Two independent routes build a draft model for an organism that has no curated GEM yet.
Transfer reactions from a curated template model to the target organism via an ortholog search.
- Run the search (or bring your own table):
{func}
raven_toolbox.reconstruction.homology.run_blast/ {func}raven_toolbox.reconstruction.homology.run_diamond, then {func}raven_toolbox.reconstruction.homology.make_ortholog_hitsto get the canonicalgene × genehits DataFrame (bidirectional / best-hits-only policies supported). - Draft the model:
{func}
raven_toolbox.reconstruction.homology.get_model_from_homology— AST-based GPR rewrite, configurable complex policy, returns a {class}cobra.Model(plus aHomologyResult).
The external aligners are resolved from the pinned binary registry
({mod}raven_toolbox.binaries).
Draft directly from KEGG orthology, either for a KEGG-listed species or from your own protein FASTA via HMM search.
- KEGG species (no FASTA):
{func}
raven_toolbox.reconstruction.kegg.get_kegg_model_for_organism. - Your sequences (HMM search): {func}
raven_toolbox.reconstruction.kegg.assign_kos→ {func}raven_toolbox.reconstruction.kegg.get_kegg_model_from_sequences. The HMM cut-off defaults are calibrated in the KEGG HMM cut-off study.
The KEGG artefact bundle (KO tables, reference model, HMM libraries) is fetched by
{mod}raven_toolbox.data; building and publishing it is a maintainer task — see
Maintaining KEGG data and the
KEGG data format.
After drafting, fill connectivity gaps with the gap-filling guide.