Table 2: QSPR models of pKa for N-base ligands.

ModelsDescriptors involved t-testStatistical parameters
Training set (N = 17)Test set (N = 5)
RMSE

Model I pKa Intercept−68.69−12.730.950.8961.790.3110.970.51
HNMVF0.021613.60
FNSA3244.0210.22
FPSA113.966.33
MNRI-C−89.88−3.92

Model II pKa Intercept16.2316.430.940.8956.620.330.851.11
MERI-C−328.98−10.45
HDSA1/TMSA26.9411.58
WPSA2−0.0163−6.51
Q max−48.37−6.38

Model III pKa Intercept8.5510.300.760.6123.371.320.910.93
HDCA1/TMSA73.896.43
FNSA3312.256.29

aC: indicates the coefficients of the descriptors involved in the models, : the square of the regression correlation coefficient, : the cross-validated square of the regression correlation coefficient, : the -value for the regression, : the standard deviation of the regression, : the predicted square of the regression correlation coefficient, RMSE: the root mean square error.