ATG2A notes

• Fetched human ATG2A with just fetch-gene human ATG2A on 2026-06-03. The scaffold seeded 30 review entries from 36 GOA rows because repeated protein binding rows were collapsed by publication in the generated review file.
• Falcon deep research was started with fallback using just deep-research-falcon human ATG2A --fallback perplexity-lite; PMID caching completed with all eight review PMIDs present locally. Falcon timed out after 600 seconds, then the perplexity-lite fallback failed with an API quota 401, so no provider deep-research file was produced in this run.
• Core function: ATG2A is best represented as a lipid-transfer/tethering factor for autophagosome assembly. Valverde et al. show that "ATG2A can bind tens of glycerophospholipids at once and transfers lipids robustly in vitro" and that ATG2-mediated lipid transfer at contact sites contributes to autophagosome formation PMID:30952800.
• Maeda et al. directly state that "human ATG2A is a lipid transfer protein" and place the "ATG2-WIPI complex between the ER and the phagophore edge" PMID:31271352.
• The PN projection for ATG2A is GO:0062079 ATG2-ATG18 complex from the PN path Autophagy-Lysosome Pathway|Autophagophore initiation and elongation|Regulation of autophagophore membrane composition|ATG2-WIPI complex component; the mapping rationale says "The GO ATG2-ATG18 complex term is the closest component-level target" [file:projects/PROTEOSTASIS/mappings/autophagy_lysosome_pathway.yaml].
• I accepted the PN-projected complex conservatively as a new cellular-component annotation, supported by ATG2A-WIPI evidence, but did not use the PN projection to broaden cargo-specific process annotations. The IBA cargo-specific rows for mitophagy, pexophagy, glycophagy, reticulophagy, and piecemeal microautophagy of the nucleus were modified toward general autophagosome assembly unless direct ATG2A cargo-specific evidence was available.
• Generic GO:0005515 protein binding rows were marked over-annotated, including WIPI-binding rows whose specific complex context is instead captured by the separate NEW GO:0062079 ATG2-ATG18 complex annotation. This follows the project guideline to avoid retaining protein binding when a more informative function or component term is available.

Falcon deep research findings (2026-06-07)

Synthesis of the Falcon (Edison Scientific) report against the existing COMPLETE review. PMIDs resolved from DOIs via PubMed ID-conversion. Only Wei 2024 (ANKFY1) is cached locally in publications/; the others are full_text_unavailable.

• CONFIRMS (no annotation change needed): Core paradigm is unchanged — ATG2A is a rod-like (~20 nm) RBG/bridge-like bulk lipid-transfer protein with an N-terminal chorein lipid-transfer module and a long hydrophobic groove; it tethers ER and phagophore and transfers glycerophospholipids for phagophore expansion, coupled to ATG9A scramblase activity, and is recruited via WIPI4/WDR45 PI3P binding [doi:10.1177/25152564231183898 Duarte & Reggiori 2023 review "bulk lipid transfer + scramblase coupling"; consistent with PMID:30952800, PMID:31271352, PMID:39174844 already in review]. ATG2A/ATG2B share ~44.5% identity (paralog redundancy already noted).

• NEW partner/donor route — ANKFY1 (endosome-to-phagophore lipid donation): Wei et al. 2024 identify ANKFY1 as a new ATG2A-binding, endosome-localized FYVE-domain protein that binds PI3P and enhances ATG2A-mediated lipid transfer between PI3P-containing liposomes; depletion phenocopies ATG2A/B loss (impaired autophagosome growth, reduced flux). Proposes endosomes (not only ER) as a lipid source for phagophore expansion via ATG2A [PMID:38622126 "ANKFY1 recruits ATG2A to PI3P-enriched endosomes and promotes ATG2A-mediated lipid transfer from endosomes to phagophores"; cached]. Strong, peer-reviewed (Cell Discovery) — added to references. Does not alter existing localization annotations (still ER/phagophore contact-site core) but broadens donor membrane context.

• NEW context — ER-side scramblases VMP1 and TMEM41B interact with the ATG2A N-terminus, forming a proposed core lipid-transfer/scrambling complex at ER–phagophore contacts; local ER lipid synthesis proposed to provide directionality [Chiduza et al. 2024 PMID:37938170; Falcon also cites a thesis nguyen2023 which I do NOT use]. PROVISIONAL as ATG2A-specific binding-site detail; consistent with existing tether/adaptor annotations, no annotation change.

