fix user provided pseudo genes in GenBank files #94

oschwengers · Jan 31, 2022 · 0f0165d · 0f0165d
1 parent 69b3103
commit 0f0165d
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 1 deletion.
diff --git a/bakta/expert/protein_sequences.py b/bakta/expert/protein_sequences.py
@@ -104,7 +104,7 @@ def write_user_protein_sequences(aa_fasta_path):
         with xopen(str(cfg.user_proteins), threads=0) as fh_in:
             for record in SeqIO.parse(fh_in, 'genbank'):
                 for feature in record.features:
-                    if(feature.type.lower() == 'cds'):
+                    if(feature.type.lower() == 'cds'  and  'pseudo' not in feature.qualifiers):
                         user_proteins.append(parse_user_protein_sequences_genbank(feature))
     except Exception as e:
         log.error('provided user proteins file GenBank format not valid!', exc_info=True)

diff --git a/test/data/user-proteins.gbff b/test/data/user-proteins.gbff
@@ -54,6 +54,21 @@ FEATURES             Location/Qualifiers
                      /transl_table=11
                      /protein_id="gnl|Bakta|hypo-mock-test"
                      /inference="ab initio prediction:Prodigal:2.6"
+     gene            1001..1038
+                     /locus_tag="pseudo-mock-test"
+                     /gene="mock2"
+                     /pseudo
+     CDS             1001..1038
+                     /locus_tag="pseudo-mock-test"
+                     /gene="mock2"
+                     /db_xref="USERDB:MOCK2"
+                     /EC_number="0.0.0.0"
+                     /product="mock pseudo user protein 2"
+                     /pseudo
+                     /codon_start=1
+                     /transl_table=11
+                     /protein_id="gnl|Bakta|hypo-mock-test"
+                     /inference="ab initio prediction:Prodigal:2.6"
 ORIGIN
         1 ttcttctgcg agttcgtgca gcttctcaca catggtggcc tgctcgtcag catcgagtgc
        61 gtccagtttt tcgagcagcg tcaggctctg gctttttatg aatcccgcca tgttgagtgc