Description:
Regression model to predict inhibitors of cyclooxygenase-2 (COX2). 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 188 training compounds and 94 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 = 0.6910, MAE = 0.4473, and RMSE = 0.5659.

External performance:
Squared R = 0.5640, MAE = 0.7987, and RMSE = 1.0451.

Regression equation:
pLC50 =

    -82.3504 * AC[5]_N2_F_AB_nCi_2_SS13_n_A_LGP[8]_v_MAS +
     -0.1209 * GV[1]_S_B_BB_nCi_2_SS15_H_D_LGP[4]_ku-dc4_MAS +
    287.3077 * GV[3]_N2_Q_AB_nCi_2_SS14_n_G_LGP[3-5]_a_MAS +
   -235.0962 * GV[7]_Q3_B_AB_nCi_2_MP11_A_LGP[3-4]_alk-hx_MAS +
  15304.0429 * IB_RA_F_AB_nCi_2_NS10_X_LGP[2-8]_p_MAS +
     -0.0485 * P3_F_AB_nCi_2_NS2_X_LGP[1]_v_MAS +
      0.1425 * TS[2]_K_B_AB_nCi_2_NS10_n_P_LGP[5]_r-c_MAS +
    -69.9184 * TS[3]_N1_Q_BB_nCi_2_MP7_X_LGP[4-6]_a_MAS +
 -17791.9006 * TS[7]_P2_B_BB_nCi_2_MP15_C_LGP[8]_e-c_MAS +
      7.7051