Consistency: ROW-1 CONFLICT. The review and notes characterize CSNK2B exclusively as the CK2 regulatory beta subunit (kinase regulator / adaptor / holoenzyme), nuclear+cytosolic, with NO ribosome-biogenesis or SSU-processome role. CSNK2B-goa.tsv has no GO:0042254/GO:0032040. The "UTP-C complex" is a yeast SSU-processome subcomplex (yeast Utp-C contains CK2); human CSNK2B SSU-processome membership is not established — PubMed returns zero hits for CSNK2B+UTP-C+processome. So row-1 is a likely yeast-orthology over-projection. Row-2 (CK2 phosphorylates ATG16L1 to promote ATG5-ATG12 binding) is biologically real but the review only captures it indirectly (kinase-regulator pleiotropy); no autophagy annotation.
PN story / NEW pressure: Row-1 over-reaches: projecting GO:0042254 "ribosome biogenesis" + GO:0032040 "small-subunit processome" (both verified real) onto human CSNK2B is not supported by human evidence and conflicts with the review's holoenzyme-centric core. Row-2 ATG16L1 phosphorylation is a defensible kinase-substrate role but is correctly left no_mapping (a regulatory, member-specific role, not a shared GO assertion).
Mapping strategy: Row-1 mapping should be reconsidered — analogous to the rejected TOMM20/HSPA8/RAB7A "too broad / orthology-driven" cases, GO:0042254/GO:0032040 over-annotate human CSNK2B. Row-2 ALP handling (all no_mapping / context_only) is appropriately conservative.
Evidence alignment: Row-1 lists no PN references (no human paper supports it). Row-2 PN cites a CSNK2/ATG16L1 cardiomyocyte paper + an autophagy-enzyme review (titles only, not PMIDs) — neither is in the review. Review's substrate/interactome PMIDs do not overlap the PN rows.
Verdict: ROW-1 OVER-REACHES (yeast UTP-C orthology); ROW-2 acceptably no_mapping. Recommended edits: [MAP] downgrade row-1 SSU-processome/ribosome-biogenesis projection (GO:0032040/GO:0042254) to context_only/no_mapping for human CSNK2B pending human evidence; [WB] human CSNK2B SSU-processome membership unverified (PubMed: 0 hits).