Collaborative Research: A Control Theoretic Framework for Guided Folding and Unfolding of Protein Molecules

Regents Of The University Of Michigan - Dearborn, Dearborn MI

Investigators

Abstract

This grant will fund research that enables accurate prediction of pathways for protein folding and unfolding, with application to computer-aided anti-viral drug design, control of protein-based nano-machines, and treatment of diseases related to protein misfolding such as Alzheimer’s, thereby promoting the progress of science, and advancing the national health and prosperity. Physics-based approaches reliably capture the processes that govern conformational changes of protein molecules, but typically do so at great computational expense. A recently developed modeling paradigm, which describes protein molecules in terms of large numbers of rigid nano-linkages that fold under the influence of interatomic forces, can significantly reduce the computational burden, but presents challenges with ensuring that the predicted folding and unfolding pathways are realistic and not artificially driven by the numerical algorithm. In this project, this challenge is overcome using an optimization-based control theoretic framework to guide both folding and unfolding dynamics while respecting biologically realistic rates of change of conformational entropy. Knowledge gained from the development of this framework will enable systematic investigation of protein conformational dynamics, including unfolding pathways of coronavirus spike proteins, while also advancing previously unexplored control tools that may help robots navigate cluttered environments. A unique approach to sonification of protein pathway data will make this knowledge broadly accessible and will be integrated in course projects for undergraduate students in engineering, computer science, and art, as well as in research activities aiming to mentor high school students in STEM. This research aims to bridge the two seemingly unrelated fields of optimization-based nonlinear control and conformational dynamics of proteins through rigorous development and investigation of computationally efficient and numerically stable algorithms that accurately predict protein folding and unfolding while avoiding pathways associated with artificially rapid loss of conformational entropy. This project will fill the critical gap in knowledge of encoding entropy-loss constraints using the kinetostatic compliance method by developing a novel non-iterative, large-scale, quadratic programming-based control scheme over hyper-ellipsoids for protein folding dynamics with large state-space dimensions; constructing a large-scale, variable-step-size, numerical integration algorithm that is expected to reduce the number of integration steps, where each step requires the burdensome computation of a very large interatomic force vector field; and developing a control theoretic approach for systematically investigating the problem of protein unfolding. Ground truth data for validation will be obtained from all-atom molecular dynamics simulations and, in the case of the model protein barnase, publicly available experimental data from optical tweezer-based mechanical unfolding experiments. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.