Computation of the molecular shapes' similarity and diversity using USR method and General Shape Index

摘要

Regarding the molecular shape, shapes' similarity and shapes' diversity the paper presents (1) a new molecular descriptor, (2) using of the new descriptor together with the previous Ultrafast Shape Recognition (USR) formula, (3) a quantitative method to verify the observance of the 'QSAR axiom', (4) a formula for identification of the activity and shape 'cliffs', (5) a method to divide in classes any group of molecules, (6) a criterion to identify the 'atypical' molecules and (7) a method, based on Shannon entropy formula, for computation of the molecules' diversity and the similarity of two groups of molecules. The proposed formulas/procedures are simple and suggestive. The algorithm which uses the proposed descriptor and USR formula describes correctly enough the molecular similarity in three analyzed groups of molecules.