Abi2 (mouse, P62484) review notes

Sources consulted

PMID:15572692 (Grove et al. 2004), the Abi2 knockout paper; the cache holds the abstract and discussion. Source of all mouse IDA/IMP rows (lens, neuronal migration, spines, memory, adherens junctions). Key quotes:
"In the absence of Abi2, secondary lens fiber orientation and migration were defective in the eye, without detectable defects in proliferation, differentiation, or apoptosis."
"Loss of Abi2 also resulted in cell migration defects in the neocortex and hippocampus, abnormal dendritic spine morphology and density, and severe deficits in short- and long-term memory."
"Abi2 was expressed in both axons and dendritic spines."
PMID:29911975 (Fan et al. 2018): "overexpression of either WAVE2 (Wasf2) or Abi2 in vivo rescues the cortical neuron migration defect of αKD neurons". This is a rescue experiment, not an Abi2 loss-of-function test.
PMID:17101133 (abstract only): "Abi-2, like Abi-1, promoted the c-Abl-mediated phosphorylation of Mena and WAVE2".
PMID:21107423 (WRC crystal structure, solved with Abi2): WAVE1:Abi2:HSPC300 trimer / four-helix bundle packing on Sra1.
PMID:11516653: EYFP-Abi-2b at the tips of lamellipodia and filopodia.
PMID:38081847 (human cells): GRAF1 phosphorylation recruits ABI2 and the WAVE2 complex to damaged mitochondria; branched actin is needed for mitochondrial clearance.
PMID:18632609: TRIM32 binds and ubiquitinates Abi2, driving its degradation.
Human ABI2 review (genes/human/ABI2) was the source of the ISO rows.

Core: signaling adaptor activity (the best available MF for a non-catalytic WRC subunit), SCAR complex, lamellipodium/filopodium tip, positive regulation of Arp2/3 nucleation and of lamellipodium assembly, Rac signalling, cell migration.
Neuronal (spine, synapse, SynGO-style rows), lens, AJ, memory, and mitophagy rows are kept as non-core: each is a tissue- or context-specific deployment of the WRC function.
dendrite development (IMP): marked over-annotated, because the measured phenotype concerns spines, not dendrite development generally; GO:0061001 captures it.
identical protein binding (ISO): marked over-annotated; the human source is high-throughput only, and ABI is a single subunit in the WRC.
small GTPase binding is kept with contributes_to, which is correct because Rac1 binds CYFIP, not ABI2.
No NEW terms were proposed. Kinase binding (Abl SH3 interaction) is by similarity only for mouse.

Abi2-deep-research-falcon.md agrees with the review: non-enzymatic WRC adaptor; Abl SH3 interaction; localization at lamellipodia, forming adherens junctions and dendritic spines. It also notes PIM1 phosphorylation of Ser183, which stabilizes ABI2 (Jensen et al. 2023, human). No change to any annotation decision.

GOA refreshed: no change; all GOA rows covered (59 rows, 51 entries, duplicates differing only in WITH/FROM).
Compared with genes/human/ABI2 and genes/human/ABI3 (no ABI1/Abi1 review exists; genes/ARATH/ABI1 is an unrelated ABA-insensitive phosphatase). All three share PANTHER PTHR10460 and treat signaling adaptor activity + SCAR complex as the core function; family IBAs (lamellipodium, SCAR complex, signaling adaptor activity, actin-based cell projection, neuron migration non-core) are handled identically.
Mitophagy (GO:0000423): mouse ISO kept as non-core, but the human source IDA (PMID:38081847) had been marked over-annotated. The full text shows ABI2 is the WRC subunit bound directly by the phosphorylated GRAF1 SH3 domain PMID:38081847, though ABI2 itself was not depleted. Harmonized the human review to KEEP_AS_NON_CORE, deferring to the experimental curator.
Small GTPase binding (GO:0031267, contributes_to): human IDA was MODIFY to signaling adaptor activity, ignoring the contributes_to qualifier; mouse ISO kept as non-core. Harmonized human to KEEP_AS_NON_CORE (complex-level activity via SRA1/CYFIP1).
Minor residual differences left as reviewer judgement: nucleus/nucleoplasm and identical protein binding are MARK_AS_OVER_ANNOTATED in mouse but KEEP_AS_NON_CORE in human; ubiquitin protein ligase binding is non-core in mouse, ACCEPT in human.