Constructing a linear QSAR for some metabolizable drugs by human or pig flavin-containing monooxygenases using some molecular features selected by a genetic algorithm trained SVM

A linear QSAR is constructed for predicting Km or whether a drug can be metabolized by human flavin-containing monooxygenases (FMOs) or not. The predicted Km (mol/L) is plotted against the measured Km for the human FMO training data set. The correlation coefficient R2 for fitting these data points to a straight line and the fitted equation for the straight line are also shown on the plot. The prediction accuracy given by the linear QSAR model is about 80% for some drugs randomly selected from the literatures.