polb review notes

• Curated in DANRE batch 04. Polb is a base-excision repair polymerase. I narrowed broad DNA metabolism, DNA biosynthesis, and generic polymerase terms toward DNA-directed DNA polymerase/5-dRP lyase activity in BER, while retaining DNA binding and metal binding as non-core biochemical properties.
• Follow-up: changed broad DNA BP annotations from cross-aspect MODIFY to REMOVE; BER remains the accepted specific BP.

Re-review 2026-09-29

Repaired a malformed row and rewrote all 26 rows. Row 16 of the yaml had a truncated cytoplasm annotation (label "cy", no evidence_type) that failed schema validation and did not match GOA; it is restored as the second GO:0005737 ISS row from UniProtKB:P06746, which is what GOA carries.

• Resolved the ten PENDING rows: GO:0003887 IDA and ISS -> ACCEPT; GO:0005737 ISS -> KEEP_AS_NON_CORE; GO:0006284 ISS -> ACCEPT; GO:0006974 ISS -> KEEP_AS_NON_CORE; GO:0016829 ISS -> MODIFY to the two specific lyase terms; GO:0034061 ISS -> MODIFY to GO:0003887; GO:0051575 ISS -> ACCEPT; GO:0140078 ISS -> KEEP_AS_NON_CORE; GO:0006303 IBA -> REMOVE (below).
• The GO:0003887 IDA is now grounded in the actual zebrafish experiment [PMID:22018137 "We have succeeded in overproducing, purifying, and characterizing Danio pol β."; "We investigated the kinetic constants Km and kcat, and the catalytic efficiency kcat/Km for Danio pol β in the presence of either Mg2+ or Mn2+."], with the Mn2+/Mg2+ result PMID:22018137.
• Re-adjudicated the three REMOVE actions. GO:0006281 DNA repair (IEA) and GO:0006259 DNA metabolic process (IEA) changed REMOVE -> MODIFY to GO:0006284 base-excision repair: these are over-general electronic assignments, not false ones, and the specific pathway is known. GO:0071897 DNA biosynthetic process changed REMOVE -> KEEP_AS_NON_CORE: pol beta genuinely synthesises DNA [file:DANRE/polb/polb-uniprot.txt "through its DNA polymerase activity, it adds one nucleotide to the 3' end of the arising single-nucleotide gap"], so removal would deny a true statement; it is simply not the informative framing.
• New REMOVE, with propagation_review: GO:0006303 double-strand break repair via nonhomologous end joining (IBA from PTN000138687). The X-family node spreads an activity of the pol lambda/pol mu/Pol4 branch across the whole family. QuickGO check: human POLB (P06746) holds GO:0006303 only by the same IBA, while POLL has IDA (PMID:30250067) and IMP (PMID:19806195) and POLM has TAS - so there is no experimental descendant on the pol beta branch. Structurally, the BRCT module that recruits pol lambda/mu to the Ku-XRCC4-ligase IV synapse is absent from pol beta, which consists of the 8-kDa lyase domain and the polymerase domain [file:DANRE/polb/polb-uniprot.txt "Belongs to the DNA polymerase type-X family."; PMID:22018137 "This substitution lies in a hairpin sequence within an 8-kDa domain, likely to be important in DNA binding."]. Recorded as PROPAGATION_BAD with FUNCTIONAL_DIVERGENCE and WRONG_ORTHOLOG_OR_PARALOG, with POLL and yeast POL4 marked SUPPORTS_SOURCE_BUT_NOT_TARGET.
• GO:0140078 class I AP endonuclease activity kept non-core as the AP-lyase side reaction [file:DANRE/polb/polb-uniprot.txt "It is also able to cleave sugar-phosphate bonds 3' to an intact AP site, acting as an AP lyase."], while GO:0051575 5'-dRP lyase is accepted as the second core activity.
• core_functions were absent; two entries added (gap-filling polymerase; 5'-dRP lyase). Description rewritten as standalone biology, suggested_questions and suggested_experiments added, PMID:22018137 given a reference_review noting it assayed polymerase activity only.