Merge pull request #96 from Knowledge-Graph-Hub/fix_data_errors_drug_…

…central Fix data errors drug central
Knowledge-Graph-Hub · Apr 15, 2020 · 487136a · 487136a
2 parents 7175d04 + 62b1a46
commit 487136a
Show file tree

Hide file tree

Showing 3 changed files with 93 additions and 29 deletions.
diff --git a/kg_covid_19/transform_utils/drug_central/drug_central.py b/kg_covid_19/transform_utils/drug_central/drug_central.py
@@ -10,7 +10,7 @@
 
 from kg_covid_19.transform_utils.transform import Transform
 from kg_covid_19.utils.transform_utils import write_node_edge_item, \
-    get_item_by_priority, ItemInDictNotFound, parse_header
+    get_item_by_priority, ItemInDictNotFound, parse_header, data_to_dict
 
 """
 Ingest drug - drug target interactions from Drug Central
@@ -28,13 +28,17 @@ def __init__(self, input_dir: str = None, output_dir: str = None) -> None:
         source_name = "drug_central"
         super().__init__(source_name, input_dir, output_dir)  # set some variables
 
-    def run(self) -> None:
-        """Method is called and performs needed transformations to process the Drug Central data, additional information
-     on this data can be found in the comment at the top of this script"""
+    def run(self, species="Homo sapiens") -> None:
+        """Method is called and performs needed transformations to process the Drug
+        Central data, additional information
+        on this data can be found in the comment at the top of this script"""
 
-        interactions_file = os.path.join(self.input_base_dir, "drug.target.interaction.tsv.gz")
+        interactions_file = os.path.join(self.input_base_dir,
+                                         "drug.target.interaction.tsv.gz")
         os.makedirs(self.output_dir, exist_ok=True)
         drug_node_type = "biolink:Drug"
+        gene_curie_prefix = "UniProtKB:"
+        drug_curie_prefix = "DrugCentral:"
         gene_node_type = "biolink:Gene"
         drug_gene_edge_label = "biolink:interacts_with"
         drug_gene_edge_relation = "RO:0002436"  # molecularly interacts with
@@ -52,21 +56,21 @@ def run(self) -> None:
             for line in interactions:
                 items_dict = parse_drug_central_line(line, header_items)
 
+                if 'ORGANISM' not in items_dict or items_dict['ORGANISM'] != species:
+                    continue
+
                 # get gene ID
                 try:
-                    gene_id = get_item_by_priority(items_dict, ['ACCESSION'])
+                    gene_id_string = get_item_by_priority(items_dict, ['ACCESSION'])
+                    gene_ids = gene_id_string.split('|')
                 except ItemInDictNotFound:
                     # lines with no ACCESSION entry only contain drug info, no target
                     # info - not ingesting these
-                    logging.info(
-                        "No gene information for this line:\n{}\nskipping".format(line))
                     continue
 
                 # get drug ID
-                drug_id = get_item_by_priority(items_dict,
-                                               ['ACT_SOURCE_URL',
-                                                'MOA_SOURCE_URL',
-                                                'DRUG_NAME'])
+                drug_id = drug_curie_prefix + get_item_by_priority(items_dict,
+                                                                   ['DRUG_NAME'])
 
                 # WRITE NODES
                 # drug - ['id', 'name', 'category']
@@ -76,21 +80,23 @@ def run(self) -> None:
                                            items_dict['DRUG_NAME'],
                                            drug_node_type])
 
-                write_node_edge_item(fh=node,
-                                     header=self.node_header,
-                                     data=[gene_id,
-                                           items_dict['GENE'],
-                                           gene_node_type])
-
-                # WRITE EDGES
-                # ['subject', 'edge_label', 'object', 'relation', 'comment']
-                write_node_edge_item(fh=edge,
-                                     header=self.edge_header,
-                                     data=[drug_id,
-                                           drug_gene_edge_label,
-                                           gene_id,
-                                           drug_gene_edge_relation,
-                                           items_dict['ACT_COMMENT']])
+                for gene_id in gene_ids:
+                    gene_id = gene_curie_prefix + gene_id
+                    write_node_edge_item(fh=node,
+                                         header=self.node_header,
+                                         data=[gene_id,
+                                               items_dict['GENE'],
+                                               gene_node_type])
+
+                    # WRITE EDGES
+                    # ['subject', 'edge_label', 'object', 'relation', 'comment']
+                    write_node_edge_item(fh=edge,
+                                         header=self.edge_header,
+                                         data=[drug_id,
+                                               drug_gene_edge_label,
+                                               gene_id,
+                                               drug_gene_edge_relation,
+                                               items_dict['ACT_COMMENT']])
 
         return None
 
@@ -106,8 +112,9 @@ def parse_drug_central_line(this_line: str, header_items: List) -> Dict:
         item_dict: A dictionary of header items and a processed Drug Central string.
     """
 
