| Popis: |
Advances in the atomistic understanding of biomolecular structure and function come with wide applications in medicine, pharmacology and biomaterials, as well as with fundamental answers about the nature of living things. Such insight can in principle be brought by molecular dynamics (MD) simulations. However, even with nowadays vast computational resources, the timescales that are usually accessible in standard MD simulations cannot reliably sample many relevant biological processes, which occur either too slowly or too infrequently. We refer to this as the "rare event" problem. To overcome it, the MD community has devised an arsenal of enhanced sampling techniques, which deliver insight in the form of free-energy landscapes, projected on key descriptive molecular degrees of freedom, i.e. collective variables (CVs). Nonetheless, the computational cost scales exponentially with the number of CVs; meaning that complex transitions are typically out of grasp. In this thesis, we introduce novel path-based enhanced sampling strategies, and also combine them with other powerful simulation techniques; thus, advancing the capabilities to navigate complex molecular transitions. Our framework enables the study of intricate conformational and chemical changes in a variety of biomolecules; including oligopeptides, sensor proteins, DNA and polysaccharides. From each system, we extract fine mechanistic details, free energies and biological insights. We hope for these advancements to contribute to the long-term goals of biomolecular simulation; enabling the atomistic understanding of ever larger, more complex and realistic living systems. |
