Merge pull request #37 from COMBAT-TB/dev

Fixes for KEGG  and Reactome.

Author: thobalose

setup.py

@@ -5,7 +5,7 @@
 
 setup(
     name='tb2neo',
-    version='0.0.8',
+    version='0.0.9',
     url='https://github.com/COMBAT-TB/tb2neo',
     description='Builds a M.tb annotation graph database from GFF files',
     long_description=long_description,

tb2neo/dbconn.py

@@ -6,7 +6,7 @@
 import sys
 import zipfile
 
-from bioservices import KEGG, ChEMBL, QuickGO, reactome
+from bioservices import KEGG, ChEMBL, QuickGO, ReactomeOld
 from pandas import read_csv
 from py2neo import Graph
 from tqdm import tqdm
@@ -28,7 +28,7 @@
 chembl = ChEMBL(verbose=False)
 quick_go = QuickGO(verbose=False)
 quick_go.url = 'http://www.ebi.ac.uk/QuickGO-Old'
-reactome_old = reactome.ReactomeOld(verbose=False)
+reactome_old = ReactomeOld(verbose=False)
 kegg = KEGG(verbose=False)
 
 # watch("neo4j.bolt")
@@ -769,35 +769,63 @@ def create_kegg_pathways_nodes():
     :return:
     """
     sys.stdout.write("Creating KEGG Pathways...")
+
+    def map_pathway_to_proteins(pathway_genes, path):
+        for g_id in pathway_genes:
+            g_id = "Rv" + \
+                g_id.strip("RVBD_") if "RV" in g_id else g_id
+            # Protein parent is stored as an array
+            gene = Gene.select(graph, g_id).first()
+            if gene:
+                for protein in gene.encodes:
+                    protein.pathway.add(path)
+                    graph.push(protein)
+                    path.protein.add(protein)
+                    graph.push(path)
+
     # TODO: Add mtc
     organisms = ['mtu']
     for organism in organisms:
         kegg.organism = organism
         pathway_ids = kegg.pathwayIds
         for path in tqdm(pathway_ids):
             data = kegg.parse(kegg.get(path))
+            accession = path[path.find(organism):].strip()
             if isinstance(data, dict) is True:
                 pathway = Pathway()
-                pathway.accession = path[path.find(organism):].strip()
+                pathway.accession = accession
                 pathway._class = data.get('CLASS')
                 pathway.name = data['PATHWAY_MAP'].get(path.strip("path:"))
                 pathway.summation = data.get('DESCRIPTION')
                 pathway.species = data.get('ORGANISM')
                 graph.create(pathway)
                 if data.get('GENE'):
-                    for g_id in data['GENE'].keys():
-                        g_id = "Rv" + \
-                               g_id.strip("RVBD_") if "RV" in g_id else g_id
-                        # Protein parent is stored as an array
-                        gene = Gene.select(graph, g_id).first()
-                        if gene:
-                            for protein in gene.encodes:
-                                protein.pathway.add(pathway)
-                                graph.push(protein)
-                                pathway.protein.add(protein)
-                                graph.push(pathway)
+                    map_pathway_to_proteins(data['GENE'].keys(), pathway)
             else:
-                sys.stderr.write("Data is: {}\n".format(data))
+                res_split = data.split("\n")
+
+                # accession = res_split[0].split()[1]
+                name = res_split[1].split("-")[0].strip('NAME').strip()
+                summation = res_split[2].strip("DESCRIPTION").strip()
+                _class = res_split[3].strip("CLASS").strip()
+                # Create Pathway
+                pathway = Pathway()
+                pathway.accession = accession
+                pathway.name = name
+                pathway.summation = summation
+                pathway._class = _class
+                graph.create(pathway)
+
+                gene_str_list = [
+                    s.split() for s in res_split if
+                    "Rv" in s and 'Myco' not in s
+                ]
+                genes = {
+                    g for l in gene_str_list for g in l if
+                    str(g).isalnum() and 'Rv' in g
+                }
+
+                map_pathway_to_proteins(genes, pathway)
     sys.stdout.write("\nCreated KEGG Pathway Nodes.")
 
