Structural Classification of Complex Molecules by Artificial Intelligence Techniques

Abstract

Algorithms for classification and taxonomy bases on criteria, e.g., information entropy. The feasibility of replacing a given molecule by similar ones in the composition of a complex drug is studied. Some local anaesthetics currently in use are classified using structural properties. In taxonomy the detailed comparison of the sequences of biomolecules, proteins or nucleic acids, allows the reconstruction of a molecular phylogenetic tree. The method is applied to the classifications of (1) indazolols (against Trichomonas vaginalis), (2) fullerenes and fullerite, (3) living and heat-inactivated lactic acid bacteria against cytokines, (4) phylogenesis of avian birds and 1918 influenza virus, (5) local anaesthetics, (6) transdermal-delivery percutaneous enhancers, (7) quantitative structure–activity relationship of anti-human immunodeficiency virus (HIV) compounds, (8) HIV inhibitors, e.g., thiocarbamates, N-aryloxazolidinone-5-carboxamides and styrylquinolines, (9) antimalarial aryltriazolylhydroxamates, (10) N-aryl-N-(3-aryl-1,2,4-oxadiazol-5-yl) amines against prostate cancer, antimitotic 2-phenylindole-3-carbaldehydes against breast cancer and anti-tubulin agents against gastric cancer with indole ring. The entropy contributions may be studied with the equipartition conjecture. It is not within the scope of our simulation method to replace biological tests of drugs or field data in palaeontology, but such simulation methods can be useful to assert priorities in detailed experimental research. Available experimental and field data should be examined by different classification algorithms to reveal possible features of real biological significance.