test with copd data

COPD_PI_test.py

@@ -0,0 +1,30 @@
+import pandas as pd
+import numpy as np
+import pathintegrate
+import sspa
+
+
+md = pd.read_csv('H:/Documents/pathway-integration/COPDgene/COPDgene_phonotype.txt', sep='\t')
+prot = pd.read_csv('H:/Documents/pathway-integration/COPDgene/COPDgene_proteomics_UniProt.csv', index_col=0)
+metab = pd.read_csv('H:/Documents/pathway-integration/COPDgene/COPDgene_metabolomics_CHEBI_mapped.csv', index_col=0)
+trans = pd.read_csv('D:/COPDgene/Processed/COPDgene_transcriptomics_filt_Q1_scaled.csv', index_col=0)
+metab['Group'] = metab.index.map(dict(zip(md['sid'], md["COPD"])))
+
+metab = metab[metab['Group'].isin([0, 1])]
+intersect_samples = set(metab.index.tolist()) & set(prot.index.tolist()) & set(trans.index.tolist())
+prot = prot.loc[intersect_samples, :]
+metab = metab.loc[intersect_samples, :]
+trans = trans.loc[intersect_samples, :]
+
+mo_paths_all = pd.read_csv('D:/Pathway_databases\Reactome_multi_omics_ChEBI_Uniprot_Ensembl.csv', index_col=0, dtype=object)
+
+pi_model = pathintegrate.PathIntegrate(omics_data={'Metabolomics': metab, 'Proteomics':prot.iloc[:, :-1], 'Transcriptomics': trans.iloc[:, :-1]}, 
+                                       metadata=metab['Group'], 
+                                       pathway_source=mo_paths_all, 
+                                       sspa_scoring='svd', 
+                                       min_coverage=2)
+
+copdgene_multi_view = pi_model.MultiView(ncomp=4)
+
+# launch the pathwy network explorer on a local server
+pathintegrate.launch_network_app(copdgene_multi_view, mo_paths_all)

PathIntegrate.py

@@ -81,9 +81,11 @@ def get_multi_omics_coverage(self):
         return coverage
 
     def MultiView(self, ncomp=2):
+        print('Generating pathway scores...')
         sspa_scores = [self.sspa_method(i, self.pathway_source, self.min_coverage, return_molecular_importance=True) for i in self.omics_data.values()]
         self.sspa_scores_mv = dict(zip(self.omics_data.keys(), [i[0] for i in sspa_scores]))
         # self.sspa_scores_mv = {k: self.sspa_method(v, self.pathway_source, self.min_coverage) for k, v in self.omics_data.items()}
+        print('Fitting MultiView model')
         mv = MBPLS(n_components=ncomp)
         mv.fit([i.copy(deep=True) for i in self.sspa_scores_mv.values()], self.labels)
 
@@ -107,13 +109,17 @@ def MultiView(self, ncomp=2):
 
     def SingleView(self, model=sklearn.linear_model.LogisticRegression, model_params=None):
         concat_data = pd.concat(self.omics_data.values(), axis=1)
+        print('Generating pathway scores...')
+
         sspa_scores =self.sspa_method(concat_data, self.pathway_source, self.min_coverage, return_molecular_importance=True)
         self.sspa_scores_sv = sspa_scores[0]
 
         if model_params:
             sv = model(**model_params)
         else:
             sv = model()
+        print('Fitting SingleView model')
+
         sv.fit(X=self.sspa_scores_sv, y=self.labels)
         sv.sspa_scores = self.sspa_scores_sv
         sv.name = 'SingleView'
@@ -142,31 +148,31 @@ def VIP_multiBlock(x_weights, x_superscores, x_loadings, y_loadings):
     vip_scores = np.sqrt(p * np.sum(sumsquares*(weights_norm**2), axis=1) / np.sum(sumsquares))
     return vip_scores
 
