Updates to WormBase processor

indra/assemblers/indranet/net.py

@@ -52,7 +52,7 @@ def from_df(cls, df):
             edge data for one edge. Indices are used to distinguish
             multiedges between a pair of nodes. Any columns not part of the
             below mentioned mandatory columns are considered extra attributes.
-            Columns starting with 'agA\_' or 'agB\_' (excluding the agA/B_name)
+            Columns starting with 'agA_' or 'agB_' (excluding the agA/B_name)
             will be added to its respective nodes as node attributes. Any other
             columns will be added as edge attributes.
 

indra/sources/wormbase/api.py

@@ -4,21 +4,33 @@
 from .processor import WormBaseProcessor
 from collections import namedtuple
 import pandas as pd
+import os
 
 # Url for all C. elegans molecular interactions data file
-wormbase_mol_file_url = ('https://fms.alliancegenome.org/download/'
+alliance_mol_int_file_url = ('https://fms.alliancegenome.org/download/'
                          'INTERACTION-MOL_WB.tsv.gz')
 # Url for all C. elegans genetic interactions data file
-wormbase_gen_file_url = ('https://fms.alliancegenome.org/download/'
+alliance_gen_int_file_url = ('https://fms.alliancegenome.org/download/'
                          'INTERACTION-GEN_WB.tsv.gz')
 # Url for wormbase-to-entrez ID mapping
 wormbase_entrez_mappings_file_url = ('https://ftp.ncbi.nih.gov/gene/'
                                      'DATA/GENE_INFO/Invertebrates/'
                                      'Caenorhabditis_elegans.gene_info.gz')
 
+wormbase_int_file_url = ('https://downloads.wormbase.org/species/c_elegans/'
+                         'annotation/interactions/'
+                         'c_elegans.PRJNA13758.current.interactions.txt.gz')
+
+# wormbase_all_genes_file_path = ("/Users/bradleybuchner/Desktop/Grad School/Research"
+#                                 "/Aging Project/indra/indra/resources/c_elegans_all_genes.txt")
+dir = os.path.dirname(os.path.abspath(__file__))
+resources_dir = os.path.dirname(os.path.dirname(dir))
+wormbase_all_genes_file_path = os.path.join(resources_dir, 'resources/c_elegans_all_genes.txt')
+
+
 # An explanation for each column of the interaction files are here:
 # https://github.com/HUPO-PSI/miTab/blob/master/PSI-MITAB27Format.md
-columns = ['ids_interactor_a', 'ids_interactor_b',
+alliance_int_columns = ['ids_interactor_a', 'ids_interactor_b',
            'alt_ids_interactor_a', 'alt_ids_interactor_b',
            'aliases_interactor_a', 'aliases_interactor_b',
            'interaction_detection_methods', 'publication_first_authors',
@@ -42,6 +54,49 @@
            'identification_method_participant_a',
            'identification_method_participant_b']
 
