Tudor Oprea, Cristian Bologa, Marius Olah
Office of Biocomputing, University of New Mexico School of Medicine
John Tallarico, Erik Brauner
Institute of Chemistry and Cell Biology, Harvard University Medical School
The Top 200 drugs (158 unique structures) were combined with 769 medicinal chemistry structures from WOMBAT (by selecting the highest active compounds for each target) to define a high-activity chemical space. These 923 compounds were clustered according to the Molecular Design Limited (MDL) "drug like" keys using Mesa Analytics & Computing software. The 1-D representation of biological activity space can be used to visualize novel compounds (single-SMILES similarity mode in a web interface) using SimNav, the Similarity Navigator. We will present our experience with SimNav, an on-line product from the University of New Mexico and Harvard Medical School.
Acknowledgments: John and Norah MacCuish (Mesa Analytics & Computing) and Andrew Dalke (Dalke Scientific Software) have contributed in this project.