Description:
Regression model to predict inhibitors of thermolysin (THERM). The model was built with the Multiple Linear Regression technique by using a total of 7 QuBiLS-MAS descriptors.
Training and testing datasets:
A total of 51 training compounds and 25 testing compounds were extracted from the Sutherland, et al., 10.1021/jm0497141
Internal performance:
For a 10-fold cross-validation repeated 100 times: Squared R = 0.7458, MAE = 0.7289, and RMSE = 0.9513.
External performance:
Squared R = 0.7925, MAE = 0.7387, and RMSE = 1.0242.
Regression equation:
pKi =
-230.7776 * AC[2]_MN_F_BB_nCi_2_SS5_A_LGP[4-8]_hx_MAS + -4.7745 * GV[4]_I50_F_AB_nCi_2_NS6_M_LGP[1-8]_c_MAS + 0.2161 * K_Q_BB_nCi_2_SS13_A_LGP[2]_c_MAS + 0.0297 * GV[5]_VC_B_BB_nCi_2_MP10_H_X_LGP[5-6]_alk-ku_MAS + 0.3356 * AC[5]_S_B_AB_nCi_2_SS15_n_C_LGP[6-7]_ku-dc2_MAS + 0.0844 * AC[1]_K_Q_AB_nCi_2_SS15_C_LGP[2-5]_c_MAS + 0.2086 * TS[2]_N2_B_AB_nCi_2_SS5_H_C_LGP[3-6]_psa-est_MAS + -2.444