A reliable and accurate method for the computational design of novel drug candidates has been a passionate pursuit of the pharmaceutical industry. Such technology would dramatically improve the efficiency of drug discovery by quickly and inexpensively providing potent molecules that can be further prioritized for synthesis based on characteristics such as patentability, specific protein-ligand interactions, ease of chemical synthesis, protein selectivity and pharmacological considerations. Described herein is the progress made at Locus Pharmaceuticals Inc toward achieving this ideal with a fragment-driven, computationally directed approach to small-molecule discovery. Specific lead identification examples from Locus Pharmaceuticals discovery programs demonstrate the efficiency and cost-effectiveness realized by such an approach.