摘要
Biological systems are composed of and regulated by intricate and diverse biomolecular interactions. Experimental and computational approaches have been developed to elucidate the mechanisms of these interactions; however, owing to cost, time, and accuracy issues, large-scale spatiotemporal analyses of molecular pairs remain challenging. Thus, the molecular recognition mechanisms underlying these diverse interactions remain unclear. We successfully simulated the large-scale molecular dynamics (MD) of 4,275 protein-compound pairs by combining a method to accelerate the MD simulations with the supercomputer Fugaku. Our spatiotemporal analysis of generated big MD data revealed universal features underlying molecular recognition and binding processes. This study expands our understanding of the concept of MD simulations from a technique to investigate the dynamic properties of individual protein-drug pairs to an approach to perform large-scale spatiotemporal analysis and compound screening. This study opens an avenue in biological research for subsequent drug discovery.