QUAISAR Platform

Traditional computational approaches to drug design treat proteins as rigid, fixed molecules, but in living organisms proteins are highly dynamic and shape-shifting molecules. Understanding these motions, and accurately predicting thermodynamic properties of interest, allows us to systematically engineer molecules with desirable properties in an atom-by-atom fashion. We perform hundreds of quantum mechanical calculations on each ligand to develop parameters for a highly accurate molecular force field that is used to run long-timescale molecular dynamics simulations of biologically relevant systems, including proteins, ligand, cofactors, membranes and ions, etc. When available, we use data to build predictive machine learning models that augment the physics-based predictions. The predictive power generated by our physics-based simulations and machine learning, coupled with our in-house experimental capabilities, gives Roivant Discovery a unique advantage in rapidly progressing molecules to the clinic. With the advancement of high-performance computers, we are able to run molecular simulations at the scale and precision necessary to impact drug discovery decisions. Our GPU-accelerated approach allows us to accurately assess thousands of design ideas on a daily basis.

QUAISAR is a flexible, scalable platform that allows our researchers to make critical drug discovery decisions, from target ID through IND nomination. We have the capability to screen billions of compounds virtually and synthesize only the most promising drug candidates by combining generative modeling, reaction-based enumeration, and accurate prediction of drug-like properties, including binding affinity, selectivity, and ADMET properties. Additionally, the simulations capabilities in QUAISAR facilitate the exploration of the conformational landscape of protein and potential drugs with great specificity, providing insights into the disease nature of proteins and a streamlined path toward therapeutic molecules. QUAISAR, in combination with genetic, biological, and biophysical data, provides a detailed understanding of the relationship between the conformational states of the protein target and disease, allowing us to focus our drug design efforts on molecules that modify disease-relevant states. Combined with experienced drug hunters, we can overcome challenges in target ID, hit finding, hit-to-lead, and lead optimization.

To understand biological systems and predict the impact of potential drug molecules, our scientists run accurate all-atom simulations at biologically meaningful timescales. Accurate free energy methods, a customized force field based on quantum mechanics, advanced molecular dynamics simulations, and a unique design interface, coupled with expansive HPC resources and an agile software engineering team, are a few of the key QUAISAR advantages. QUAISAR also produces intuitive visualizations and data analyses that provide our veteran drug hunters with actionable information to guide the next steps of drug design.

Integration of the QUAISAR platform with our world-class team of drug discovery scientists facilitates the rapid discovery of next generation medicines for the most challenging diseases.