Update scripts for version 2.0

alzkb/populate_edge_weights.py

@@ -0,0 +1,215 @@
+import pandas as pd
+import os
+
+
+path = './data/alzkb_v2-populated.csv'
+df= pd.read_csv(path)
+df= pd.concat([df,pd.DataFrame(columns=['sourceDB','unbiased','affinity_nM','p_fisher','z_score','correlation','score','confidence'])])
+
+# hetionet-custom-edges.tsv
+data_dir = "./AlzKB_Raw_Data"
+hetionet_custom = pd.read_table(os.path.join(data_dir,'hetionet/hetionet-custom-edges.tsv'))
+
+hetio_custom = {
+    'CbG':'CHEMICALBINDSGENE', 
+    'DrD':'DISEASEASSOCIATESWITHDISEASE', # no results
+    'DlA':'DISEASELOCALIZESTOANATOMY',
+    'DpS':'SYMPTOMMANIFESTATIONOFDISEASE'
+}
+
+
+affinity_nM = hetionet_custom[hetionet_custom['metaedge']=='CbG']
+affinity_nM['xrefDrugbank'] = affinity_nM['source'].str.split('::').str[-1]
+affinity_nM['xrefNcbiGene'] = affinity_nM['target'].str.split('::').str[-1].astype(int)
+affinity_nM = affinity_nM.merge(df[['_id','xrefDrugbank']].rename(columns={'_id':'_start'}), on='xrefDrugbank', how='left')
+affinity_nM = affinity_nM.merge(df[['_id','xrefNcbiGene']].rename(columns={'_id':'_end'}), on='xrefNcbiGene', how='left')
+affinity_nM['_type'] = hetio_custom['CbG']
+merged_df = df.merge(affinity_nM, on=['_start', '_end', '_type'], suffixes=('', '_new'), how='left')
+for column in ['sourceDB', 'unbiased', 'affinity_nM']:
+    df[column] = merged_df[column + '_new'].combine_first(df[column])
+df.shape
+
+
+disgenet = pd.read_table('./AlzKB_Raw_Data/disgenet/CUSTOM/disease_mappings_alzheimer.tsv')
+disgenet = disgenet[disgenet['vocabulary']=='DO']
+
+
+p_fisher_DlA = hetionet_custom[hetionet_custom['metaedge']=='DlA']
+
+p_fisher_DlA['do_id'] = p_fisher_DlA['source'].str.split('::').str[-1].str.split(':').str[-1]
+p_fisher_DlA['xrefUberon'] = p_fisher_DlA['target'].str.split('::').str[-1]
+
+p_fisher_DlA = p_fisher_DlA.merge(disgenet, left_on='do_id', right_on= 'code')
+p_fisher_DlA['_start'] = 'disease_'+p_fisher_DlA['diseaseId'].str.lower()
+p_fisher_DlA = p_fisher_DlA.merge(df[['_id','xrefUberon']].rename(columns={'_id':'_end'}), on='xrefUberon', how='left')
+p_fisher_DlA['_type'] = hetio_custom['DlA']
+
+p_fisher_DpS = hetionet_custom[hetionet_custom['metaedge']=='DpS']
+
+p_fisher_DpS['xrefMeSH'] = p_fisher_DpS['target'].str.split('::').str[-1]
+p_fisher_DpS['do_id'] = p_fisher_DpS['source'].str.split('::').str[-1].str.split(':').str[-1]
+
+p_fisher_DpS = p_fisher_DpS.merge(df[['_id','xrefMeSH']].rename(columns={'_id':'_start'}), on='xrefMeSH', how='left')
+p_fisher_DpS = p_fisher_DpS.merge(disgenet, left_on='do_id', right_on= 'code')
+p_fisher_DpS['_end'] = 'disease_'+p_fisher_DpS['diseaseId'].str.lower()
+p_fisher_DpS['_type'] = hetio_custom['DpS']
+
+p_fisher = pd.concat([p_fisher_DlA, p_fisher_DpS])
+
+merged_df = df.merge(p_fisher, on=['_start', '_end', '_type'], suffixes=('', '_new'), how='left')
+for column in ['sourceDB', 'unbiased', 'p_fisher']:
+    df[column] = merged_df[column + '_new'].combine_first(df[column])
+df.shape
+
+
+# hetionet-v1.0-edges.sif
+#https://github.com/dhimmel/integrate/blob/master/integrate.ipynb
+
+import hetio.hetnet
+import hetio.readwrite
+import hetio.stats
+
+path = 'https://raw.githubusercontent.com/dhimmel/integrate/master/data/hetnet.json.bz2'
+graph = hetio.readwrite.read_graph(path, formatting=None)
+
+
+#https://github.com/hetio/hetnetpy/blob/main/hetnetpy/readwrite.py
+import collections
+import operator
+import pandas as pd
+
+def write_nodetable(graph):
+    """Write a tabular encoding of the graph nodes."""
+    rows = list()
+    for node in graph.node_dict.values():
+        row = collections.OrderedDict()
+        row["kind"] = node.metanode.identifier
+        row["id"] = str(node)
+        row["name"] = node.name
+        row["source"] = node.data['source']
+        rows.append(row)
+    rows.sort(key=operator.itemgetter("kind", "id"))
+    fieldnames = ["id", "name", "kind", "source"]
+    df_nodes_tsv = pd.DataFrame(rows, columns=fieldnames)
+    print(df_nodes_tsv.shape)
+    return df_nodes_tsv
+
+
+def write_edgetable(graph):
+    """Write a tsv of the graph edges."""
