계산생물학 연구실

The ultimate goal of our research is to understand and to predict the functions of biomolecules in atomic details with theoretical and computational methods. Such research can also provide basic tools for the design of small molecules or proteins that can contribute to promoting human health. For such purposes, we have been developing a biomolecular modeling program package GALAXY and a web server GalaxyWeb implementing the GALAXY methods1 (click here to connect to the web server). GALAXY is a software package for biomolecular modeling which consists of core programs that perform basic modeling tasks such as protein structure prediction, protein-ligand docking, and protein-protein docking and applied programs aimed at more specific goals such as ligand binding site prediction or protein homo-oligomeric state prediction. The GALAXY program package has been designed to be modular, enabling convenient addition of new programs in the future. The philosophy underlying the GALAXY development is to provide more “practically useful” information to experimentalists, such as more “reliable” model structures, by fusion of available tools and knowledge in diverse disciplines such as chemistry, physics, computer science, or bioinformatics. GALAXY currently utilizes various scoring functions (e.g. molecular mechanics, implicit solvation models, statistical scores) and sampling techniques (e.g. molecular dynamics, genetic algorithm, Monte Carlo simulation, conformational space annealing) to achieve the best performance in each modeling task. One of the merits of GalaxyTBM is that it automatically refines unreliable model regions in relatively short computation time. Unreliable loops, termini, or side-chains are automatically detected and reconstructed by Galaxy modules. The generated