Merge pull request #222 from Knowledge-Graph-Hub/fix_sars_cov_2_taxon…

…_ids Prettify taxon id in sars-cov-2 gene annotations
Knowledge-Graph-Hub · Jun 12, 2020 · 9cab8b4 · 9cab8b4
2 parents c983cfd + 0c91a69
commit 9cab8b4
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 4 deletions.
diff --git a/kg_covid_19/transform_utils/sars_cov_2_gene_annot/sars_cov_2_gene_annot.py b/kg_covid_19/transform_utils/sars_cov_2_gene_annot/sars_cov_2_gene_annot.py
@@ -21,15 +21,15 @@ def __init__(self, input_dir: Optional[str] = None, output_dir: str = None):
         source_name = "sars_cov_2_gene_annot"
         super().__init__(source_name, input_dir, output_dir)
 
-        self.node_header = ['id', 'name', 'category', 'synonym', 'taxon', 'provided_by']
+        self.node_header = ['id', 'name', 'category', 'synonym', 'in_taxon', 'provided_by']
         self.edge_header = ['subject', 'edge_label', 'object', 'relation',
                             'provided_by', 'DB_References', 'ECO_code', 'With',
                             'Interacting_taxon_ID',
                             'Date', 'Assigned_by', 'Annotation_Extension',
                             'Annotation_Properties']
 
         self.protein_node_type = "biolink:Protein"
-        self.ncbi_taxon_prefix = "NCBITaxon:"
+        self.ncbi_taxon_prefix = "NCBITaxon"
 
         # translate edge labels to RO term, for the 'relation' column in edge
         self.edge_label_prefix = "biolink:"  # prepend to edge label
@@ -88,6 +88,8 @@ def gpa_to_edge_data(self, rec: dict) -> list:
                 item = get_item_by_priority(rec, [key])
                 if type(item) is list:
                     item = item[0]
+                if key == 'Interacting_taxon_ID':
+                    item = ":".join([self.ncbi_taxon_prefix, item])
             except (ItemInDictNotFound, IndexError):
                 item = ''
             edge_data.append(item)
@@ -109,7 +111,7 @@ def gpi_to_gene_node_data(self, rec: dict) -> list:
         :param rec: record from gpi iterator
         :return: list of node items, one for each thing in self.node_header
         """
-        # ['id', 'name', 'category', 'synonym', 'taxon']
+        # ['id', 'name', 'category', 'synonym', 'in_taxon']
         id: str = self._rec_to_id(rec)
 
         try:
@@ -127,6 +129,7 @@ def gpi_to_gene_node_data(self, rec: dict) -> list:
         except (IndexError, ItemInDictNotFound):
             synonym = ''
         taxon = get_item_by_priority(rec, ['Taxon'])
+        taxon = ":".join([self.ncbi_taxon_prefix, taxon.split(":")[1]])
         return [id, name, category, synonym, taxon, self.source_name]
 
 

diff --git a/tests/test_sars_cov_2_gene_annot.py b/tests/test_sars_cov_2_gene_annot.py
@@ -44,7 +44,7 @@ def test_gpi_to_gene_node(self):
         node = self.sc2ga.gpi_to_gene_node_data(item)
         self.assertEqual(len(self.sc2ga.node_header), len(node))
         self.assertEqual(node,
-                         ['UniProtKB:P0DTD2', 'Protein 9b', 'biolink:Protein', '', 'taxon:2697049', 'sars_cov_2_gene_annot'])
+                         ['UniProtKB:P0DTD2', 'Protein 9b', 'biolink:Protein', '', 'NCBITaxon:2697049', 'sars_cov_2_gene_annot'])
 
     def test_gpa_to_edge_data(self):
         gpa_iter = _gpa11iterator(self.gpa_fh)