Consistency: Strong agreement across all sources. Deep research (falcon), review YAML, and PN annotations all converge on AUP1 as an ER/lipid-droplet monotopic membrane adaptor that recruits the E2 UBE2G2 (via G2BR) and binds ubiquitin (CUE) to drive ERAD, with secondary LD-clustering and (virus-induced) lipophagy roles. No contradictions. The three PN branch rows (ER/ALP/UPS) each map to a function the review independently accepts as core (ERAD, ubiquitin binding/E2 recruitment) or non-core (lipophagy).
PN story / NEW pressure: No NEW GO pressure. All three projected terms are already captured. GO:0036503 ERAD pathway and GO:0061724 lipophagy are present in GOA and ACCEPT/KEEP_AS_NON_CORE in the review (lipophagy correctly flagged virus-context, non-core). GO:0097466 ubiquitin-dependent glycoprotein ERAD pathway (verified real) is NOT in GOA and is more specific; the review does not list it but its parent GO:0036503 is accepted. Review's proposed_new_terms is empty — appropriate.
Mapping strategy: No change warranted. ERAD group=mapped/exact is well justified (AUP1 is a bona fide ERAD factor). The two propagation-scope mappings (lipophagy, glycoprotein-ERAD) are defensible at gene level for AUP1; the no_mapping/context_only calls on broad UPS/ALP ancestors are consistent with the TOMM20/HSPA8 "too broad" precedent.
Evidence alignment: PN row 2 cites Spandl 2011 (PMID:21127063), which is in the review (G2BR/Ube2g2 binding). PN otherwise cites few papers; the review is far richer (Klemm 2011 PMID:21857022, Jo 2013 PMID:23223569, Smith 2021 PMID:34879065, Frachon 2024 PMID:38474353, flavivirus PMID:29902443). No divergence — review is a strict superset.
Verdict: Consistent and well-curated; no NEW term, no mapping change. Optional: GO:0097466 could be added to the review as a more-specific ERAD term but is not required.