+wormbase_int_columns_raw = [
+    "WBInteractionID", "Interaction_type", "Interaction_subtype", "Summary",
+    "Citation", "Interactor1", "Common_name1", "Role1",
+    "Interactor2", "Common_name2", "Role2",
+    "Interactor3", "Common_name3", "Role3",
+    "Interactor4", "Common_name4", "Role4",
+    "Interactor5", "Common_name5", "Role5",
+    "Interactor6", "Common_name6", "Role6",
+    "Interactor7", "Common_name7", "Role7",
+    "Interactor8", "Common_name8", "Role8",
+    "Interactor9", "Common_name9", "Role9",
+    "Interactor10", "Common_name10", "Role10",
+    "Interactor11", "Common_name11", "Role11",
+    "Interactor12", "Common_name12", "Role12",
+    "Interactor13", "Common_name13", "Role13",
+    "Interactor14", "Common_name14", "Role14",
+    "Interactor15", "Common_name15", "Role15",
+    "Interactor16", "Common_name16", "Role16",
+    "Interactor17", "Common_name17", "Role17",
+    "Interactor18", "Common_name18", "Role18",
+    "Interactor19", "Common_name19", "Role19",
+    "Interactor20", "Common_name20", "Role20",
+    "Interactor21", "Common_name21", "Role21",
+    "Interactor22", "Common_name22", "Role22",
+    "Interactor23", "Common_name23", "Role23",
+    "Interactor24", "Common_name24", "Role24",
+    "Interactor25", "Common_name25", "Role25",
+    "Interactor26", "Common_name26", "Role26",
+    "Interactor27", "Common_name27", "Role27",
+    "Interactor28", "Common_name28", "Role28",
+    "Interactor29", "Common_name29", "Role29",
+    "Interactor30", "Common_name30", "Role30",
+    "Interactor31", "Common_name31", "Role31",
+    "Interactor32", "Common_name32", "Role32"
+    # May need to add more column names if the dataset is updated
+    # with interactions having more than 32 interactors
+]
+
+wormbase_int_columns = ["WBInteractionID", "Type", "Subtype",
+                                  "Effector_ID", "Effector", "Effector_Role",
+                                  "Affected_ID", "Affected", "Affected_Role",
+                                  "Direction"]
+
 mapping_columns = ['tax_id', 'GeneID', 'Symbol',
                    'LocusTag', 'Synonyms', 'dbXrefs', 'chromosome',
                    'map_location', 'description', 'type_of_gene',
@@ -50,19 +105,28 @@
                    'Nomenclature_status', 'Other_designations',
                    'Modification_date', 'Feature_type']
 
-_WormBaseRow = namedtuple('WormBaseRow', columns)
+wormbase_all_genes_columns = ['source', 'bioentity_internal_id', 'bioentity_label',
+                              'synonym', 'database_xref', 'type', 'taxon',
+                              'taxon_label']
 
+_AllianceGenomeRow = namedtuple('AllianceGenomeRow', alliance_int_columns)
+_WormBaseRow = namedtuple('WormBaseRow', wormbase_int_columns)
 
-def process_from_files(wormbase_gen_data_file, wormbase_mol_data_file,
-                       wb_to_entrez_mappings_file):
+
+def process_from_files(alliance_gen_data_file, alliance_mol_data_file,
+                       wb_all_int_data_file, wb_to_entrez_mappings_file):
     """Process WormBase interaction data from TSV files.
 
     Parameters
     ----------
-    wormbase_gen_data_file : str
-        Path to the WormBase genetic interactions data file in TSV format.
-    wormbase_mol_data_file : str
-        Path to the WormBase molecular interactions data file in TSV format.
+    alliance_gen_data_file : str
+        Path to the Alliance Genome dataset of C. elegans genetic interactions
+        in TSV format.
+    alliance_mol_data_file : str
+        Path to the Alliance Genome dataset of C. elegans molecular interactions
+        in TSV format.
+    wb_all_int_data_file : str
+        Path to the WormBase dataset of all C. elegans interactions
     wb_to_entrez_mappings_file : str
         Path to the WormBase-to-Entrez ID mapping file in TSV format.
 
@@ -72,13 +136,21 @@ def process_from_files(wormbase_gen_data_file, wormbase_mol_data_file,
         WormBaseProcessor containing Statements extracted from the
         interactions data.
     """
-    gen_iter = pd.read_csv(wormbase_gen_data_file, sep='\t', comment='#',
+    gen_int_iter = pd.read_csv(alliance_gen_data_file, sep='\t', comment='#',
                            dtype=str).values.tolist()
-    mol_iter = pd.read_csv(wormbase_mol_data_file, sep='\t', comment='#',
+    mol_int_iter = pd.read_csv(alliance_mol_data_file, sep='\t', comment='#',
                            dtype=str).values.tolist()
+
+    all_wb_int = pd.read_csv(wb_all_int_data_file, sep='\t', comment='#',
+                             header=None, names=wormbase_int_columns_raw,
+                             dtype=str, compression='gzip')
+    all_wb_int_iter = _process_wormbase_interactions(all_wb_int).values.tolist()
+
     mappings_df = pd.read_csv(wb_to_entrez_mappings_file, sep='\t',
                               comment='#', dtype=str, names=mapping_columns)
-    return _processor_from_data(gen_iter, mol_iter, mappings_df)
+
+    # return _processor_from_data(gen_int_iter, mol_int_iter, mappings_df)
+    return _processor_from_data(gen_int_iter, mol_int_iter, all_wb_int_iter, mappings_df)
 
