Quantum mechanics, or quantum physics, describes nature at the fundamental level of atomic nuclei and electrons. Quantum mechanics is the most accurate way to simulate the behavior of molecules and interactions between molecules and allows us to predict with remarkable accuracy the complex biomolecular interactions critical to biological systems. We deploy quantum physics methods to tackle the challenges in protein targets previously thought to be “undruggable,” yielding the most accurate predictions and facilitate our mission to deliver new safe and effective medicines to patients.
Building on Ground Truth
By building on the foundation of quantum mechanics, we are secure in knowing we are operating based on ground truth for how biological systems work at the molecular level. This approach allows Roivant Discovery to move beyond static representations of biological molecules like proteins and ligands. At 37 ℃, the temperature of human biology, atomic interactions are driven by a balance of entropy and enthalpy, captured by a thermodynamic quantity known as the Gibbs free energy, which can be computed using molecular dynamics and statistical thermodynamics.
In the complex world of human biology, the Gibbs free energy is the ultimate arbiter of conformational states and defines the energetic landscape that makes biology possible. By accurately predicting the Gibbs free energy through the lens of molecular dynamics and statistical thermodynamics, scientists at Roivant Discovery can assess the full role of protein motion and dynamic interactions between biomolecules to overcome critical bottlenecks in the drug discovery process.
Understanding How Biology Works, Atom-by-Atom
The complexity of protein-sized molecules prevents full quantum mechanical simulations from being run directly on biologically relevant system. As such, we have taken an industry-leading approach where we apply the appropriate level of theory that strikes a balance between throughput and accuracy. We employ:
- Quantum physics to understand the properties of potential drug molecules.
- Molecular dynamics and statistical thermodynamics to predict how potential drug molecules bind to and modulate therapeutic protein targets.
- High-performance computing and software engineering to make accurate predictions at scale.
These result in a clearer understanding of the biological system and the relationship to disease at an atomic level.
Better Medicines through Physics and Predictive Sciences
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.
Traditional computational approaches to drug design treat proteins as rigid, fixed molecules, but in living organisms proteins are in fact highly dynamic and shape shifting molecules. This new paradigm of drug design is founded on the principles of quantum physics, but not physics in a vacuum. We perform hundreds of quantum mechanical calculations on each ligand to parameterize 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, ions, etc. It’s the predictive power generated by our physics-based simulations and machine learning coupled with our in-house experimental efforts in chemistry and biology that propels our advances toward better medicines through physics.