## FAF2
- **UniProt:** Q96CS3 (UBXD8/ETEA) · **batch:** proteostasis-batch-2026-06-11 · **review status:** COMPLETE (very thorough; UBA-UAS-UBX p97/VCP cofactor — ERAD + lipid-droplet/ATGL + stress-granule + pexophagy + ERMCS)
- **PN placement:** 5 rows — `ER proteostasis|...|ERAD|Retrotranslocation channel complex`; `Mitochondrial proteostasis|Organelle-specific protein degradation|mitoTAD pathway`; three UPS rows (`UBX domain|VCP adaptors`, `VCP and associated proteins|adaptors|UBX`, `Ubiquitin and UBL binding|VCP-associated adaptor`). **PN-node mapping:** ERAD group→GO:0036503 exact (in_goa); Mito-degradation class→GO:0035694 mitochondrial protein catabolic process (new_to_goa); the three UPS adaptor rows all no_mapping/context_only (GO:0019787, GO:0034098, GO:0043335, GO:0140036 as too_broad context only). Projected: GO:0036503 (in GOA), GO:0035694 (new).
- **Consistency:** Excellent. Deep research, review and PN concur on the core: membrane-tethered UBX p97/VCP substrate-recruiting cofactor (UBA binds ubiquitin, UBX docks p97) in SEL1L/HRD1 ERAD dislocation, plus lipid-droplet/ATGL inhibition, stress-granule disassembly, Insig-1/SREBP, and newer pexophagy (PMID:39472561) and ER-mitochondria contact-site (PMID:36746962) roles. The review correctly avoids `protein binding` as core, using GO:0030674 adaptor, GO:0043130 ubiquitin binding, GO:0034098 VCP-NPL4-UFD1 complex. No contradictions.
- **PN story / NEW pressure:** PN's only NEW projection is GO:0035694 mitochondrial protein catabolic process (verified real; NOT in GOA) from the mitoTAD-pathway row. FAF2/UBXD8 does participate in mitochondrial outer-membrane-associated degradation (mitoTAD) and p97-dependent extraction at mitochondria (PMID:41258083 "extract membrane proteins from the ER and mitochondria"), so this is a DEFENSIBLE (if non-core) ADD — but the review captures FAF2's degradation roles only as ERAD/retrotranslocation, not a mitochondrial-catabolism BP. Conclude: ERAD captured; GO:0035694 a defensible NEW (non-core), mildly broader than the review's ER-centric framing.
- **Mapping strategy:** Correct. All UBX/VCP-adaptor UPS rows are conservatively no_mapping/context_only (avoids the TOMM20/HSPA8-style over-broad propagation of GO:0043335 protein unfolding / GO:0034098 to a mere adaptor). ERAD group→GO:0036503 exact is right. Mito-class→GO:0035694 is the conservative shared target. No node-status change warranted.
- **Evidence alignment:** Good. PN UBX rows cite "18438607 / rev" and "18775313" — PMID:18775313 IS in the review (UBX/p97 cofactor family, HIGH). PMID:18438607 is NOT in the review (likely a UBX-domain review). The review's ERAD/LD/SG evidence (PMID:18711132, 23297223, 25660456, 34739333, + Falcon-added 26389662/35920641/36746962/39472561/41258083) is far richer than the PN rows.
- **Verdict:** Fully consistent; ERAD already captured, GO:0035694 mitochondrial protein catabolic process a defensible (non-core) NEW from the mitoTAD role. PN correctly declines to over-propagate the VCP-adaptor architecture rows. No YAML change strictly required.
- **Recommended edits:** [YAML] (optional) add GO:0035694 mitochondrial protein catabolic process (involved_in; mitoTAD/mito membrane-protein extraction, PMID:41258083) as a non-core BP to mirror the PN projection. [REF] PN-cited PMID:18438607 (UBX row) absent from review — verify whether gene-specific or a UBX-family review.