()
| 26 | 'Customized':'process_customized'} |
| 27 | |
| 28 | def process_drugcomb(): |
| 29 | data = pd.read_csv(os.path.join(ROOT_DIR, 'data', 'synergy_data','p13','p13_trueset.csv')) |
| 30 | data = data[['drug_row', 'drug_col','cell_line_name','study_name','tissue_name',\ |
| 31 | 'synergy_zip','synergy_loewe','synergy_hsa','synergy_bliss','DepMap_ID','RRID',\ |
| 32 | 'pubchemID_x','compound0_x','pubchemID_y','compound0_y',\ |
| 33 | 'ri_row', 'ri_col']] |
| 34 | |
| 35 | data_trim = data[['compound0_x','compound0_y','DepMap_ID','tissue_name','synergy_loewe',\ |
| 36 | 'ri_row', 'ri_col']] |
| 37 | |
| 38 | ## clean the scores |
| 39 | data_trim = data_trim[data_trim['synergy_loewe'] != '\\N'] |
| 40 | data_trim['synergy_loewe'] = data_trim['synergy_loewe'].astype(float) |
| 41 | |
| 42 | data_trim['compound0_x'] = data_trim['compound0_x'].apply(lambda x: split_it(x)) |
| 43 | data_trim['compound0_y'] = data_trim['compound0_y'].apply(lambda x: split_it(x)) |
| 44 | |
| 45 | |
| 46 | # # summarize 3*3 or 5*3 data into one by calculating the mean score |
| 47 | # summary_data = data_trim.groupby(['compound0_x','compound0_y','DepMap_ID_x']).agg({\ |
| 48 | # "synergy_loewe":'mean',"study_name":'count'}).reset_index().rename(columns={\ |
| 49 | # 'synergy_loewe':'MEAN_SCORE','study_name':'count'}).astype({'compound0_x':'int32','compound0_y':'int32'}) |
| 50 | |
| 51 | # # some experiments may fail and get NA values, drop these experiments |
| 52 | summary_data = data_trim.dropna() |
| 53 | |
| 54 | summary_data = summary_data[['compound0_x','compound0_y','DepMap_ID','tissue_name','synergy_loewe','ri_row', 'ri_col']].rename(columns={\ |
| 55 | 'compound0_x':'drug1','compound0_y':'drug2','DepMap_ID':'cell','synergy_loewe':'score',\ |
| 56 | 'ri_row':'ic_1', 'ri_col':'ic_2'}) |
| 57 | |
| 58 | # proessed_dpi = pd.read_csv(os.path.join(ROOT_DIR, 'results','proessed_dpi_extended.csv'),index_col=0) |
| 59 | # proessed_dpi.columns = proessed_dpi.columns.astype(int) |
| 60 | # library = set(proessed_dpi.columns) |
| 61 | # summary_data = summary_data[summary_data['drug1'].isin(library)] |
| 62 | # summary_data = summary_data[summary_data['drug2'].isin(library)] |
| 63 | |
| 64 | return summary_data.head(10000) |
| 65 | |
| 66 | def process_sanger2022(): |
| 67 | data = pd.read_csv(os.path.join(ROOT_DIR, 'data', 'synergy_data','Sanger2022','Jaaks2022_trueset_xiabo.csv')) |
no test coverage detected