Computer Simulations Decode Biological Phenomenon and Drug Prediction

Understanding the underlying mechanisms that govern the biological process requires scrutiny at spatial and temporal resolutions that can be challenging for current experimental techniques. Over the past three decades, molecular dynamics (MD) simulation has evolved to become an indispensable tool for studying biological phenomena with atomic precision and at relevant timescales. Such simulations have served as a “computational microscope”, providing information previously unattainable by experiments. Today, MD simulations are having a profound impact on the treatment of diseases and the development of new, potent drugs. For example, researchers have utilized MD simulations to capture the assembly of a whole virus inside host cells and tried to explain how virus cause diseases. Insight provided by these findings could be particularly informative for developing new drugs. Currently, an experimental drug entering Phase I clinical trial has only a 10% chance of reaching the market, and lack of clinical efficacy has become the most frequent cause. A deeper understanding of the fundamental mechanisms that underlie disease pathologies will enhance the success rate. To this effect, MD simulations have and will continue to make vital contributions.

In general, all biochemical processes in MD are governed by physical equations termed force field (FF). The FF mimic the forces acting on a molecule, and the accuracy of such forces governs the predictive capability of all MD simulations. As such, iteratively developing more precise FFs is vital to ensure that the theoretically derived models map onto physical reality. My current research is focused on improving the predictive power of FF by performing quantum calculations and numerical optimizations using state-of-the-art supercomputers. Specifically, I proposed a robust modelling framework that is capable of providing more rigorous description of the field of forces, followed by extensive testing against experimental benchmarks. My research has revealed that this ‘new’ FF can provide more reliable description of reality which is indeed a prerequisite for any MD simulation. Moreover, the new FF is superior to our lab’s prior FF that is the standard in the field with over 4000 citations. The application of more accurate model will greatly enhance our understanding of biological processes and aid in our ability to control, manipulate and arrest such processes, which will profoundly contribute to new drug development.