| import pandas as pd
|
| import numpy as np
|
| from sklearn.metrics import mean_squared_error, mean_absolute_error
|
| from sklearn.metrics import r2_score
|
| from scipy.stats import pearsonr
|
|
|
|
|
| res = np.array(pd.read_excel('1_all_samples_metrics.xlsx', sheet_name='Sheet1')).T
|
| sequence = res[1]
|
| smiles = res[2]
|
| Type = res[6]
|
| Value = res[7]
|
| Predict_Label = res[8]
|
| Training_test = res[9]
|
| print(sequence[0], smiles[0], Type[0], Value[0], Predict_Label[0], Training_test[0])
|
|
|
|
|
| def Whole_dataset():
|
|
|
| Pcc = pearsonr(Value, Predict_Label)[0]
|
| RMSE = np.sqrt(mean_squared_error(Value, Predict_Label))
|
| MAE = mean_absolute_error(Value, Predict_Label)
|
| r2 = r2_score(Value, Predict_Label)
|
| print('***The whole set***')
|
| print('Pcc:', Pcc, 'RMSE:', RMSE, 'MAE:', MAE, 'r2:', r2)
|
|
|
|
|
| def test_dataset():
|
|
|
| Value_test = []
|
| Predict_Label_test = []
|
| for i in range(len(Training_test)):
|
| if Training_test[i] == 1:
|
| Value_test.append(Value[i])
|
| Predict_Label_test.append(Predict_Label[i])
|
| Value_test = np.array(Value_test)
|
| Predict_Label_test = np.array(Predict_Label_test)
|
| Pcc_test = pearsonr(Value_test, Predict_Label_test)[0]
|
| RMSE_test = np.sqrt(mean_squared_error(Value_test, Predict_Label_test))
|
| MAE_test = mean_absolute_error(Value_test, Predict_Label_test)
|
| r2_test = r2_score(Value_test, Predict_Label_test)
|
| print('***Test set***')
|
| print('Pcc:', Pcc_test, 'RMSE:', RMSE_test, 'MAE:', MAE_test, 'r2:', r2_test)
|
|
|
|
|
| def Wildtype_all_dataset():
|
|
|
| Value_wildtype = []
|
| Predict_Label_wildtype = []
|
| for i in range(len(Type)):
|
| if Type[i] == 'wildtype':
|
| Value_wildtype.append(Value[i])
|
| Predict_Label_wildtype.append(Predict_Label[i])
|
| Value_wildtype = np.array(Value_wildtype)
|
| Predict_Label_wildtype = np.array(Predict_Label_wildtype)
|
| Pcc_test = pearsonr(Value_wildtype, Predict_Label_wildtype)[0]
|
| RMSE_test = np.sqrt(mean_squared_error(Value_wildtype, Predict_Label_wildtype))
|
| MAE_test = mean_absolute_error(Value_wildtype, Predict_Label_wildtype)
|
| r2_test = r2_score(Value_wildtype, Predict_Label_wildtype)
|
| print('***The whole wildtype set***')
|
| print('Pcc:', Pcc_test, 'RMSE:', RMSE_test, 'MAE:', MAE_test, 'r2:', r2_test)
|
|
|
|
|
| def Wildtype_test_dataset():
|
|
|
| Value_wildtype = []
|
| Predict_Label_wildtype = []
|
| for i in range(len(Type)):
|
| if Type[i] == 'wildtype' and Training_test[i] == 1:
|
| Value_wildtype.append(Value[i])
|
| Predict_Label_wildtype.append(Predict_Label[i])
|
| Value_wildtype = np.array(Value_wildtype)
|
| Predict_Label_wildtype = np.array(Predict_Label_wildtype)
|
| Pcc_test = pearsonr(Value_wildtype, Predict_Label_wildtype)[0]
|
| RMSE_test = np.sqrt(mean_squared_error(Value_wildtype, Predict_Label_wildtype))
|
| MAE_test = mean_absolute_error(Value_wildtype, Predict_Label_wildtype)
