On the Equivalence Between Different Descriptions of Molecules: Value for Computational Approaches
Romualdo Benigni, Grazia Gallo, Fabrizio Giorgi, Alessandro Giuliani · Journal of Chemical Information and Computer Sciences · 1999
The correlation between two different sets of chemical descriptors for the same heterogeneous set of molecules was investigated at both local and global levels. The two descriptions were well correlated at the global level, whereas they showed different characteristics on the detailed scale. This result has important consequences for both combinatorial chemistry strategies and QSAR analyses of congeneric series. It implies that the “maximal diversity” optimization programs are largely independent from the type of chemical descriptors only in large ensembles of molecules, but not in individual series. On the other hand, the general isomorphism between different chemical descriptions may give rise, for QSAR analyses of congeneric series, to a multiplicity of formally equivalent solutions for the same QSAR problem.