Facet diagrams for quantum similarity data.

The objective of this work is to demonstrate that an appropriate treatment of quantum similarity matrices can reveal hidden data grouping related to relevant structural features and even to biological properties of interest. Classical scaling is used here to extract the information contained in the similarity relationships between the elements of a molecular set. Facet theory is invoked to relate, in a qualitative way, the spatial regions to structural characteristics as well as to properties of interest. Two application examples are discussed: the Cramer steroid set and a benzene, toluene and xylene derivatives set.