Description:
Regression model to predict inhibitors of thrombin (THR). The model was built with the Multiple Linear Regression technique by using a total of 9 QuBiLS-MAS descriptors.
Training and testing datasets:
A total of 59 training compounds and 29 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 = 0.8483, MAE = 0.3114, and RMSE = 0.3736.
External performance:
Squared R = 0.8493, MAE = 0.3125, and RMSE = 0.4352.
Regression equation:
pKi =
-1.7565 * AC[4]_I50_Q_AB_nCi_2_MP15_H_n_D_NSRW_ec_MAS + -1.9806 * TS[7]_MX_Q_BB_nCi_2_MP4_H_T_LGP[3-4]_dc4_MAS + 0.042 * TS[7]_I50_Q_AB_nCi_2_NS4_n_D_LGP[3-7]_dc3_MAS + -14412.2052 * TS[6]_V_Q_BB_nCi_2_NS4_D_LGP[2-3]_c_MAS + 0.0744 * TS[2]_MX_Q_AB_nCi_2_SS2_H_n_M_LGP[2-7]_est_MAS + -0.8592 * AC[6]_S_Q_AB_nCi_2_MP5_H_T_LGP[5-6]_alk_MAS + -0.0012 * AC[7]_VC_B_AB_nCi_2_NS15_n_X_LGP[2]_c-ku_MAS + 0.0664 * GV[5]_K_F_BB_nCi_2_MP12_H_T_LGP[2-7]_est_MAS + -0.0845 * AC[5]_S_Q_BB_nCi_2_MP7_H_T_LGP[3]_est_MAS + 7.7484