gempro.py 文件源码

def set_representative_sequence(self, force_rerun=False):
        """Automatically consolidate loaded sequences (manual, UniProt, or KEGG) and set a single representative sequence.

        Manually set representative sequences override all existing mappings. UniProt mappings override KEGG mappings
        except when KEGG mappings have PDBs associated with them and UniProt doesn't.

        Args:
            force_rerun (bool): Set to True to recheck stored sequences

        """

        # TODO: rethink use of multiple database sources - may lead to inconsistency with genome sources

        sequence_missing = []
        successfully_mapped_counter = 0
        for g in tqdm(self.genes):
            repseq = g.protein.set_representative_sequence(force_rerun=force_rerun)

            if not repseq:
                sequence_missing.append(g.id)
            elif not repseq.sequence_file:
                sequence_missing.append(g.id)
            else:
                successfully_mapped_counter += 1

        log.info('{}/{}: number of genes with a representative sequence'.format(len(self.genes_with_a_representative_sequence),
                                                                                len(self.genes)))
        log.info('See the "df_representative_sequences" attribute for a summary dataframe.')