| We describe a novel de novo design algorithm to build compounds that are combinatorially accessible. The method called the Combinatorial Assembly Method (or CASM), uses compounds and reactions desired by the chemist to be used to construct new and novel inhibitors. This method uses aspects of Pitman's FASM (Fragment Assembly Method), to ensure that invented compounds are able to fit into the active site pocket and function as good inhibitors.; We then apply CASM to the invention of candidates to Checkpoint Kinase-1 (Chk1). Chk1 is a pharmaceutically interesting target as suppression of it, with siRNA, has shown to hold cells in the G1 phase. CASM constructed candidates from two sets of reactions and reactants, the 1rxnlib and the 14rxnlib. The 1rxnlib, virtual library was also completely enumerated by PIPELINE PILOT and the candidates invented by CASM were compared to the high scoring candidates from the virtual library. The candidates from the virtual library are evaluated with ROCS, FRED, and GLIDE. We have found that the candidates invented by CASM are enriched with candidates that score as well as micromolar strength inhibitors over random picking from the whole potential combinatorial library. CASM also invents candidates in poses that are similar to both their docked pose and their minimized pose. The second virtual library, the 14rxnlib, used fourteen reactions and covered all the combinatorial chemistry of interest on this project and included products of the 1rxnlib as subset members.; Finally, we use OMEGA, FRED, and GLIDE to propose candidates for screening against Senp2. Senp2 has been shown to be a target of interest in cancer. We evaluated two flavors of candidates: a non-covalent, and a covalent set. The non-covalent inhibitors are from a commercially available library. Seventy-three candidates have been recommended, of which 40 have been screened. Thirty of the screened candidates have shown activity against Senp2, equating to a 75% hit rate. |