-    items = this_line.strip().split("\t")
-    item_dict = dict(zip(header_items, items))
+    data = this_line.strip().split("\t")
+    data = [i.replace('"', '') for i in data]
+    item_dict = data_to_dict(header_items, data)
 
     return item_dict
 
diff --git a/tests/resources/drug.target.interaction_SNIPPET.tsv b/tests/resources/drug.target.interaction_SNIPPET.tsv
@@ -0,0 +1,16 @@
+"DRUG_NAME"	"STRUCT_ID"	"TARGET_NAME"	"TARGET_CLASS"	"ACCESSION"	"GENE"	"SWISSPROT"	"ACT_VALUE"	"ACT_UNIT"	"ACT_TYPE"	"ACT_COMMENT"	"ACT_SOURCE"	"RELATION"	"MOA"	"MOA_SOURCE"	"ACT_SOURCE_URL"	"MOA_SOURCE_URL"	"ACTION_TYPE"	"TDL"	"ORGANISM"
+"levobupivacaine"	4	"Sodium channel protein type 4 subunit alpha"	"Ion channel"	"P35499"	"SCN4A"	"SCN4A_HUMAN"					"WOMBAT-PK"		1	"CHEMBL"		"https://www.ebi.ac.uk/chembl/compound/inspect/CHEMBL1200749"	"BLOCKER"	"Tclin"	"Homo sapiens"
+"levobupivacaine"	4	"Cytochrome P450 2D6"	"Enzyme"	"P10635"	"CYP2D6"	"CP2D6_HUMAN"	6.706858517		"IC50"	"DRUGMATRIX: CYP450, 2D6 enzyme inhibition (substrate: 3-Cyano-7-ethoxycoumarin)"	"DRUG MATRIX"	"="						"Tclin"	"Homo sapiens"
+"levobupivacaine"	4	"Potassium voltage-gated channel subfamily H member 2"	"Ion channel"	"Q12809"	"KCNH2"	"KCNH2_HUMAN"	4.89		"IC50"	"Inhibition of wild-type human ERG channel expressed in HEK293 cells assessed as blockade of potassium tail current by whole-cell patch clamp technique"	"CHEMBL"	"="			"https://www.ebi.ac.uk/chembl/compound/inspect/CHEMBL1200749"			"Tclin"	"Homo sapiens"
+"levobupivacaine"	4	"Potassium voltage-gated channel subfamily D member 3"	"Ion channel"	"Q9UK17"	"KCND3"	"KCND3_HUMAN"	4.5		"IC50"		"WOMBAT-PK"	"="						"Tclin"	"Homo sapiens"
+"levobupivacaine"	4	"Prostaglandin E2 receptor EP1 subtype"	"GPCR"	"P34995"	"PTGER1"	"PE2R1_HUMAN"					"WOMBAT-PK"							"Tclin"	"Homo sapiens"
+"levobupivacaine"	4	"Sodium channel protein type 1 subunit alpha"	"Ion channel"	"P35498"	"SCN1A"	"SCN1A_HUMAN"	5.79		"IC50"		"WOMBAT-PK"	"="						"Tclin"	"Homo sapiens"
+"levobupivacaine"	4	"Potassium voltage-gated channel subfamily A member 5"	"Ion channel"	"P22460"	"KCNA5"	"KCNA5_HUMAN"					"WOMBAT-PK"							"Tclin"	"Homo sapiens"
+"levobupivacaine"	4	"5-hydroxytryptamine receptor 3A"	"Ion channel"	"P46098"	"HTR3A"	"5HT3A_HUMAN"					"WOMBAT-PK"							"Tclin"	"Homo sapiens"
+"(S)-nicardipine"	5	"Voltage-gated L-type calcium channel"	"Ion channel"	"Q01668|Q13936"	"CACNA1C|CACNA1D"	"CAC1C_HUMAN|CAC1D_HUMAN"				"Mechanism of Action"	"DRUG LABEL"		1	"DRUG LABEL"	"http://www.accessdata.fda.gov/drugsatfda_docs/label/2009/022276s003lbl.pdf"	"http://www.accessdata.fda.gov/drugsatfda_docs/label/2009/022276s003lbl.pdf"	"BLOCKER"	"Tclin|Tclin"	"Homo sapiens"
+"(S)-nitrendipine"	6	"Intermediate conductance calcium-activated potassium channel protein 4"	"Ion channel"	"O15554"	"KCNN4"	"KCNN4_HUMAN"	7.599999905		"IC50"		"IUPHAR"	"="			"http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2334"		"BLOCKER"	"Tchem"	"Homo sapiens"
