Conference: Computer Aided Drug Design Gordon Research Conference

Title: Accurate binding free energy calculations with MovableType and density-driven refinement coupled with enhanced sampling

Abstract: 

In this poster, we will discuss how MovableType (MT) and density-driven structural refinement enhance the accuracy and efficiency of absolute binding free energy calculations. We will explore MovableType’s unique sampling framework, which divides sampling into local and global regimes, demonstrate its integration with high-quality, X-ray/Cryo-EM-refined structural models, and examine the impact of the newly available OpenMM-enhanced protocol on prediction accuracy. Case studies will illustrate the computational efficiency, predictive power, and drug discovery applications enabled by these innovations.

Conventional free energy methods, which rely on computationally intensive sampling techniques like molecular dynamics (MD), often face challenges in balancing accuracy and efficiency. MovableType addresses these limitations through a hybrid sampling approach. Local sampling, or “MT proper,” decomposes inter- and intramolecular energies into pairwise statistical or energetic potentials, enabling efficient exploration of local conformational space. Global sampling incorporates docking, rotamer enumeration, enhanced MD, and full MD/MC simulations to capture larger-scale conformational changes. With OpenMM now integrated as an optional MD engine in the DivCon Suite, users can leverage enhanced MD capabilities directly within MovableType workflows for greater flexibility and precision.

To ensure reliable predictions, the DivCon Suite employs X-ray/Cryo-EM density-driven refinement to generate chemically accurate, structurally complete models. Powered by semiempirical quantum mechanics (SE-QM) and mixed QM/MM methods, these refinements address key challenges, including protonation, truncations, and tautomeric/rotameric states, providing robust inputs for free energy workflows and AI/ML training datasets.