New computer-based modeling and data mining tools are continuously being developed to help assess and exploit targets uncovered in genomic research. Traditional computer-aided drug design software has several serious shortcomings. Cross-vendor incompatibilities induce users to use single software exclusively. The users ability of Mix & Match software, unattainable with any current offering, has been shown to enhance in silico hits discovery and enrichments and to lead to increased hit rates. Additionally, low throughput of compute-intensive jobs in commercial seriously limits Docking-scoring-hit identification due to its inability to partition and streamline jobs execution across a cluster of compute servers. These limitations inspired TransTech Pharma to introduce innovations to circumvent them. The proposed system features a graphical user interface, Mixing & Matching cross-vendor software, uses a divide-and-conquer strategy to increase throughput across a multi-unit/multi operating system distributed processing architecture to parallelize job execution across a cluster of compute servers. This parallelization would maximize throughput of complex and compute-intensive computational chemistry tasks. The system will be validated using structure-activity information for three cyclin dependent kinase inhibitors PfPK5, PfPK6, and Pfmrk as targets. Initial work showed that Pfmrk has 40% sequence identity to MAP-Kinase P38 whose crystal structure would serve as threading & homology model template for Pfmrk.
Benefits: In an era where efficiency and timeliness are integral part for the success of any organization, the system for performing integrated structure-based and ligand-based drug design is appealing, as the approach would significantly improve the overall process efficiency. The parallelization of several tasks using the underutilized computing resources in a pharmaceutical setting avoids high-powered hardware purchases. This feature would attract the companies as it results in significant cost savings. The scientific merits, efficiency, and the financial advantages of the discovery engine would be projected as key for commercializing the product. Initial discussions have taken place with representatives of IBM who have expressed an interest in forming a three party consortium whereby TransTech (the company) would develop and support the software, IBM would provide the hardware, and IBM would coordinate with one of its existing health care software distributors to market and sell a turn key solution.
Keywords: computer-aided, drug design, bioinformatics, homology, threading