+"(S)-nitrendipine"	6	"Voltage-dependent L-type calcium channel subunit alpha-1F"	"Ion channel"	"O60840"	"CACNA1F"	"CAC1F_HUMAN"	6		"IC50"		"IUPHAR"	"="			"http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2334"		"NEGATIVE MODULATOR"	"Tchem"	"Homo sapiens"
+"(S)-nitrendipine"	6	"Voltage-dependent L-type calcium channel subunit alpha-1C"	"Ion channel"	"P22002"	"Cacna1c"	"CAC1C_RAT"	6		"IC50"		"IUPHAR"	"="			"http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2334"		"NEGATIVE MODULATOR"		"Rattus norvegicus"
+"(S)-nitrendipine"	6	"Voltage-dependent L-type calcium channel subunit alpha-1S"	"Ion channel"	"Q02485"	"Cacna1s"	"CAC1S_RAT"	6		"IC50"		"IUPHAR"	"="			"http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2334"		"NEGATIVE MODULATOR"		"Rattus norvegicus"
+"(S)-nitrendipine"	6	"Voltage-dependent L-type calcium channel subunit alpha-1D"	"Ion channel"	"Q01668"	"CACNA1D"	"CAC1D_HUMAN"	8.39999961853027		"IC50"	"Mechanism of Action"	"IUPHAR"	"="	1	"IUPHAR"	"http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2334"	"http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2334"	"BLOCKER"	"Tclin"	"Homo sapiens"
+"(S)-nitrendipine"	6	"Voltage-dependent L-type calcium channel subunit alpha-1C"	"Ion channel"	"Q13936"	"CACNA1C"	"CAC1C_HUMAN"				"Mechanism of Action"	"SCIENTIFIC LITERATURE"		1	"SCIENTIFIC LITERATURE"		"http://www.ncbi.nlm.nih.gov/pubmed/17276408"	"BLOCKER"	"Tclin"	"Homo sapiens"
diff --git a/tests/test_drug_central.py b/tests/test_drug_central.py
@@ -0,0 +1,41 @@
+import unittest
+
+from kg_covid_19.transform_utils.drug_central.drug_central import \
+    parse_drug_central_line
+from kg_covid_19.utils.transform_utils import parse_header
+from parameterized import parameterized
+
+
+class TestDrugCentral(unittest.TestCase):
+
+    def setUp(self) -> None:
+        self.dti_fh = open('tests/resources/drug.target.interaction_SNIPPET.tsv', 'rt')
+
+    @parameterized.expand([
+     ('STRUCT_ID', '4'),
+     ('TARGET_NAME', 'Sodium channel protein type 4 subunit alpha'),
+     ('TARGET_CLASS', 'Ion channel'),
+     ('ACCESSION', 'P35499'),
+     ('GENE', 'SCN4A'),
+     ('SWISSPROT', 'SCN4A_HUMAN'),
+     ('ACT_VALUE', ''),
+     ('ACT_UNIT', ''),
+     ('ACT_TYPE', ''),
+     ('ACT_COMMENT', ''),
+     ('ACT_SOURCE', 'WOMBAT-PK'),
+     ('RELATION', ''),
+     ('MOA', '1'),
+     ('MOA_SOURCE', 'CHEMBL'),
+     ('ACT_SOURCE_URL', ''),
+     ('MOA_SOURCE_URL', 'https://www.ebi.ac.uk/chembl/compound/inspect/CHEMBL1200749'),
+     ('ACTION_TYPE', 'BLOCKER'),
+     ('TDL', 'Tclin'),
+     ('ORGANISM', 'Homo sapiens')])
+    def test_parse_drug_central_line(self, key, value):
+        header = parse_header(self.dti_fh.readline())
+        line = self.dti_fh.readline()
+        parsed = parse_drug_central_line(line, header)
+        self.assertTrue(key in parsed)
+        self.assertEqual(value, parsed[key])
+
+