The Application of Pattern Recognition to Screening Prospective Anti-Cancer Drugs: Adenocarcinoma 775Biological Activity Test.

摘要

A novel application of pattern recognition to the screening of potential anti-cancer drugs is presented. Structural features of 200drugs previously tested by the National Cancer Institute for activity in the solid tumor adenocarcinoma 755screening test are input to a master program of pattern recognition methods. The programs were 93.5% accurate in discriminating drugs with positive anti-neoplastic activity versus those with no anti-cancer activity. Extensions to a more rational approach to 'drug design'are also discussed. (Author)