Biophysical Society Thematic Meeting | Hamburg 2022
Biophysics at the Dawn of Exascale Computers
Wednesday Speaker Abstracts
EXPECTATION MAXIMIZED MOLECULAR DYNAMICS: RAPID ESTIMATION OF TRANSITION BARRIERS IN BIOMOLECULAR FREE ENERGY LANDSCAPES Neelanjana Sengupta 1 ; Pallab Dutta 1 ; 1 Indian Institute of Science Education and Research (IISER) Kolkata, Biological Sciences, Mohanpur, India Modern-day efforts towards uncovering emergent protein structure-function relationships incorporate methods that leverage data from both experiments and computational studies. A large part of such efforts is devoted to capturing the infrequent transitions that control complex biophysical processes over long timescales. Such efforts, however, are often challenged by the lack of objective definitions of system-specific slow modes, and sometimes by lack of access to high-end computational resources. This talk will present our recently proposed novel algorithm termed 'Expectation Maximized Molecular Dynamics' (EMMD) that incorporates a statistical inference-based approach in estimating free energy barriers related to rate-limiting transitions. The method bypasses conventional thermodynamic sampling by connecting metastable basins using Bayesian likelihood maximization. Furthermore, the method incorporates a tunable self- feedback protocol to prevent unnecessary sampling that does not effectively contribute to the underlying distributions of metastable states. Termed 'Expectation Maximized Molecular Dynamics' (EMMD), this algorithm has demonstrated significant efficacy in predicting the experimentally known free energy barriers in putative biomolecular systems. Ongoing efforts are geared towards leveraging the method to unravel key descriptive features that accompany conformational changes during the catalytic activity of a tyrosine kinase. Finally, as we step into the era of ‘exascale computing’, a glimpse of the emergent possibilities from integration of EMMD with other state of the art machine learning and enhanced sampling schemes will be discussed. Reference: P. Dutta and N. Sengupta, J. Chem. Phys. 153, 154104 (2020)
23
Made with FlippingBook Learn more on our blog