DISCRIMINATION TECHNIQUES APPLIED TO THE NCI IN-VITRO ANTITUMOR DRUG SCREEN - PREDICTING BIOCHEMICAL-MECHANISM OF ACTION

The National Cancer Institute currently tests approximately 400 compou nds per week against a panel of human tumour cell lines in order to id entify potential anti-cancer drugs. We describe several approaches, ba sed on these in vitro data, to the problem of identifying the primary biochemical mechanism of action of a compound. Using linear and non-pa rametric discriminant procedures and cross-validation, we find that ac curate identification of the mechanism of action is achieved for appro ximately 90 per cent of a diverse collection of 141 known compounds, r epresenting six different mechanistic categories. We demonstrate that two-dimensional graphical displays of the compounds in terms of the in itial three principal components (of the original data) result in sugg estive visual clustering according to mechanism of action. Finally, we compare the classification accuracy of the statistical discrimination procedures with the accuracy obtained from a neural network approach and, for our example, we find that the results obtained from the vario us approaches are similar.