Merge pull request #21 from PathwayMerger/bel-specification

Bel specification

setup.cfg

@@ -55,8 +55,8 @@ install_requires =
     bio2bel==0.2.1
     bio2bel_hgnc==0.2.2
     bio2bel_chebi==0.2.1
-    bio2bel_kegg==0.2.3
-    bio2bel_wikipathways==0.2.2
+    bio2bel_kegg==0.2.5
+    bio2bel_wikipathways==0.2.3
     bio2bel_reactome==0.2.3
     pybel==0.13.2
     pybel-tools>=0.7.2

src/pathme/constants.py

@@ -90,15 +90,16 @@ def ensure_pathme_folders(): # TODO why is this a function?
     'methylation': 'Me',
 }
 KEGG_CITATION = '10592173'
+REACTOME_CITATION = '29145629'
 
 # FIXME why doesn't this just import the compath_resources package?
 KEGG_WIKIPATHWAYS_MAPPINGS = 'https://github.com/ComPath/curation/raw/master/mappings/kegg_wikipathways.xlsx'
 KEGG_REACTOME_MAPPINGS = 'https://github.com/ComPath/curation/raw/master/mappings/kegg_reactome.xlsx'
 WIKIPATHWAYS_REACTOME_MAPPINGS = 'https://github.com/ComPath/curation/raw/master/mappings/wikipathways_reactome.xlsx'
 
 KEGG_KGML_URL = 'http://rest.kegg.jp/get/{}/kgml'
-RDF_REACTOME = ' ftp://ftp.ebi.ac.uk/pub/databases/RDF/reactome/r67/reactome-biopax.tar.bz2'
-RDF_WIKIPATHWAYS = 'http://data.wikipathways.org/20190310/rdf/wikipathways-20190310-rdf-wp.zip'
+RDF_REACTOME = 'ftp://ftp.ebi.ac.uk/pub/databases/RDF/reactome/r67/reactome-biopax.tar.bz2'
+RDF_WIKIPATHWAYS = 'http://data.wikipathways.org/20190610/rdf/wikipathways-20190610-rdf-wp.zip'
 
 KEGG_STATS_COLUMN_NAMES = {
     'nodes': 'BEL Nodes',

src/pathme/export_utils.py

@@ -4,7 +4,7 @@
 
 import logging
 import os
-from typing import List, Iterable
+from typing import Iterable, List, Tuple
 
 import click
 import networkx as nx
@@ -36,7 +36,7 @@ def add_annotation_key(graph):
             graph[u][v][k][ANNOTATIONS] = {}
 
 
-def get_all_pickles(kegg_path, reactome_path, wikipathways_path):
+def get_all_pickles(kegg_path: str, reactome_path: str, wikipathways_path: str) -> Tuple[List, List, List]:
     """Return a list with all pickle paths."""
     kegg_pickles = get_paths_in_folder(kegg_path)
 

src/pathme/kegg/convert_to_bel.py

@@ -10,7 +10,7 @@
 from pybel import BELGraph, to_pickle
 from pybel.dsl.edges import activity
 from pybel.dsl.node_classes import CentralDogma
-from pybel.dsl.nodes import abundance, bioprocess, complex_abundance, composite_abundance, pmod, protein, reaction
+from pybel.dsl.nodes import abundance, bioprocess, complex_abundance, composite_abundance, pmod, protein, reaction, rna
 from pybel.struct.summary import count_functions, edge_summary
 
 from pathme.constants import *
@@ -171,7 +171,8 @@ def gene_to_bel_node(graph, node):
                 return protein_node
 
             elif UNIPROT in attribute:
-                protein_node = protein(namespace=UNIPROT.upper(), name=attribute[UNIPROT], identifier=attribute[UNIPROT])
+                protein_node = protein(namespace=UNIPROT.upper(), name=attribute[UNIPROT],
+                                       identifier=attribute[UNIPROT])
                 graph.add_node_from_data(protein_node)
                 return protein_node
 
@@ -505,13 +506,23 @@ def add_simple_edge(graph, u, v, relation_type):
 
                 # Add increases edge if pmod subtype is coupled with activation subtype
                 if relation_type[0] == 'activation':
-                    graph.add_increases(u, v_modified, citation='', evidence='', subject_modifier=activity(),
-                                        annotations={})
+                    graph.add_increases(
+                        u, v_modified,
+                        citation=KEGG_CITATION, evidence='Extracted from KEGG',
+                        subject_modifier=activity() if u in {protein, complex_abundance, rna} else None,
+                        # Add the activity function if subject is one of the following nodes (BEL 2.0 specifications)
+                        annotations={},
+                    )
 
                 # Add decreases edge if pmod subtype is coupled with inhibition subtype
                 elif relation_type[0] == 'inhibition':
-                    graph.add_decreases(u, v_modified, citation='', evidence='', subject_modifier=activity(),
-                                        annotations={})
+                    graph.add_decreases(
+                        u, v_modified,
+                        citation=KEGG_CITATION, evidence='Extracted from KEGG',
+                        subject_modifier=activity() if u in {protein, complex_abundance, rna} else None,
+                        # Add the activity function if subject is one of the following nodes (BEL 2.0 specifications)
+                        annotations={},
+                    )
 
