MCPcopy Create free account
hub / github.com/Mew233/pairwise / process_CCLE

Function process_CCLE

pairwise/prepare_data.py:188–273  ·  view source on GitHub ↗
()

Source from the content-addressed store, hash-verified

186 function_mapping = {'CCLE':'process_CCLE', 'Customized':'process_Customized'}
187
188 def process_CCLE():
189
190 def load_file(postfix):
191 if args.train_test_mode == 'fine_tune':
192 df = pd.read_csv(os.path.join(ROOT_DIR, 'data', 'cell_line_data','CCLE','CCLE_%s.csv' % postfix),sep=',')
193 else:
194 df = pd.read_csv(os.path.join(ROOT_DIR, 'data', 'cell_line_data','CCLE','CCLE_%s.csv' % postfix),sep=',')
195
196
197 ## need to transform mut into one-hot dataframe
198 if postfix == 'mut':
199 # remove entre_id which is 0
200 mu = df.loc[df['Entrez_Gene_Id'].values !=0]
201
202 # transform data to one_hot format
203 clean_cells_ALL = list(set(list(mu['DepMap_ID'])))
204 clean_genes_ALL = list(set(list(mu['Entrez_Gene_Id'])))
205 CCLE_mu = pd.DataFrame(one_hot(mu,clean_cells_ALL,clean_genes_ALL), columns=['Entrez gene id']+clean_genes_ALL)
206
207
208 # use entrez gene id as index
209 if postfix != "mut":
210 df.columns = ['Entrez gene id']+[split_it_cell(_) for _ in list(df.columns)[1:]]
211 df_transpose = df.T
212 # set first row as column
213 df_transpose.columns = df_transpose.iloc[0]
214 processed_data = df_transpose.drop(df_transpose.index[0])
215 # use entrez gene id as index
216 if postfix != "mut":
217 if postfix == "GSVA_scores":
218 # X_tr, means1, std1, means2, std2, feat_filt = normalize(df, norm='tanh_norm')
219 # X_tr = pd.DataFrame(X_tr,columns=df.columns,index=df.index)
220 processed_data = df
221
222 else:
223 df.columns = ['Entrez gene id']+[split_it_cell(_) for _ in list(df.columns)[1:]]
224 df_transpose = df.T
225 # set first row as column
226 df_transpose.columns = df_transpose.iloc[0]
227 processed_data = df_transpose.drop(df_transpose.index[0])
228
229 else:
230 ## same as exp
231 mu_transpose = CCLE_mu.T
232 # set first row as column
233 mu_transpose.columns = mu_transpose.iloc[0]
234 processed_data = mu_transpose.drop(mu_transpose.index[0])
235 return processed_data
236
237 # load all cell line features
238 save_path = os.path.join(ROOT_DIR, 'data', 'cell_line_data','CCLE')
239 save_path = os.path.join(save_path, 'input_cellline_data.npy')
240 if not os.path.exists(save_path):
241 data_dicts = {}
242 for file_type in ['exp', 'cn', 'mut']:
243 data_dicts[ file_type ] = load_file(file_type)
244 # load GNN_cell
245 gnn_path = os.path.join(ROOT_DIR, 'data', 'cell_line_data','CCLE')

Callers

nothing calls this directly

Calls 1

load_fileFunction · 0.85

Tested by

no test coverage detected