-metab = pd.read_csv('data/metabolomics_example.csv', index_col=0)
-prot = pd.read_csv('data/proteomics_example.csv', index_col=0)
+# metab = pd.read_csv('data/metabolomics_example.csv', index_col=0)
+# prot = pd.read_csv('data/proteomics_example.csv', index_col=0)
 
-# make possible to download MO paths from reactome
-# mo_paths = sspa.process_reactome(
-#     organism='Homo sapiens',
-#     download_latest=True,
-#     omics_type='multiomics',
-#     filepath='data/')
+# # make possible to download MO paths from reactome
+# # mo_paths = sspa.process_reactome(
+# #     organism='Homo sapiens',
+# #     download_latest=True,
+# #     omics_type='multiomics',
+# #     filepath='data/')
 
-# load pre-loaded pathways 
-mo_paths = sspa.process_gmt(infile='data/Reactome_Homo_sapiens_pathways_multiomics_R85.gmt')
+# # load pre-loaded pathways 
+# mo_paths = sspa.process_gmt(infile='data/Reactome_Homo_sapiens_pathways_multiomics_R85.gmt')
 
-pi_model = PathIntegrate({'Metabolomics': metab, 'Proteomics':prot.iloc[:, :-1]}, metadata=prot['Group'], pathway_source=mo_paths, sspa_scoring='svd', min_coverage=2)
+# pi_model = PathIntegrate({'Metabolomics': metab, 'Proteomics':prot.iloc[:, :-1]}, metadata=prot['Group'], pathway_source=mo_paths, sspa_scoring='svd', min_coverage=2)
 
 # covid_multi_view = pi_model.MultiView(ncomp=5)
 
-# launch the pathwy network explorer on a local server
+# # launch the pathwy network explorer on a local server
 # launch_network_app(covid_multi_view, mo_paths)
 
 # print(covid_multi_view.A_corrected_)
 # print(covid_multi_view.vip)
 
 # plot_functs.plot_block_importance(covid_multi_view)
 
-covid_single_view = pi_model.SingleView(model_params={'random_state':0})
-launch_network_app(covid_single_view, mo_paths)
+# covid_single_view = pi_model.SingleView(model_params={'random_state':0})
+# launch_network_app(covid_single_view, mo_paths)
 # print(covid_single_view.intercept_)

app.py

@@ -58,7 +58,7 @@ def find_root(G,child):
 # the default graph is the pathway hierarchy coloured by root pathway membership as defined by Reactome
 G = nx.from_pandas_edgelist(hierarchy_hsa, source=0, target=1, create_using=nx.DiGraph())
 hierarchy_hsa_all['Root'] = [find_root(G, i) for i in hierarchy_hsa_all[1]]
-root_cmap = dict(zip(set(hierarchy_hsa_all['Root']), sns.color_palette("hls", len(set(hierarchy_hsa_all['Root']))).as_hex()))
+root_cmap = dict(zip(set(hierarchy_hsa_all['Root']), sns.color_palette("husl", len(set(hierarchy_hsa_all['Root']))).as_hex()))
 cy_mo = nx.readwrite.json_graph.cytoscape_data(G)
 
 
@@ -346,7 +346,7 @@ def update_bar_chart(pathway):
                 pass
 
         pathway_df_molec = pd.concat(pathways_dfs, axis=1)
-        fig = make_subplots(rows=1, cols=len(pathways_dfs), shared_xaxes=False)
+        fig = make_subplots(rows=1, cols=len(pathways_dfs), shared_xaxes='rows')
 
         for i in range(0, len(pathways_dfs)):
             fig.add_trace(go.Bar(x=pathway_df_molec.index, y=pathway_df_molec.iloc[:,i], name=pathway_df_molec.columns[i]), row=1, col=i+1)
@@ -428,8 +428,6 @@ def launch_network_app(pi_model, pathway_source, hierarchy_source='preloaded'):
     # style=CONTENT_STYLE
     )
 
-
-
     app.layout = html.Div([
         navbar,
         sidebar,