## VPS28
- **UniProt:** Q9UK41 · **batch:** proteostasis-pr-1217 · **review status:** COMPLETE
- **PN placement:** ALP `…|Sealing of autophagophore membrane|ESCRT-I complex component` ; **PN-node mapping:** leaf mapped/ok_for_propagation GO:0000813 ESCRT I complex (already_in_goa_exact); group mapped GO:0000045 autophagosome assembly (more_specific_than_existing_goa).
- **Consistency:** Strong agreement. Deep research/notes, review, and PN all treat VPS28 as a constitutive structural/scaffold ESCRT-I subunit acting in MVB sorting plus ESCRT-I-dependent phagophore/autophagosome closure. The review adds nuance the PN does not contradict: VPS28 itself does not bind ubiquitin (TSG101 UEV does), so it MARKs GO:0043130 over-annotated.
- **PN story / NEW pressure:** ESCRT-I membership (GO:0000813) is already_in_goa. For the autophagy role, GOA only has broad GO:0016236 macroautophagy; the review MODIFYs that to GO:0000045 autophagosome assembly AND adds GO:0000045 as a NEW annotation (IMP, PMID:32424346 helical-interface mutant → closure defect). This matches the PN projection exactly. Verdict: PN story is real and the review correctly adds the more specific GO:0000045 (new to direct GOA).
- **Mapping strategy:** VPS28 legitimately supports the ALP leaf (GO:0000813) and group (GO:0000045) projections — unlike UBAP1/UMAD1 it has direct VPS28 closure evidence (PMID:31519728, PMID:32424346). Scope ok_for_propagation is justified; PN-projected GO:0000045 is appropriately narrower than parent macroautophagy and aligns with the review.
- **Evidence alignment:** PN cites only PMID:32424346. Review's autophagy evidence pairs PMID:32424346 with PMID:31519728 (VPS37A directs ESCRT recruitment for phagophore closure) and PMID:20588296 — broader and concordant. ESCRT-I/MVB core uses PMID:11134028/11916981/18005716/20654576/21757351. Good overlap, review is richer.
- **Verdict:** Consistent and well-supported; the VPS28 review is the model case for this ESCRT-I leaf — GO:0000045 NEW is justified and matches the PN mapping.