         # TODO: add pmod of v activates v
         # TODO: how to represent abundance modification in BEL?
@@ -522,48 +533,72 @@ def add_simple_edge(graph, u, v, relation_type):
         # If the object is a gene, miRNA, RNA, or protein, add protein modification
         if isinstance(v, CentralDogma):
             v_modified = v.with_variants(pmod(KEGG_MODIFICATIONS[relation_type]))
-            graph.add_increases(u, v_modified, citation='', evidence='', subject_modifier=activity(),
-                                annotations={})
+            graph.add_increases(
+                u, v_modified,
+                citation=KEGG_CITATION, evidence='Extracted from KEGG',
+                subject_modifier=activity() if u in {protein, complex_abundance, rna} else None,
+                annotations={},
+            )
 
     # Subject activity decreases protein modification (i.e. dephosphorylation) of object
     elif relation_type == 'dephosphorylation':
 
         # If the object is a gene, miRNA, RNA, or protein, add protein modification
         if isinstance(v, CentralDogma):
             v = v.with_variants(pmod('Ph'))
-        graph.add_decreases(u, v, citation=KEGG_CITATION, evidence='', subject_modifier=activity(), annotations={})
+        graph.add_decreases(
+            u, v,
+            citation=KEGG_CITATION, evidence='Extracted from KEGG',
+            subject_modifier=activity() if u in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     # Subject increases activity of object
     elif relation_type == 'activation':
-        graph.add_increases(u, v, citation=KEGG_CITATION, evidence='', object_modifier=activity(), annotations={})
+        graph.add_increases(
+            u, v,
+            citation=KEGG_CITATION, evidence='Extracted from KEGG',
+            object_modifier=activity() if v in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     # Catalytic activity of subject increases transformation of reactant(s) to product(s)
     elif relation_type in {'reversible', 'irreversible'}:
-        graph.add_increases(u, v, citation=KEGG_CITATION, evidence='', subject_modifier=activity('cat'), annotations={})
+        graph.add_increases(
+            u, v,
+            citation=KEGG_CITATION, evidence='Extracted from KEGG',
+            subject_modifier=activity('cat') if u in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     # Subject decreases activity of object
     elif relation_type == 'inhibition':
-        graph.add_decreases(u, v, citation=KEGG_CITATION, evidence='', object_modifier=activity(), annotations={})
+        graph.add_decreases(
+            u, v,
+            citation=KEGG_CITATION, evidence='Extracted from KEGG',
+            object_modifier=activity() if v in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     # Indirect effect and binding/association are noted to be equivalent relation types
     elif relation_type in {'indirect effect', 'binding/association'}:
-        graph.add_association(u, v, citation=KEGG_CITATION, evidence='', annotations={})
+        graph.add_association(u, v, citation=KEGG_CITATION, evidence='Extracted from KEGG', annotations={})
 
     # Subject increases expression of object
     elif relation_type == 'expression':
 
         # Expression object is converted to RNA abundance
         if isinstance(v, CentralDogma):
             v = v.get_rna()
-        graph.add_increases(u, v, citation=KEGG_CITATION, evidence='', annotations={})
+        graph.add_increases(u, v, citation=KEGG_CITATION, evidence='Extracted from KEGG', annotations={})
 
     # Subject decreases expression of object
     elif relation_type == 'repression':
 
         # Repression object is converted to RNA abundance
         if isinstance(v, CentralDogma):
             v = v.get_rna()
-        graph.add_decreases(u, v, citation=KEGG_CITATION, evidence='', annotations={})
+        graph.add_decreases(u, v, citation=KEGG_CITATION, evidence='Extracted from KEGG', annotations={})
 
     elif relation_type in {'dissociation', 'hidden compound', 'missing interaction', 'state change'}:
         pass

src/pathme/kegg/kegg_xml_parser.py

@@ -6,8 +6,8 @@
 import json
 import logging
 import os
-from xml.etree.ElementTree import parse
 from collections import defaultdict
+from xml.etree.ElementTree import parse
 
 import requests
 from bio2bel_kegg.constants import API_KEGG_GET
@@ -92,10 +92,12 @@ def _post_process_api_query(node_meta_data, hgnc_manager, chebi_manager):
                     for chebi_id in identifier.split(' '):
                         chebi_entry = chebi_manager.get_chemical_by_chebi_id(chebi_id)
 
-                        if not chebi_entry:
-                            continue
-
-                        node_dict[CHEBI_NAME] = chebi_entry.name
+                        # If the id is found in the database stick the name
+                        if chebi_entry:
+                            node_dict[CHEBI_NAME] = chebi_entry.name
+                        # Else use the default name by KEGG to ensure the name makes it into the graph
+                        elif "ENTRY_NAME" in node_meta_data:
+                            node_dict[CHEBI_NAME] = node_meta_data["ENTRY_NAME"]
 
     return node_dict
 

src/pathme/reactome/convert_to_bel.py

@@ -7,7 +7,6 @@
 
 from bio2bel_chebi import Manager as ChebiManager
 from bio2bel_hgnc import Manager as HgncManager
-
 from pybel import BELGraph
 from pybel.dsl import (
     abundance,
@@ -23,7 +22,7 @@
     NamedComplexAbundance
 )
 
