抄録
A simple method of virtual screening is proposed. This method uses only chemical characters calculated from two dimensional chemical structures. Local and global chemical characters are represented by molecular fingerprint and trait, respectively. The trait is a newly introduced concept in this paper and it is expressed by a set of two dimensional (2D) molecular descriptors. In this study, both the molecular fingerprint and the trait were used to represent drug-likeness of a group of molecules with a particular pharmacological activity. To learn about the molecular fingerprint and trait specific to a particular group of drugs, we used a database of drugs that are clinically used in Japan now. The molecular fingerprint and trait trained on these real drugs were used to predict drug-likeness of molecules in other chemical databases. In these chemical databases, an appreciable number of compounds that show the relevant pharmacological activity are contained. Some of these compounds are drugs clinically used abroad, but not in Japan. The prediction rate was judged by an enrichment factor. Despite the simplicity of the methodology, practical results were obtained. In the case of β-adrenergic blockers, the enrichment factor of 66 was attained and nearly 57% of active molecules in the chemical databases were successfully covered.
