PN placement: ALP …|Sealing of autophagophore membrane|ESCRT-I complex component + UPS …|trafficking|ESCRT-I complex|UBA (UBAP1) ; PN-node mapping: leaf mapped/ok_for_propagation GO:0000813 ESCRT I complex; group mapped GO:0000045 autophagosome assembly; UPS nodes no_mapping/context_only.
Consistency: Deep research (notes), review YAML, and PN annotation agree that UBAP1 is the endosome-specific fourth ESCRT-I subunit (TSG101-VPS28-VPS37A-UBAP1) doing ubiquitin binding + MVB cargo sorting. One internal tension: the PN places UBAP1 under autophagophore-sealing, but the review deliberately argues the autophagy/membrane-fission evidence (PMID:32424346) used VPS37B-MVB12A ESCRT-I, not UBAP1, and so MARK_AS_OVER_ANNOTATED membrane fission. Defensible and well-sourced.
PN story / NEW pressure: ESCRT-I membership (GO:0000813), ubiquitin binding (GO:0043130), and MVB catabolism (GO:0043162) are all already captured. The PN-projected GO:0000045 autophagosome assembly is new_to_goa but the review does NOT add it for UBAP1 (no UBAP1-specific phagophore-closure evidence). Verdict: already captured for the endosomal role; autophagosome-assembly over-reaches for UBAP1.
Mapping strategy:UBAP1 does not justify the ALP group's GO:0000045 projection — the review's evidence argument is that autophagic ESCRT-I uses MVB12A/VPS37B not UBAP1. PN-projected GO:0000045 is therefore broader/mis-attributed for this gene (TOMM20/HSPA8-style over-reach). GO:0000813 leaf mapping is correct.
Evidence alignment: PN cites only PMID:32424346 (the membrane-scission structure paper). Review's core ESCRT-I evidence is PMID:21757351 (UBAP1 endosome-specific ESCRT-I) + PMID:22405001 (SOUBA), and it treats PMID:32424346 as the over-annotation source — direct divergence in how the shared paper is used.
Verdict: Internally consistent and defensible; flag that the PN autophagosome-sealing placement / GO:0000045 projection is not supported by UBAP1-specific evidence (review correctly withholds it).