-from pathme.constants import UNKNOWN
+from pathme.constants import UNKNOWN, REACTOME_CITATION
 from pathme.reactome.utils import get_valid_node_parameters, process_multiple_proteins
 from pathme.utils import parse_id_uri
 
@@ -36,6 +35,7 @@
 
 def convert_to_bel(nodes: Dict[str, Dict], interactions: List[Tuple[str, str, Dict]], pathway_info: Dict,
                    hgnc_manager: HgncManager, chebi_manager: ChebiManager) -> BELGraph:
+    """Convert RDF graph dictionary into BEL graph."""
     uri_id = pathway_info['uri_reactome_id']
 
     if uri_id != UNKNOWN:
@@ -69,7 +69,8 @@ def convert_to_bel(nodes: Dict[str, Dict], interactions: List[Tuple[str, str, Di
     return graph
 
 
-def nodes_to_bel(nodes: Dict[str, Dict], graph: BELGraph, hgnc_manager: HgncManager, chebi_manager: ChebiManager) -> Dict[str, BaseEntity]:
+def nodes_to_bel(nodes: Dict[str, Dict], graph: BELGraph, hgnc_manager: HgncManager, chebi_manager: ChebiManager) -> \
+        Dict[str, BaseEntity]:
     """Convert dictionary values to BEL nodes."""
     return {
         node_id: node_to_bel(node_att, graph, hgnc_manager, chebi_manager)
@@ -125,18 +126,16 @@ def node_to_bel(node: Dict, graph, hgnc_manager: HgncManager, chebi_manager: Che
                 namespace=namespace.upper()
             )
 
-
     elif 'Pathway' in node_types:
         bioprocess_node = bioprocess(identifier=identifier, name=name, namespace=namespace.upper())
         graph.add_node_from_data(bioprocess_node)
         return bioprocess_node
-
     else:
         log.warning('Entity type not recognized', node_types)
 
 
 def add_edges(graph: BELGraph, participants, nodes, att: Dict):
-    uri_id = att['uri_id']
+    """Add edges into the graph."""
     edge_types = att['interaction_type']
 
     if isinstance(participants, dict):
@@ -157,17 +156,25 @@ def add_edges(graph: BELGraph, participants, nodes, att: Dict):
     elif isinstance(participants, tuple):
         u = nodes[participants[0]]
         v = nodes[participants[1]]
-        add_simple_edge(graph, u, v, edge_types, uri_id)
+        add_simple_edge(graph, u, v, edge_types)
 
 
-def add_simple_edge(graph: BELGraph, u, v, edge_types, uri_id):
+def add_simple_edge(graph: BELGraph, u, v, edge_types):
+    """Add a simple edge into the graph."""
     if 'ACTIVATION' in edge_types:
-        # TODO anadir pubmed y descripcion
-        graph.add_increases(u, v, citation=uri_id, evidence='', object_modifier=activity(), annotations={})
+        graph.add_increases(
+            u, v,
+            citation=REACTOME_CITATION, evidence='Extracted from Reactome',
+            object_modifier=activity() if v in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     elif 'INHIBITION' in edge_types:
-        # TODO anadir pubmed y descripcion
-        graph.add_decreases(u, v, citation=uri_id, evidence='', object_modifier=activity(), annotations={})
-
+        graph.add_decreases(
+            u, v,
+            citation=REACTOME_CITATION, evidence='Extracted from Reactome',
+            object_modifier=activity() if v in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
     else:
         log.warning('edge type %s', edge_types)

src/pathme/wikipathways/convert_to_bel.py

@@ -200,20 +200,35 @@ def add_simple_edge(graph: BELGraph, u, v, edge_types, uri_id):
     :param uri_id: citation URI
     """
     if 'Stimulation' in edge_types:
-        graph.add_increases(u, v, citation=uri_id, evidence='', object_modifier=activity())
+        graph.add_increases(
+            u, v,
+            citation=uri_id, evidence='Extracted from WikiPathways',
+            object_modifier=activity() if v in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     elif 'Inhibition' in edge_types:
-        graph.add_decreases(u, v, citation=uri_id, evidence='', object_modifier=activity())
+        graph.add_decreases(
+            u, v,
+            citation=uri_id, evidence='Extracted from WikiPathways',
+            object_modifier=activity() if v in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     elif 'Catalysis' in edge_types:
-        graph.add_increases(u, v, citation=uri_id, evidence='', object_modifier=activity())
+        graph.add_increases(
+            u, v,
+            citation=uri_id, evidence='Extracted from WikiPathways',
+            object_modifier=activity() if v in {protein, complex_abundance, rna} else None,
+            annotations={},
+        )
 
     elif 'DirectedInteraction' in edge_types:
         graph.add_qualified_edge(
             u, v,
             relation=REGULATES,
             citation=uri_id,
-            evidence='',
+            evidence='Extracted from WikiPathways',
             annotations={
                 'EdgeTypes': edge_types,
             },