Research Projects

Modeling chemical reactions and structural transformations in large biological systems remains challenging. The slow exploration of conformations in molecular dynamics (MD) simulations makes it difficult to provide accurate free energies and kinetics. While several enhanced sampling methods are available, there is still a need for efficient techniques that can accelerate conformational sampling and leverage modern high-performance computing platforms to speed up calculations. To address these, we propose the development of two new methods: Solute Tempered – Temperature Accelerated Sliced Sampling (ST-TASS) and Slice Exchange ST-TASS (SEST-TASS). Our group has previously demonstrated that Temperature Accelerated Sliced Sampling (TASS) is an efficient approach for modeling chemical reactions and conformational changes in systems with tens of thousands of atoms. TASS achieves controlled sampling and can accommodate a wide range of collective variables (CVs), which can lead to quicker convergence in free energy estimates. In this project, we aim to enhance the TASS to effectively handle larger systems and model more complex processes. We consider two strategies: 1) combining TASS with solute-tempering, and 2) implementing a slice-exchange protocol alongside ST-TASS. Since TASS is a collective variable (CV)-based sampling method, integrating it with a generalized ensemble approach, which can accelerate global sampling, will be highly beneficial. Such hybrid strategies are known to significantly improve sampling efficiency. The implementation of these two methods will be done within an open-source software designed for biomolecular simulations. To take advantage of the capabilities of parallel computing platforms, replicas and slices can be distributed across GPU and CPU nodes. We propose new approaches for analyzing the high-dimensional free energy surfaces generated by these techniques, as well as for calculating kinetics. These features will help us to compare our computational results with experimental data on structure and kinetics. We will explore the applications of these methods that are crucial for therapeutic purposes, particularly in the treatment of obesity and diabetes. Our first focus will be on glucagon-like peptide-1 (GLP-1) receptors (GLP1R). Shorter peptides inspired by GLP-1 have been found to stimulate insulin release and inhibit glucagon release, making them useful in treating type 2 diabetes and obesity. One specific peptide of interest is Peptide 20, which has demonstrated triagonist properties and is currently undergoing clinical trials. However, the detailed molecular interactions involving Peptide 20 that result in structural changes and receptor activation remain unclear. We aim to investigate this through MD simulations and free energy calculations. Given the size of Peptide 20 and the membrane-bound receptor, as well as the influence of the membrane and transmembrane domains, we anticipate that advanced techniques like ST-TASS and SEST-TASS will be essential for obtaining reliable information about the binding of the peptide and its subsequent activation. The second research problem of interest is the aggregation of intrinsically disordered proteins. Specifically, we are focused on simulating the initial interactions between amylin proteins during the formation of a stable dimeric structure. We will consider both human and rat amylin. The aggregation of human amylin is associated with type 2 diabetes. Its adapted metastable alpha-helical and beta-sheet conformations trigger further aggregation processes. Therefore, it is crucial to explore these conformations along with the mutual interactions between the two protein chains. We think ST-TASS and SEST-TASS methods are ideal for such simulations and for obtaining configurational free energy landscapes. These calculations will provide insights into the mechanisms, energetics, and kinetics of aggregation during the initial phases.