 

tb2neo/stringstitch.py

@@ -45,17 +45,14 @@ def fetch_string_data(gene, output_format='json', method='network',
 def load_string_data():
     df = read_csv(UNIPROT_DATA).fillna("")
     for entry in df.values:
-        try:
-            gene = eu_mapping(from_=entry[0], to='TUBERCULIST_ID')[0]
-        except TypeError:
-            print(f"A TypeError occurred for: {entry[0]}")
+        rv_tag = eu_mapping(from_=entry[0], to='TUBERCULIST_ID')
+        gene = rv_tag[0] if rv_tag else None
+        data = fetch_string_data(gene=gene) if gene else None
+        if data:
+            for ppi in data:
+                create_ppi(ppi["stringId_A"], ppi["stringId_B"], ppi["score"])
         else:
-            data = fetch_string_data(gene=gene) if gene else None
-            if data:
-                for ppi in data:
-                    create_ppi(
-                        ppi["stringId_A"], ppi["stringId_B"], ppi["score"]
-                    )
+            print(f"Nothing was found for {gene} with entry: {entry[0]}")
 
 
 def fetch_stitch_data():

tb2neo/uniprot.py

@@ -120,5 +120,5 @@ def eu_mapping(from_, to):
         if len(_map) != 0:
             xref_id = _map[from_]
     else:
-        raise ValueError("Can't map {} to {}".format(from_, to))
+        print(f"Can't map {from_} to {to}!")
     return xref_id

test/test_cli.py

@@ -102,7 +102,7 @@ def test_load_kegg_pathways(cli_runner):
     assert result.exit_code == 0
 
 
-@pytest.mark.skip(reason="'NoneType' object is not subscriptable")
+# @pytest.mark.skip(reason="'NoneType' object is not subscriptable")
 def test_load_reactome_pathways(cli_runner):
     result = cli_runner.invoke(load_reactome_pathways)
     assert result.exit_code == 0

test/test_dbconn.py

@@ -3,9 +3,9 @@
 """
 import pytest
 
-from tb2neo.dbconn import graph, split_gene_names, create_chromosome_nodes, \
-    create_publication_nodes
-from tb2neo.model.core import Gene, Chromosome
+from tb2neo.dbconn import (create_chromosome_nodes, create_publication_nodes,
+                           graph, split_gene_names)
+from tb2neo.model.core import Chromosome, Gene
 from test_cli import UNIPROT_DATA
 
 
@@ -29,8 +29,8 @@ def test_db_nodes():
 def test_rv0001():
     gene = Gene.select(graph, "Rv0001").first()
     assert gene.name == "dnaA"
-    assert gene.category is not ""
-    assert gene.residues is not ""
+    assert gene.category != ""
+    assert gene.residues != ""
 
 
 @pytest.mark.parametrize("test_input,expected", [
@@ -58,3 +58,13 @@ def test_split_gene_name(test_input, expected):
 def test_create_publication_nodes():
     res = create_publication_nodes(uniprot_data=UNIPROT_DATA)
     assert isinstance(res, set)
+
+
+# @pytest.mark.skip(reason="")
+def test_ppi_score():
+    cypher_q = f"MATCH (pa:Protein {{ uniquename: 'P9WQD9' }})"
+    cypher_q += f"-[r:INTERACTS_WITH]->"
+    cypher_q += f"(pb:Protein {{ uniquename: 'P9WIE5' }}) "
+    cypher_q += f"RETURN r.score"
+    data = graph.run(cypher_q).data()[0]
+    assert data['r.score'] == 0.881

test/test_uniprot.py

@@ -3,7 +3,7 @@
 import pytest
 
 from tb2neo.gffproc import get_locus_tags
-from tb2neo.uniprot import query_uniprot, eu_mapping
+from tb2neo.uniprot import eu_mapping, query_uniprot
 from test_cli import TEST_GFF
 
 
@@ -21,8 +21,3 @@ def test_search_uniprot():
 ])
 def test_eu_mapping(test_input, expected):
     assert isinstance(test_input, expected) is True
-
-
-def test_eu_mapping_error():
-    with pytest.raises(ValueError):
-        eu_mapping(None, to="PDB_ID")