 
 def process_from_web():
@@ -89,37 +161,182 @@ def process_from_web():
     indra.sources.wormbase.WormBaseProcessor
         WormBaseProcessor containing Statements extracted from the interactions data.
     """
-    gen_iter = pd.read_csv(wormbase_gen_file_url, sep='\t', comment='#',
+    gen_int_iter = pd.read_csv(alliance_gen_int_file_url, sep='\t', comment='#',
                            dtype=str).values.tolist()
-    mol_iter = pd.read_csv(wormbase_mol_file_url, sep='\t', comment='#',
+    mol_int_iter = pd.read_csv(alliance_mol_int_file_url, sep='\t', comment='#',
                            dtype=str).values.tolist()
-    mappings_df = pd.read_csv(wormbase_entrez_mappings_file_url, sep='\t',
+
+    all_wb_int = pd.read_csv(wormbase_int_file_url, sep='\t', comment='#',
+                             header=None, names=wormbase_int_columns_raw,
+                             dtype=str, compression='gzip')
+    all_wb_int_iter = _process_wormbase_interactions(all_wb_int).values.tolist()
+
+    entrez_mappings_df = pd.read_csv(wormbase_entrez_mappings_file_url, sep='\t',
                               comment='#', dtype=str,
                               names=mapping_columns)
 
-    return _processor_from_data(gen_iter, mol_iter, mappings_df)
+    all_genes_df = pd.read_csv(wormbase_all_genes_file_path, sep="\t",
+                               header=None, dtype=str,
+                               names=wormbase_all_genes_columns)
+
+    # return _processor_from_data(gen_int_iter, mol_int_iter, entrez_mappings_df)
+    return _processor_from_data(gen_int_iter, mol_int_iter, all_wb_int_iter,
+                                entrez_mappings_df, all_genes_df)
+
+
+def _process_wormbase_interactions(df):
+    """Process raw interaction data file from WormBase. Expand all interactions
+    into two-way interactions.
+
+    Parameters
+    ----------
+    df : pandas.DataFrame
+        Raw interaction data file from WormBase.
+
+    Returns
+    -------
+    processed_df : pandas.DataFrame
+        Processed (pairwise) interaction data file from WormBase.
+    """
+
+    from itertools import product
+
+    # Get relevant columns
+    interactor_cols = [col for col in df.columns if col.startswith('Interactor')]
+    common_name_cols = [col for col in df.columns if col.startswith('Common_name')]
+    role_cols = [col for col in df.columns if col.startswith('Role')]
+
+    # Create one row per effector-affected pair
+    seen = set()
+    table_rows = []
+    for _, row in df.iterrows():
+        participants = []
+        effectors = []
+        affected = []
+        non_directional = []
+        gene_ids = {}
+        effector_roles = {}
+        affected_roles = {}
+
+        for i in range(len(interactor_cols)):
+            gene_id = row.get(interactor_cols[i])
+            gene_name = row.get(common_name_cols[i])
+            role_raw = row.get(role_cols[i])
+
+            if pd.isnull(gene_id) or pd.isnull(role_raw):
+                continue  # skip incomplete entries
+
+            role = str(role_raw).strip().lower()
+            gene_display = gene_name if pd.notnull(gene_name) else gene_id
+            gene_display = str(gene_display).strip()
+
+            if not gene_display:
+                continue  # skip blank names
+
+            gene_ids[gene_display] = gene_id
+            participants.append(gene_display)
+
+            if role in ["effector", "target", "trans_regulator"]:
+                effectors.append(gene_display)
+                effector_roles[gene_display] = role
+            elif role in ["affected", "bait", "trans_regulated"]:
+                affected.append(gene_display)
+                affected_roles[gene_display] = role
+            elif role in ["non_directional"]:
+                non_directional.append(gene_display)
 
+        interaction_id = row.get("WBInteractionID", "-")
+        interaction_type = row.get("Interaction_type", "-").lower()
+        interaction_subtype = row.get("Interaction_subtype", "-").lower()
 