• NEW context — ATG9B paralog: Chiduza et al. 2024 show ATG9B is a tissue-specific homotrimeric lipid scramblase that can compensate for ATG9A and can form a heteromeric complex with ATG2A PMID:37938170. Relevant to ATG2A–ATG9 coupling generally; ATG2A-shared, not ATG2A-unique. Added to references.

• NEW context — non-canonical autophagy / lysosome damage (CASM): Cross et al. 2023 report lysosome damage triggers CASM and promotes ATG2/ATG8 engagement; the ATG2–LC3 interaction required the ATG2 LIR rather than the WIPI4-binding region. Note the robustly demonstrated interaction was GFP-LC3A–ATG2B (paralog), with ATG2A inferred — so this is an ATG2-family/ATG2B-leaning finding, not firmly ATG2A-specific PMID:37796195. Suggests a possible LIR/ATG8-dependent recruitment mode and lysosome-repair role; PROVISIONAL for ATG2A. Captured as a suggested question rather than an annotation change.

• PROVISIONAL (preprint, do NOT use for annotations) — lipid droplets: Korfhage et al. 2023 (bioRxiv, NOT peer-reviewed) report ATG2A bridge-like phospholipid transport from LD monolayers regulates LD accumulation, with a transport-dead ATG2A mutant failing to rescue LD accumulation; LD-like monolayer donors accelerated lipid mixing ≥4-fold vs ~100 nm liposomes [doi:10.1101/2023.08.14.553257 preprint]. Supports the existing KEEP_AS_NON_CORE lipid-droplet annotations; preprint status means it is NOT used to change annotations or as supporting_text.

• van Vliet et al. 2024 (ATG9A interactome, PMID:38294121, DOI:10.1242/jcs.261081, PubMed-verified title "Exploring the ATG9A interactome uncovers interaction with VPS13A.") reaffirms ATG9A–ATG2A complex (scramblase function "thought to require" ATG2A interaction) and reports a distinct ATG9A–VPS13A complex; ATG2A role here is confirmatory of existing ATG9A coupling. Added to references (statement-only, full_text_unavailable) on 2026-06-07.

• PMID verification (2026-06-07, PubMed ID-converter): DOI 10.1038/s41421-024-00659-y -> PMID:38622126 (ANKFY1/Wei 2024); DOI 10.1080/15548627.2023.2275905 -> PMID:37938170 (ATG9B/Chiduza 2024); DOI 10.1083/jcb.202303078 -> PMID:37796195 (CASM/Cross 2023); DOI 10.1242/jcs.261081 -> PMID:38294121 (van Vliet 2024). All four already-or-now present in references are confirmed correct; no guessed PMIDs.

• Disease/physiology: no ATG2A-specific Mendelian disease established in the retrieved evidence; relevance is mechanistic (autophagy/lipid homeostasis, neurodegeneration context) [doi:10.1038/s41576-022-00562-w Yamamoto et al. 2023 Nat Rev Genet review, general]. No new disease annotation warranted.

Full-gene IBA re-review, 2026-09-20

All 31 annotation rows were screened. The selective macroautophagy IBAs for mitophagy, pexophagy, glycophagy and reticulophagy are retained as non-core contexts: ATG2A performs membrane-supply work and need not recognize each cargo itself. Cached PTHR13190 places these assertions at PTN000324023. Nuclear microautophagy is separately UNDECIDED because conservation of that particular route remains unresolved; a focused OpenScientist assessment is queued.

The PI3P-binding source was recovered directly: QuickGO resolves yeast ATG2/P53855 IDA to PMID:27498190, which assays a phosphoinositide-binding APT1-containing region. WIPI-enhanced recruitment of human ATG2A to PI3P membranes (PMID:31271352) is not proof that ATG2 lacks direct lipid binding, so the IBA is retained. The established core lipid-transfer role now includes ANKFY1-dependent endosomal donor contacts, directly supported by PMID:38622126, as well as ER contacts. The existing NEW ATG2-ATG18-family complex claim is grounded in human ATG2A-WIPI4 assemblies, including PMID:39174844.

Live QuickGO checked 2026-09-20 marks GO:0034045 obsolete and provides GO:7770114 phagophore membrane. Both original source rows are preserved with MODIFY replacements. The earlier 'no destination exists' rationale and redundant proposal for a new phagophore membrane term are superseded. The rim/contact-site proposals remain separate. Generic interaction rows are removed for informativeness, without rejecting their assays.

Recovery PR evidence follow-up (2026-09-22)

Incorporate the existing OpenScientist report and its route-specific divergence argument, recording the remaining lysosome/junction equivalence question for a second opinion.