Nov 10 2020

11/10 – William Noid, The Pennsylvania State University

CBQB Seminar

November 10, 2020

1:30 PM - 2:30 PM

Address

Chicago, IL

Watch this webinar live with Zoom>>

Title: Exploring the landscape of coarse-grained representations for proteins

Abstract: Coarse-grained (CG) models are widely adopted for simulating complex biological systems. By representing molecules in reduced detail, CG models enable simulations on length- and time-scales that are inaccessible to conventional atomistic models. In this talk, we first examine how best to represent proteins in reduced detail, i.e., which features should be retained in the CG model. We employ Monte Carlo methods to statistically sample the space of CG representations. We quantitatively assess the quality of each representation based upon its information content and spectral quality. We find that these two metrics are anti-correlated in low-resolution representations. Interestingly, below a critical resolution, we observe a phase transition distinguishing "good" and "bad" CG representations. As time allows, we will also discuss a dual potential approach for CG models that accurately describe both structural and energetic properties, while even reproducing the thermodynamic specific heat.

Contact

cbqb@uic.edu

Date posted

Oct 29, 2020

Date updated

Oct 29, 2020