+    rows = list()
+    edge_properties=["sourceDB", "unbiased", "affinity_nM", "z_score", "p_fisher", "correlation"]
+    fieldnames =["source", "metaedge", "target"]
+    fieldnames = fieldnames+edge_properties
+    metaedge_to_edges = graph.get_metaedge_to_edges(exclude_inverts=True)
+    for metaedge, edges in metaedge_to_edges.items():
+        for edge in edges:
+            row = collections.OrderedDict()
+            row["source"] = edge.source
+            row["metaedge"] = edge.metaedge.abbrev
+            row["target"] = edge.target
+            for pro in edge_properties:
+                if pro =='sourceDB':
+                    if 'source' in edge.data.keys():
+                        row[pro]=edge.data['source']
+                    else:
+                        row[pro]=None
+                else:
+                    if pro in edge.data.keys():
+                        row[pro]=edge.data[pro]
+                    else:
+                        row[pro]=None
+            rows.append(row)
+    df_edges_tsv = pd.DataFrame(rows, columns=fieldnames)
+    print(df_edges_tsv.shape)
+    return df_edges_tsv
+
+hetionet = write_edgetable(graph)
+hetionet['source']=hetionet['source'].astype(str)
+hetionet['target']=hetionet['target'].astype(str)
+hetionet
+
+hetio = {
+    'CuG':'CHEMICALINCREASESEXPRESSION', 
+    'CdG':'CHEMICALDECREASESEXPRESSION',
+    'GcG':'GENECOVARIESWITHGENE',
+    'Gr>G':'GENEREGULATESGENE'
+}
+
+
+z_score = hetionet[hetionet['metaedge']=='CuG']
+z_score['xrefDrugbank'] = z_score['source'].str.split('::').str[-1]
+z_score['xrefNcbiGene'] = z_score['target'].str.split('::').str[-1].astype(int)
+
+z_score = z_score.merge(df[['_id','xrefDrugbank']].rename(columns={'_id':'_start'}), on='xrefDrugbank', how='left')
+z_score = z_score.merge(df[['_id','xrefNcbiGene']].rename(columns={'_id':'_end'}), on='xrefNcbiGene', how='left')
+z_score['_type'] = hetio['CuG']
+
+z_score_all = z_score
+
+z_score = hetionet[hetionet['metaedge']=='CdG']
+z_score['xrefDrugbank'] = z_score['source'].str.split('::').str[-1]
+z_score['xrefNcbiGene'] = z_score['target'].str.split('::').str[-1].astype(int)
+
+z_score = z_score.merge(df[['_id','xrefDrugbank']].rename(columns={'_id':'_start'}), on='xrefDrugbank', how='left')
+z_score = z_score.merge(df[['_id','xrefNcbiGene']].rename(columns={'_id':'_end'}), on='xrefNcbiGene', how='left')
+z_score['_type'] = hetio['CdG']
+
+z_score_all = pd.concat([z_score_all,z_score])
+
+merged_df = df.merge(z_score_all, on=['_start', '_end', '_type'], suffixes=('', '_new'), how='left')
+for column in ['sourceDB', 'unbiased', 'z_score']:
+    df[column] = merged_df[column + '_new'].combine_first(df[column])
+df.shape
+
+
+correlation = pd.read_table(os.path.join(data_dir,'hetionet/geneCovariesWithGene_correlation.tsv'))
+
+correlation = correlation.merge(df[['_id','xrefNcbiGene']].rename(columns={'_id':'_start'}), left_on='source_entrez', right_on='xrefNcbiGene', how='left')
+correlation = correlation.merge(df[['_id','xrefNcbiGene']].rename(columns={'_id':'_end'}), left_on='target_entrez', right_on='xrefNcbiGene', how='left')
+correlation['_type'] = hetio['GcG']
+correlation['sourceDB'] = 'Hetionet - ERC'
+correlation['unbiased'] = True
+
+merged_df = df.merge(correlation, on=['_start', '_end', '_type'], suffixes=('', '_new'), how='left')
+for column in ['sourceDB', 'unbiased', 'correlation']:
+    df[column] = merged_df[column + '_new'].combine_first(df[column])
+df.shape
+df.loc[~df['correlation'].isna()]
+
+
+#DisGeNET
+score = pd.read_table('./AlzKB_Raw_Data/disgenet/curated_gene_disease_associations.tsv')
+score['sourceDB'] = 'DisGeNET - '+score['source']
+
+score = score.merge(df[['_id','xrefNcbiGene']].rename(columns={'_id':'_start'}), left_on='geneId', right_on='xrefNcbiGene', how='left')
+score['_end'] = 'disease_'+score['diseaseId'].str.lower()
+score['_type'] = 'GENEASSOCIATESWITHDISEASE'
+
+merged_df = df.merge(score, on=['_start', '_end', '_type'], suffixes=('', '_new'), how='left')
+for column in ['sourceDB', 'score']:
+    df[column] = merged_df[column + '_new'].combine_first(df[column])
+df.shape
+
+
+#TF
+confidence = pd.read_table('./AlzKB_Raw_Data/dorothea/tf.tsv')
+confidence
+
+confidence = pd.read_table('./AlzKB_Raw_Data/dorothea/tf.tsv')
+
+confidence = confidence.merge(df[['_id','TF']].rename(columns={'_id':'_start'}), on='TF', how='left')
+confidence = confidence.merge(df[['_id','geneSymbol']].rename(columns={'_id':'_end'}), left_on='Gene', right_on='geneSymbol', how='left')
+
+confidence['_type'] = 'TRANSCRIPTIONFACTORINTERACTSWITHGENE'
+
+merged_df = df.merge(confidence, on=['_start', '_end', '_type'], suffixes=('', '_new'), how='left')
+for column in ['sourceDB', 'confidence']:
+    df[column] = merged_df[column + '_new'].combine_first(df[column])
+df.shape
+
+#save data file
+df.to_csv('./data/alzkb_v2.0.0_with_edge_properties.csv')
+
+
+