plus a secondary shed-ectodomain/phagocytosis role kept non-core.
- **PN story / NEW pressure:** No NEW GO pressure; functions are captured (IMP D-mannose binding, retrograde transport, ER-Golgi cycling). Falcon shedding cleavage-site detail (PMID:33796845) refines the already-non-core shedding role only. Conclude: already captured.
- **Mapping strategy:** Same over-reach as LMAN1 — **group→GO:0006487 over-reaches for VIP36**, a glycan-reader/sorter that neither installs nor trims N-glycans. `new_to_goa` would introduce an unsupported biosynthesis assertion. The review's own BP terms (GO:0006888, GO:0006890) are transport/QC, conceptually disjoint from GO:0006487. PN node correctly leaves "Lectin chaperone" unmapped; the inherited group term should not propagate.
- **Evidence alignment:** PN dossier is mapping-only (no titles). Review evidence PubMed-verified and centered on PMID:20477988 (the key functional paper), 23701871, 22016386, 15308636. No bibliographic conflict; divergence is conceptual (group term vs. lectin function).
- **Verdict:** Review strong and PN-consistent (lectin, not mannosidase; distinct retrograde-QC angle). Inherited group→GO:0006487 projection is an over-reach for VIP36.
**Recommended edits:** [MAP] Do not propagate GO:0006487 (protein N-linked glycosylation) to LMAN2/VIP36 — it is a high-mannose glycan-reading sorting receptor (transport/retrograde QC), not an N-linked-glycosylation enzyme; flag the "N-glycosylation system" group node as over-broad.
