Description:
Regression model to predict inhibitors of the benzodiazepine receptor (BZR). 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 98 training compounds and 48 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 = 0.6915, MAE = 0.2952, and RMSE = 0.3733.
External performance:
Squared R = 0.6598, MAE = 0.4679, and RMSE = 0.6353.
Regression equation:
pLC50 =
-0.5247 * AC[1]_VC_B_AB_nCi_2_MP12_n_X_LGP[5-6]_p-ec_MAS + -0.0424 * AC[1]_VC_F_BB_nCi_2_NS12_H_P_LGP[5-6]_c_MAS + -0.0494 * AC[3]_AM_F_AB_nCi_2_NS9_H_n_D_LGP[7]_c_MAS + 0.4086 * AC[3]_N1_Q_BB_nCi_2_SS10_A_LGP[1-2]_est_MAS + 73.6293 * AC[5]_Q2_B_AB_nCi_2_MP4_H_n_P_LGP[2-4]_ku-dc2_MAS + 0.3251 * AC[6]_MX_Q_BB_nCi_2_MP11_P_LGP[3-4]_alk_MAS + -0.4995 * ES_S_F_BB_nCi_2_SS14_P_LGP[6-8]_ec_MAS + 0.017 * GV[7]_VC_F_AB_nCi_2_NS15_H_n_G_LGP[5]_est_MAS + -0.3415 * TS[6]_S_Q_AB_nCi_2_NS11_n_X_LGP[1-8]_p_MAS + 8.0994