Skip navigation







FIELD Computational Sciences
DATE November 21 (Thu), 2019
TIME 10:30-12:00
PLACE 7323
SPEAKER Jaehyeok Jin
HOST Hyeon, Changbong
INSTITUTE University of Chicago
TITLE Advances and Challenges in High Fidelity Bottom-up Coarse-Graining: Accuracy, Representability, and Transferability
ABSTRACT Advances in theoretical and computational methodologies have relied on coarse-grained (CG) models to facilitate efficient simulation of complex systems by averaging the details beneath the CG resolution. Systematically-derived CG models from FG simulations (so-called bottom-up models) have been shown to reproduce microscopic properties yet face major limitations in accuracy, representability, and transferability. This talk will focus on the recent developments in CG theory to resolve the aforementioned issues. Despite its efficiency, a CG model may lose important information below the CG resolution during the CG process, resulting in an inaccurate model. To surmount this limitation, the Ultra-Coarse-Graining (UCG) theory has been developed to incorporate discrete state variables within conventional configurational variables. The presence of internal states in the CG site greatly expands the possible range of systems amenable to accurate CG modeling, including solvophobic association, interfacial systems, and even hydrogen bonding interactions. In regards to the applicability of the CG model, both representability and transferability are the bottlenecks in the versatility of the CG model and are closely related problems. Based on the approximated CG potential of mean force (PMF), we utilize the energy-entropy decomposition of the CG PMF to address these fundamental issues. We find that the entropic component to the CG PMF can be represented using additive pairwise contributions, which we show is highly coupled to the CG configurational entropy. This approach rigorously connects the missing entropy that is lost, due to the degrees of freedom that are thrown away and lost during the CG process, to the CG entropy. By design, the developed framework imparts transferable CG interactions across different state points i.e., transferability over different temperatures or across bulk liquids and their mixtures. Structural correlations of the fitted CG liquid model are found to corroborate a high-fidelity combining rule. Our findings elucidate the physical nature of the CG entropy for the first time and suggest a novel approach for resolving the transferability problem toward the predictive multiscale modeling.

References [* contributed equally]
(1) J. F. Dama,* J. Jin*, G. A. Voth, The Theory of Ultra-Coarse-Graining. 3. Coarse-Grained Sites with Rapid Local Equilibrium of Internal States, J. Chem. Theory Comput. 2017, 13, 1010
(2) J. W. Wagner, T. Dannenhoffer-Lafage, J. Jin, G. A. Voth, Extending the range and physical accuracy of coarse-grained models: Order parameter dependent interactions. J. Chem. Phys. 2017, 147, 044113
(3) J. Jin and G. A. Voth, Ultra-Coarse-Grained Models Allow for an Accurate and Transferable Treatment of Interfacial Systems, J. Chem. Theory. Comput. 2018, 14, 2180
(4) J. Jin, Y. Han, G. A. Voth, Ultra-Coarse-Grained Liquid State Models with Implicit Hydrogen Bonding, J. Chem. Theory. Comput. 2018, 14, 6159
(5) J. Jin,* Y. Han,* G. A. Voth, Coarse-graining involving virtual sites: Centers of symmetry coarse-graining. J. Chem. Phys. 2019, 150, 154103
(6) J. Jin, A. J. Pak, G. A. Voth, Understanding missing entropy in coarse-grained systems: Addressing issues of representability and transferability. J. Phys. Chem. Lett. 2019, 10, 4549
(7) J. Jin, G. A. Voth, Communication: Resolving Paradox of the Configurational Entropy in the Coarse-Grained System. 2019, In preparation
  • list