High-performance computing for drug design

High performance computing has become a major focus of attention by government, industry, medical centers and academic institutions. The U.S. government has made this a “top national priority”, linking the development of a “data superhighway system” to national competitiveness and national research interest. High performance supercomputing approaches have been used for sequence analysis, gene finding, protein structural prediction, all-atom simulation such as molecular dynamics and quantum calculations, modeling biological networks such as systems biology and more recently drug design and drug discovery. All these approaches are highly computationally demanding, in terms of compute load, communication speed, and memory load. Supercomputing based drug design and drug discovery use high-performance super-computers and bioinformatics approaches to discover, enhance, and study drugs and related biologically active molecules as well as the sites of protein interactions. Methods include molecular modeling using biophysical approaches such as molecular dynamics, semi-empirical quantum mechanics methods, ab initio quantum chemistry methods, density functional theory, receptor - ligand interactions and protein docking and so on. The success of the high-throughput drug design and drug discovery now directly relies on the high-performance computing capabilities. Many research and computational products that were used to be considered impossible now proved to be feasible and effective with the help of today’s supercomputing techniques. In particular, the identification of diseases relating to protein structural changes challenges biomedicine as the result of the sophisticated protein interaction networks that demand effective drug design using supercomputing based on mathematical, computational and biophysical models and algorithms for solving the model equations, and the bioinformatics techniques to analyze and validate the results. Those will need our deeper studies of biophysical phenomena and interesting biophysical and algorithmic problems using supercomputing. In this keynote lecture, we follow the scenario of Koshland’s “induced-fit” to demonstrate that the identification of intrinsically disordered/unstructured proteins will became increasingly important to the drug design and discovery, because many proteins are folding upon interaction with drugs. The High-performance computing approaches help to reduce the number of targets for a good drug that has to be eventually tested by expensive and time-consuming synthesis and laboratory and toxicology experiments. High-performance computing study of intrinsically unstructured protein related pathways focuses on the discovery and development of novel drugs for the treatment of metabolic and mis-folding related diseases by targeting the metabolic and biological pathways in the protein interaction network. The advances of supercomputing will foster the synergistic relation between biophysical and computational sciences towards translational bioinformatics in the future.