- Export of cluster sequences with genes and domain annotations in GenBank format.
- Proper data model to allow retaining more information from input.
- Dedicated functions to load the KNN training matrix and the CRF model.
- Sorting issue causing
ClusterCRFto be passed a feature table with potentially scrambled proteins.
- All internal classes but
ClusterCRFexpect objects from
KNNtraining matrix is now stored as a compressed sparse matrix.
- Internal data is stored in the relevant module next to the code needing it instead of in a dedicated
gecco.orf.ORFFinderworks at the single record level.
- Reduce code complexity in
- Export of individual cluster proteins in FASTA files.
PyrodigalFinderis used instead).