Pseudomonas putida KT2440

Module-first GO review of a whole bacterial proteome

AI Gene Review · projects/P_PUTIDA · 2026

Bottom line

  • KT2440 is a versatile soil bacterium whose GO annotation is almost entirely automated.
  • Since July 2026 we review it pathway by pathway: start from a curated module, ask which steps KT2440 can satisfy, review only the genes each step needs.
  • September 2026 snapshot: 921 PSEPK reviews (5,324 rows) and 138 batch pages, still growing; unresolved steps are kept as explicit holes.

Why a module-first pass

  • 5,527 proteins; only ~50 GO annotations cite a paper.
  • Gene-by-gene review does not scale to a genome, and treats every gene as equally uncertain.
  • A module says what the organism must do; review effort goes to steps that are missing, ambiguous or over-propagated.
  • Early phase: 18 selected genes, then batches of 50 and 16 (TCA, stress, DNA repair, aromatic amino acids).

The workflow

A hole kept as a hole

What a batch page records

projects/P_PUTIDA/batches/ppu00470_d_amino_acid_cell_wall_precursor_supply.md: selected genes and their module assessment.

Results so far

857 rows marked over-annotated, 243 removed, 85 left undecided.

Status and next steps

  • ✅ Whole-proteome metadata, gene list, 161-bucket partition, pathway worklist.
  • ✅ 138 pathway batches, from the ppu00400 tryptophan pilot (PR #1874) to the most recent, purine-base oxidation (PR #2643).
  • ⬜ The status columns in data/psepk_pathway_worklist.tsv and the "Completed Reviews" tables predate most of this work.
  • ⬜ 777 of 921 reviews are still status: DRAFT. The 1,149 orphan and 825 unknown-function genes have no pathway module to seed them.

Read more: projects/P_PUTIDA.md · projects/P_PUTIDA/P_PUTIDA_MODULE_PLAN.md · projects/P_PUTIDA/batches/