|
| r2_test = r2_score(Value_wildtype, Predict_Label_wildtype)
|
| print('***The test wildtype set***')
|
| print('Pcc:', Pcc_test, 'RMSE:', RMSE_test, 'MAE:', MAE_test, 'r2:', r2_test)
|
|
|
|
|
| def Mutant_all_dataset():
|
|
|
| Value_wildtype = []
|
| Predict_Label_wildtype = []
|
| for i in range(len(Type)):
|
| if Type[i] != 'wildtype':
|
| Value_wildtype.append(Value[i])
|
| Predict_Label_wildtype.append(Predict_Label[i])
|
| Value_wildtype = np.array(Value_wildtype)
|
| Predict_Label_wildtype = np.array(Predict_Label_wildtype)
|
| Pcc_test = pearsonr(Value_wildtype, Predict_Label_wildtype)[0]
|
| RMSE_test = np.sqrt(mean_squared_error(Value_wildtype, Predict_Label_wildtype))
|
| MAE_test = mean_absolute_error(Value_wildtype, Predict_Label_wildtype)
|
| r2_test = r2_score(Value_wildtype, Predict_Label_wildtype)
|
| print('***The whole mutant set***')
|
| print('Pcc:', Pcc_test, 'RMSE:', RMSE_test, 'MAE:', MAE_test, 'r2:', r2_test)
|
|
|
|
|
| def Mutant_test_dataset():
|
|
|
| Value_wildtype = []
|
| Predict_Label_wildtype = []
|
| for i in range(len(Type)):
|
| if Type[i] != 'wildtype' and Training_test[i] == 1:
|
| Value_wildtype.append(Value[i])
|
| Predict_Label_wildtype.append(Predict_Label[i])
|
| Value_wildtype = np.array(Value_wildtype)
|
| Predict_Label_wildtype = np.array(Predict_Label_wildtype)
|
| Pcc_test = pearsonr(Value_wildtype, Predict_Label_wildtype)[0]
|
| RMSE_test = np.sqrt(mean_squared_error(Value_wildtype, Predict_Label_wildtype))
|
| MAE_test = mean_absolute_error(Value_wildtype, Predict_Label_wildtype)
|
| r2_test = r2_score(Value_wildtype, Predict_Label_wildtype)
|
| print('***The test mutant set***')
|
| print('Pcc:', Pcc_test, 'RMSE:', RMSE_test, 'MAE:', MAE_test, 'r2:', r2_test)
|
|
|
|
|
| def New_substrate_enzyme_dataset():
|
|
|
| Trainingset_seq_smiles = []
|
| for i in range(len(Training_test)):
|
| if Training_test[i] == 0:
|
| Trainingset_seq_smiles.append(sequence[i])
|
| Trainingset_seq_smiles.append(smiles[i])
|
| Value_test = []
|
| Predict_Label_test = []
|
| for i in range(len(Training_test)):
|
| if Training_test[i] == 1 and (sequence[i] not in Trainingset_seq_smiles or smiles[i] not in Trainingset_seq_smiles):
|
| Value_test.append(Value[i])
|
| Predict_Label_test.append(Predict_Label[i])
|
| Value_test = np.array(Value_test)
|
| Predict_Label_test = np.array(Predict_Label_test)
|
| Pcc_test = pearsonr(Value_test, Predict_Label_test)[0]
|
| RMSE_test = np.sqrt(mean_squared_error(Value_test, Predict_Label_test))
|
| MAE_test = mean_absolute_error(Value_test, Predict_Label_test)
|
| r2_test = r2_score(Value_test, Predict_Label_test)
|
| print('***The Test new_substrate_enzyme dataset***')
|
| print('Pcc:', Pcc_test, 'RMSE:', RMSE_test, 'MAE:', MAE_test, 'r2:', r2_test)
|
|
|
|
|
| if __name__ == '__main__':
|
|
|
|
|
|
|
|
|
|
|
|
|
| New_substrate_enzyme_dataset()
|
|
|
