Gerald Lushington presentation on chemical informatics as applied to discovering biologically relevant chemical diversity metrics from HTS data. The presentation outlines simple criteria for identifying molecular properties that tend to discriminate between compounds that are pharmacologically active toward specific phenotypes vs. those that have no effect. The implication of the work is that one can identify simple parameters that can be used to rapidly screen large compound collections for subset of chemical that have an elevated chance of being relevant in the treatment of specific diseases.