## AMBRA1
- **UniProt:** Q9C0C7 · **batch:** proteostasis-batch-2026-06-03 · **review status:** COMPLETE
- **PN placement:** three rows (branches ALP+UPS) — Class-3 PI3K complex-1 / BECN1-localization modulator; `Chaperone-mediated autophagy|...|CMA enhancer|Enhancer of substrate uptake`; and `UPS|E3 ubiquitin and UBL ligases|Cul4A/Cul4B substrate receptor|WD40|other` ; **PN-node mapping:** PI3K subtree context_only (too_broad, GO:0035032 / GO:0016236, no CC/process projection); CMA-enhancer *type* mapped→GO:1904716 (positive regulation of CMA); CRL4-receptor *group* mapped→GO:1990756 (UBL ligase-substrate adaptor). Projections: **GO:1904716 more_specific_than_existing_goa**, **GO:1990756 already_in_goa_exact**.
- **Consistency:** Largely consistent on the UPS/autophagy axes; one divergence on CMA. CRL4 adaptor (GO:1990756) and autophagy/mitophagy roles match across deep research, review, PN, and mapping. The **CMA-enhancer projection (GO:1904716) is NOT adopted** by the review: it has no proposed_new_terms, and explicitly asks (suggested_question/experiment) whether AMBRA1 should be excluded from GO:1904716 until direct LAMP2A/HSPA8 substrate-uptake evidence exists. Review notes correctly state the AMBRA1 CMA-enhancer **leaf (subtype) is no_mapping**, but the dossier shows the parent *type* is mapped and projects GO:1904716 — so the projection still fires.
- **PN story / NEW pressure:** Two PN MFs. GO:1990756 — already in GOA, abundant direct IDA support (PMID:23524951, 33854232/35/39, 30166453) → already captured (ACCEPT). GO:1904716 (verified real) — asserted via the CMA-enhancer bucket but supported only by the "CMA prevents neuronal proteome collapse" background paper, not AMBRA1-specific CMA substrate-uptake data. Verdict: GO:1990756 already captured; GO:1904716 over-reaches as a propagated annotation (candidate pending direct evidence).
- **Mapping strategy:** PI3K subtree context_only is correctly conservative (no spurious GO:0035032 complex membership — avoids the TOMM20/RAB7A broader-than-review failure mode). The CMA-enhancer *type*→GO:1904716 mapping is the weak point: it propagates a positive-CMA process to AMBRA1 on thin evidence; review withholds it. CRL4 group→GO:1990756 is appropriately specific.
- **Evidence alignment:** UPS row PN ref PMID:33854239 = review IDA reference PMID:33854239 (CRL4 cyclin-D) — strong overlap. CMA/autophagy PN titles (AMBRA1 network review; mTOR-ULK1-AMBRA1-TRAF6; CMA proteome-collapse) are background; review's autophagy/mitophagy calls rest on dedicated PMIDs (20921139, 21358617, 23524951, 25215947, 30217973).
- **Verdict:** GO:1990756 consistent/already captured; GO:1904716 CMA-enhancer projection over-reaches (review withholds pending direct CMA evidence).
- **Recommended edits:** [MAP] Downgrade the CMA-enhancer *type*→GO:1904716 to context_only (or mark AMBRA1 a gene-level exclusion) so positive-CMA is not auto-projected without LAMP2A/HSPA8 substrate-uptake evidence, consistent with the review's withholding.