-def _processor_from_data(gen_iter, mol_iter, mappings_df):
+        if len(affected) > 0 and len(effectors) < 1:
+            continue
+
+        if len(effectors) > 0 and len(affected) < 1:
+            continue
+
+        if len(participants) < 2:
+            continue
+
+        # For directional edges
+        for eff, aff in product(effectors, affected):
+            key = (eff, aff, "Effector->Affected")
+            if key not in seen:
+                seen.add(key)
+                table_rows.append({
+                    "WBInteractionID": interaction_id,
+                    "Type": interaction_type,
+                    "Subtype": interaction_subtype,
+                    "Effector_ID": gene_ids.get(eff, "-"),
+                    "Effector": eff,
+                    "Effector_Role": effector_roles.get(eff, "-"),
+                    "Affected_ID": gene_ids.get(aff, "-"),
+                    "Affected": aff,
+                    "Affected_Role": affected_roles.get(aff, "-"),
+                    "Direction": "Effector->Affected"
+                })
+
+        # For non-directional edges
+        if not effectors and not affected and len(non_directional) >= 2:
+            for i in range(len(non_directional)):
+                for j in range(i + 1, len(non_directional)):
+                    eff, aff = non_directional[i], non_directional[j]
+                    key = tuple(sorted([eff, aff])) + ("non-directional",)
+                    if key not in seen and eff != aff:
+                        seen.add(key)
+                        table_rows.append({
+                            "WBInteractionID": interaction_id,
+                            "Type": interaction_type,
+                            "Subtype": interaction_subtype,
+                            "Effector_ID": gene_ids.get(eff, "-"),
+                            "Effector": eff,
+                            "Effector_Role": "non-directional",
+                            "Affected_ID": gene_ids.get(aff, "-"),
+                            "Affected": aff,
+                            "Affected_Role": "non-directional",
+                            "Direction": "non-directional"
+                        })
+
+    # Convert to DataFrame
+    processed_df = pd.DataFrame(table_rows)
+
+    return processed_df
+
+
+def _processor_from_data(gen_int_iter, mol_int_iter, all_wb_int_iter,
+                         entrez_mappings_df, all_genes_df):
     """Create a WormBaseProcessor from the interaction data and ID mappings.
 
     Parameters
     ----------
-    gen_iter : list
-        Iterable of rows in the genetic interactions data file.
-    mol_iter : list
-        Iterable of rows in the molecular interactions data file.
-    mappings_df : pd.DataFrame
+    gen_int_iter : list
+        Iterable of rows in the Alliance Genome genetic interactions data file.
+    mol_int_iter : list
+        Iterable of rows in the Alliance Genome molecular interactions data file.
+    all_wb_int_iter : list
+        Iterable of rows in the WormBase interactions data file.
+    entrez_mappings_df : pd.DataFrame
         DataFrame containing associated WormBase and Entrez IDs.
+    all_genes_df : pd.DataFrame
+        DataFrame of all WormBase genes and their WormBase IDs,
+        symbols, and synonyms.
 
     Returns
     -------
     indra.sources.wormbase.WormBaseProcessor
         WormBaseProcessor containing Statements extracted from the interactions data.
     """
     # Process into a list of WormBaseRow namedtuples
-    all_rows = gen_iter + mol_iter
-    data = [_WormBaseRow(*[None if item == '-' else item
-                           for item in row][:len(columns)])
-            for row in all_rows]
-    return WormBaseProcessor(data, mappings_df)
+    alliance_rows = gen_int_iter + mol_int_iter
+    alliance_data = [_AllianceGenomeRow(*[None if item == '-' else item
+                           for item in row][:len(alliance_int_columns)])
+            for row in alliance_rows]
+    wormbase_rows = all_wb_int_iter
+    wormbase_data = [_WormBaseRow(*[None if item == '-' else item
+                           for item in row][:len(wormbase_int_columns)])
+            for row in wormbase_rows]
+
+    # return WormBaseProcessor(data, entrez_mappings_df)
+    return WormBaseProcessor(alliance_data, wormbase_data,
+                             entrez_mappings_df, all_genes_df)
+