Merge pull request #196 from phac-nml/dev/columns

Adding Additional Columns to Pointfinder Output

CHANGELOG.md

@@ -2,7 +2,7 @@
 
 * Fixed an issue where the string "None" in the drug table would be parsed differently by different versions of pandas (#175).
 * Upgraded to pandas version 2.
-* Added CGE-predicted phenotypes to Pointfinder output.
+* Added the CGE-predicted phenotypes, PMID, Mechanism, and the Required Mutation columns to the Pointfinder output.
 * The resfinder.tsv and pointfinder.tsv outputs now contain a Notes column.
 * Updated the help description of the --mlst-scheme parameter to include a more useful link for available schemas.
 * Switched to only officially supporting Python 3.7+ due to recent incompatibilities with Python 3.6 and some Python packages (numpy, biopython, and others).

staramr/blast/pointfinder/PointfinderBlastDatabase.py

@@ -94,27 +94,62 @@ def get_organism(self):
     def get_name(self):
         return 'pointfinder'
 
-    def get_notes(self, gene, mutation):
+    def get_cge_notes(self, gene, mutation):
         """
         Gets the note associated with a particular mutation from the Pointfinder Database table.
 
         :param gene: The gene.
         :param mutation: The mutation.
-        :return: A string containtain the Note, if it exists.
+        :return: A string containing the note, if it exists.
         """
 
-        return self._pointfinder_info.get_notes(gene, mutation)
+        return self._pointfinder_info.get_value(gene, mutation, "Notes")
 
     def get_cge_phenotype(self, gene, mutation):
         """
         Gets the phenotype associated with a particular mutation from the Pointfinder Database table.
 
         :param gene: The gene.
         :param mutation: The mutation.
-        :return: A string containtain the phenotype, if it exists.
+        :return: A string containing the phenotype, if it exists.
         """
 
-        return self._pointfinder_info.get_phenotype(gene, mutation)
+        return self._pointfinder_info.get_value(gene, mutation, "Resistance")
+
+    def get_cge_pmid(self, gene, mutation):
+        """
+        Gets the PMID associated with a particular mutation from the Pointfinder Database table.
+
+        :param gene: The gene.
+        :param mutation: The mutation.
+        :return: A string (not integer) containing the PMID, if it exists.
+        """
+
+        return self._pointfinder_info.get_value(gene, mutation, "PMID")
+
+    def get_cge_required_mutation(self, gene, mutation):
+        """
+        Gets the required mutation associated with a particular mutation from
+        the Pointfinder Database table.
+
+        :param gene: The gene.
+        :param mutation: The mutation.
+        :return: A string containing the required mutation, if it exists.
+        """
+
+        return self._pointfinder_info.get_value(gene, mutation, "Required_mut")
+
+    def get_cge_mechanism(self, gene, mutation):
+        """
+        Gets the mechanism associated with a particular mutation from the
+        Pointfinder Database table.
+
+        :param gene: The gene.
+        :param mutation: The mutation.
+        :return: A string containing the mechanism, if it exists.
+        """
+
+        return self._pointfinder_info.get_value(gene, mutation, "Mechanism")
 
     @classmethod
     def get_available_organisms(cls):

staramr/blast/pointfinder/PointfinderDatabaseInfo.py

@@ -187,38 +187,22 @@ def get_resistance_nucleotides(self, gene, nucleotide_mutations):
 
         return resistance_mutations
 
-    def get_notes(self, gene, mutation):
-        """
-        Gets the note associated with a particular mutation from the Pointfinder Database table.
-
-        :param gene: The gene.
-        :param mutation: The mutation.
-        :return: A string containtain the Notes, if they exist, or the empty string ("") if
-                 there are no Notes.
-        """
-
-        matches = self._get_resistance_codon_match(gene, mutation)
-        matches = matches.fillna("")
-
-        # There's a chance of having multiple matches:
-        notes = ';'.join(matches["Notes"])
-
-        return notes
-
-    def get_phenotype(self, gene, mutation):
+    def get_value(self, gene, mutation, attribute):
         """
         Gets the phenotype associated with a particular mutation from the Pointfinder Database table.
 
