Molecules in Motion: Computing Structural Flexibility

Amarda Shehu

PhD Candidate
Department of Computer Science
Rice University

Abstract

Growing databases of genomic data call for computational methods to extract structural and functional information. Elucidating how a protein sequence determines a three-dimensional structure that carriesout biological function remains a main challenge in molecular biology. In flexible molecules like proteins, biological function cannot be reliably extracted from static structure data. Structural information needs to be enhanced with motion data to understand how motions are employed to modulate function and inter-molecular associations.

This talk presents a novel framework to characterize functionally-relevant motions by computing the ensemble of conformations assumed by a protein under physiological (native) conditions. Probabilistic exploration conducted hierarchically and at multiple resolutions is proposed to efficiently probe the vast high-dimensional protein conformational space and detect emerging energy minima. Data mining techniques combined with a statistical mechanics formulation allow to quantify the relative population of the different substates comprising the native ensemble of conformations.

Applications to proteins of various native topologies show that the proposed framework reproduces with high accuracy wet-lab data corresponding to a broad range of timescales. In addition, the framework is able to extract from sequence data diverse conformational states both in cyclic cysteine-rich peptides and in proteins where large-scale concerted motions connect different functional states.

The proposed framework provides a bridge between computer science, biophysical theory, and wet-lab applications. By obtaining an in-silico view of molecular motions, the framework has the potential to complement wet-lab experiments in the study of function and mechanism in naturally-occurring and engineered biomacromolecules.

Speaker Bio

Amarda Shehu is a Ph.D. candidate in the department of Computer Science at Rice University and an NIH fellow of the Nanobiology Training Program of the Gulf Coast Consortia. Her research focuses on the development and application of accurate and reliable computational methods to conduct biological research in silico. Her research interests encompass computational biology, bioinformatics, nanobiology, and biomimetics.