EIF2AK2 (PKR) review notes
Literature search (no deep-research provider available)
Sources checked: the UniProt entry P19525 (curated FUNCTION, SUBUNIT, SUBCELLULAR LOCATION
and ACTIVITY REGULATION text); every GOA-cited publication cached by
just fetch-gene-pmids human EIF2AK2 (49 papers; most are abstract-only); and the cached
Reactome reactions for the PKR pathway (R-HSA-98336xx/9835xxx/9839xxx). Context only: the integrated stress response module (module proposed in ai4curation/ai-gene-review#5904),
which models PKR as one realizer of the eIF2alpha kinase step, and GO-CAM 66b5638000001589, which is about PERK and contains no PKR activity.
Key findings
Curation decisions
- The 58 bare protein-binding IPI rows are REMOVEd under the protein-binding policy. Identical protein binding is MODIFYed to protein homodimerization activity.
- The general protein kinase activity and kinase activity rows are MODIFYed to protein serine/threonine kinase activity, or to eIF2alpha kinase activity where eIF2alpha was the substrate assayed.
- Cytoplasm rows are MODIFYed to cytosol. Nucleus and nucleoplasm rows are kept as non-core.
- Protein phosphatase regulator activity is REMOVEd, because PKR regulates PP2A by phosphorylating B56alpha, not as a regulatory subunit.
- Cellular response to amino acid starvation: the ARBA IEA row is marked as over-annotated, because amino-acid sensing is the GCN2 branch. The IMP row from PMID:25329545 is UNDECIDED, because the abstract does not describe a PKR experiment.
- NEW: integrated stress response signaling (GO:0140467). Comparator: EIF2AK1 carries it by IDA.