         :param gene: The gene.
         :param mutation: The mutation.
-        :return: A string containing the phenotype, if it exists, or the empty string ("") if
-                 there is no phenotype.
+        :param attribute: The attribute to get (ex: "Resistance", "PMID", etc.)
+        :return: A string containing the attribute for the passed gene and
+                 mutation, if it exists, or the empty string ("") if the
+                 attribute is missing.
         """
 
         matches = self._get_resistance_codon_match(gene, mutation)
         matches = matches.fillna("")
 
         # There's a chance of having multiple matches:
-        resistances = ';'.join(matches["Resistance"])
+        results = ';'.join(matches[attribute])
 
-        return resistances
+        return results

staramr/blast/results/pointfinder/BlastResultsParserPointfinder.py

@@ -68,7 +68,7 @@ def _get_result(self, hit, db_mutation):
                 hit.get_genome_contig_start(),
                 hit.get_genome_contig_end(),
                 db_mutation.get_pointfinder_mutation_string(),
-                self._blast_database.get_notes(hit.get_amr_gene_id(), db_mutation)
+                self._blast_database.get_cge_notes(hit.get_amr_gene_id(), db_mutation)
                 ]
 
         return result

staramr/blast/results/pointfinder/BlastResultsParserPointfinderResistance.py

@@ -28,6 +28,9 @@ class BlastResultsParserPointfinderResistance(BlastResultsParserPointfinder):
     End
     Pointfinder Position
     CGE Notes
+    CGE Required Mutation
+    CGE Mechanism
+    CGE PMID
     '''.strip().split('\n')]
 
     def __init__(self, file_blast_map, arg_drug_table, blast_database, pid_threshold, plength_threshold,
@@ -84,7 +87,10 @@ def _get_result(self, hit, db_mutation):
                 hit.get_genome_contig_start(),
                 hit.get_genome_contig_end(),
                 db_mutation.get_pointfinder_mutation_string(),
-                self._blast_database.get_notes(hit.get_amr_gene_id(), db_mutation)
+                self._blast_database.get_cge_notes(hit.get_amr_gene_id(), db_mutation),
+                self._blast_database.get_cge_required_mutation(hit.get_amr_gene_id(), db_mutation),
+                self._blast_database.get_cge_mechanism(hit.get_amr_gene_id(), db_mutation),
+                self._blast_database.get_cge_pmid(hit.get_amr_gene_id(), db_mutation),
                 ]
 
         return result

staramr/blast/results/resfinder/BlastResultsParserResfinderResistance.py

@@ -66,6 +66,6 @@ def _get_result_rows(self, hit, database_name):
                  hit.get_genome_contig_end(),
                  hit.get_amr_gene_accession(),
                  hit.get_genome_contig_hsp_seq(),
-                 self._cge_drug_table.get_notes(hit.get_amr_gene_name_with_variant(),
+                 self._cge_drug_table.get_cge_notes(hit.get_amr_gene_name_with_variant(),
                                                 hit.get_amr_gene_accession())
                  ]]

staramr/databases/resistance/cge/CGEDrugTableResfinder.py

@@ -49,7 +49,7 @@ def get_drug(self, drug_class, gene_plus_variant, accession):
         else:
             return drug.iloc[0]
 
-    def get_notes(self, gene_plus_variant, accession):
+    def get_cge_notes(self, gene_plus_variant, accession):
         """
         Gets the notes for a gene (plus variant of gene encoded in ResFinder database) and accession.
         :param gene_plus_variant: The gene plus variant (e.g., {gene}_{variant} = {blaIMP-58}_{1}).

staramr/databases/resistance/pointfinder/complex/ComplexMutations.py

@@ -54,7 +54,7 @@ def get_matches(self, results_table, hit):
                 result = [hit.get_genome_id(),
                         ", ".join(intersection), # Technically these are also Pointfinder co-ords
                         row.phenotype,
-                        None,  # CGE-predicted phenotype
+                        pd.NA,  # CGE-predicted phenotype
                         "complex",  # Type
                         ", ".join(mutation_positions),
                         "complex", # Creating a mutation string would be confusing for this.
@@ -65,7 +65,10 @@ def get_matches(self, results_table, hit):
                         hit.get_genome_contig_start(),
                         hit.get_genome_contig_end(),
                         ", ".join(intersection),
-                        "This mutation represents a combination of multiple individual mutations."]  # The notes.
+                        "This mutation represents a combination of multiple individual mutations.", # The notes.
+                        pd.NA,
+                        pd.NA,
+                        pd.NA]
 
                 matches.append(result)