# Hand-adjudicated failure cases, P. putida KT2440, NCBI RefSeq vs GOA. # Generated tables (ncbi_review_cases.tsv, goa_errors_vs_ncbi.tsv) give every # row; this file records the cases that were checked by hand, 2026-10-18. # # ncbi_specific: NCBI terms absent from GOA, so never shown to reviewers # ("unassessed" in ncbi_review_cases.tsv). Each was checked against the gene's # review (description/core_functions) and, for HMM models, against the model's # own go_terms in hmm_PGAP.tsv. "model_go_confirmed" means the term is listed on # the model itself, so fixing the model fixes every genome. # goa_specific: GOA terms reviewers removed that NCBI did not assert # ("ncbi_avoided" in goa_errors_vs_ncbi.tsv). # shared: GOA terms reviewers removed that NCBI also asserts. # naming: protein-name disagreements found along the way. ncbi_specific: - gene: secG locus: PP_5706 model: TIGR00810 terms: [GO:0015136 sialic acid transmembrane transporter activity, GO:0015739 sialic acid transport] failure: wrong_function_on_model model_go_confirmed: true note: SecG is a SecYEG channel subunit (review core BP is protein transport by the Sec complex). - gene: queF locus: PP_2160 model: TIGR03138 terms: [GO:0043867 7-cyano-7-deazaguanine tRNA-ribosyltransferase activity] failure: wrong_function_on_model model_go_confirmed: true note: This is the activity of the next enzyme in the pathway (Tgt-type), not of the preQ1 synthase QueF. - gene: lgt locus: PP_5142 model: TIGR00544 terms: [GO:0009249 protein lipoylation] failure: wrong_function_on_model model_go_confirmed: true note: Lipoylation attaches lipoic acid; Lgt performs diacylglyceryl transfer in lipoprotein biosynthesis (GO:0042158). - gene: pqqA locus: PP_0380 model: TIGR02107 terms: [GO:0003824 catalytic activity] failure: substrate_given_activity model_go_confirmed: true note: PqqA is the precursor peptide that PqqBCDE modify; it catalyses nothing. - gene: aacs locus: PP_3071 model: TIGR01217 terms: [GO:0019287 isopentenyl diphosphate biosynthetic process, mevalonate pathway] failure: pathway_context_ignored model_go_confirmed: true note: KT2440 makes isoprenoids via the MEP pathway; the review places Aacs in 3-hydroxybutyrate utilisation. - gene: purE locus: PP_5336 model: TIGR01162 terms: [GO:0004638 phosphoribosylaminoimidazole carboxylase activity] failure: name_and_go_disagree model_go_confirmed: true note: The model is named N5-CAIR mutase, but its GO is the class II carboxylase activity; correct term is GO:0034023. - gene: yajC locus: PP_0834 model: TIGR00739 terms: [GO:0015450 protein-transporting ATPase activity] failure: complex_activity_transfer model_go_confirmed: true note: YajC is a SecDF-associated accessory subunit with no ATPase. - gene: imuB locus: PP_3118 model: ArchId:10131054 terms: [GO:0003887 DNA-directed DNA polymerase activity] failure: pseudoenzyme_overannotation note: Review describes ImuB as a catalytically inactive polymerase-like adaptor. - gene: PP_3384 locus: PP_3384 model: ArchId:10613030 terms: [GO:0008873 gluconate 2-dehydrogenase (NADP+) activity] failure: complex_activity_transfer note: Gamma subunit of the periplasmic FAD-dependent gluconate 2-dehydrogenase; no independent activity, and the term names the unrelated cytosolic NADP+ enzyme. - gene: glgE locus: PP_4060 model: ArchId:10571012 terms: [GO:0009313 oligosaccharide catabolic process] failure: wrong_direction note: GlgE elongates alpha-glucan (review core BP alpha-glucan biosynthetic process). - gene: PP_2928 locus: PP_2928 model: ArchId (saccharopine dehydrogenase family) terms: [GO:0009085 L-lysine biosynthetic process] failure: family_level_substrate_transfer note: Review identifies a carboxyspermidine dehydrogenase (spermidine biosynthesis). - gene: PP_1969 and PP_2482 locus: PP_1969, PP_2482 model: ArchId (MoaA) terms: [GO:1990140 molybdopterin synthase complex] failure: wrong_complex note: MoaA is not a subunit of the MoaD/MoaE molybdopterin synthase complex. - gene: oprE, nicP, galP model: ArchId (OprD family porin) terms: [GO:0015802 basic amino acid transport] failure: family_level_substrate_transfer note: Correct for OprD itself; copied to every OprD/Occ-family porin, whose substrates differ. - gene: sdhC and sdhD model: TIGR02970, TIGR02968 terms: [GO:0000104 succinate dehydrogenase activity] failure: complex_activity_transfer model_go_confirmed: true note: Membrane anchor subunits; review core MF for SdhC is electron transfer activity. - gene: tatC-I and tatC-II model: TIGR00945 terms: [GO:0015628 protein secretion by the type II secretion system] failure: wrong_function_on_model model_go_confirmed: true note: Tat translocation is a separate system (review core BP protein transport by the Tat complex). - gene: ttg2D model: ArchId:10006734 terms: [GO:0140359 ABC-type transporter activity] failure: wrong_function_on_model note: MlaC is a periplasmic phospholipid shuttle, not a transporter (review core BP intermembrane phospholipid transfer). # Checked and NOT counted as errors: # - wspR GO:0071111 c-di-GMP phosphodiesterase: the review notes a REC-PAS-GGDEF-EAL architecture, so PDE is not ruled out. # - oprD GO:0015802: correct for OprD itself. # - 64 of the 224 unassessed NCBI terms are obsolete GO terms (GAF built on GO 2025-03-16); stale, not wrong mappings. goa_specific: # Paralog/cross-family transfers that the specific NCBI equivalog avoided. - {gene: mqo1/mqo2/mqo3, goa_source: TreeGrafter, term: GO:0047545 (S)-2-hydroxyglutarate dehydrogenase activity} - {gene: dapE, goa_source: TreeGrafter, term: GO:0008777 acetylornithine deacetylase activity} - {gene: metXS, goa_source: TreeGrafter, term: GO:0004414 homoserine O-acetyltransferase activity} - {gene: kdsC, goa_source: TreeGrafter, term: GO:0008781 N-acylneuraminate cytidylyltransferase activity} - {gene: phnX, goa_source: TreeGrafter, term: GO:0008967 phosphoglycolate phosphatase activity} - {gene: gcl, goa_source: TreeGrafter, term: GO:0009099 L-valine biosynthetic process} - {gene: glyS, goa_source: InterPro2GO, term: GO:0004814 arginine-tRNA ligase activity} - {gene: thiI, goa_source: InterPro2GO, term: GO:0004810 CCA tRNA nucleotidyltransferase activity} - {gene: purC, goa_source: InterPro2GO, term: GO:0009236 cobalamin biosynthetic process} - {gene: prpB, goa_source: Combined IEA, term: GO:0004451 isocitrate lyase activity} - {gene: putA, goa_source: Combined IEA, term: GO:0055129 L-proline biosynthetic process} - {gene: hisI, goa_source: ARBA, term: GO:0004636 phosphoribosyl-ATP diphosphatase activity} shared: # Reviewer-removed GOA terms that NCBI also asserts (66 terms). Recurrent patterns: complex_activity_transfer: [dnaN/dnaQ/dnaX GO:0003887, eutC GO:0008851, carA GO:0004088, nirD GO:0008942, soxG GO:0008115, ureA GO:0016151, acoA GO:0016624] wrong_cofactor: [icd GO:0051287 NAD binding (NADP enzyme), gapA GO:0050661 NADP binding, prpD and nirD GO:0051537 2Fe-2S, dnaJ and groES GO:0005524 ATP binding] paralog_substrate: [pssA GO:0008444 (PgsA reaction), nadE GO:0003952/GO:0004359 (NH3-dependent enzyme), murA GO:0019277 (UDP-GalNAc), purT GO:0004644, sbcB GO:0000175 (RNA exonuclease), cysH GO:0004604] wrong_direction: [gpsA GO:0046168 catabolic, dfp GO:0015941 pantothenate catabolic, secA GO:0017038 protein import] ncbi_name_wrong: [pycB named sodium-extruding oxaloacetate decarboxylase; review assigns pyruvate carboxylase subunit B, pyrC' named dihydroorotase but lacks catalytic histidines] naming: - gene: puuD (PP_3099) winner: NCBI note: UniProt keeps the 2002 submitter name (Name=puuD, uricase); NCBI names it the T6SS contractile sheath large subunit, which the review independently confirms.