Research Progress | Structural Coarse-Graining via Multiobjective Optimization with Differentiable Simulation

Recently, A study titled "Structural Coarse-Graining via Multiobjective Optimization with Differentiable Simulation" was recently published in the Journal of Chemical Theory and Computation.

The results of the study, led by Dr. Zhenghao Wu from the School of Science at Xi 'an Jiaotong-Liverpool University, introduce a versatile method called differentiable coarse-graining (DiffCG), which combines multiobjective optimization and differentiable simulation techniques. This approach aims to optimize effective potentials to match multiple target properties simultaneously, offering a novel strategy for creating efficient coarse-grained (CG) representations of soft matter systems.

Read the Article: https://pubs.acs.org/doi/10.1021/acs.jctc.3c01348

Dr. Wu mentioned that as a complement to experimental methods, molecular simulation has become a powerful tool for studying the behavior of complex soft matter systems at the microscopic level. Although all-atom models can provide direct and accurate predictions of the structure-property relationships of complex systems with rich atomic details, there are still a considerable number of phenomena in systems such as macromolecules that cannot be achieved through all-atom simulations. To overcome these limitations, CG models have been developed, in which multiple atoms or molecules are represented by a single interacting entity, reducing computational costs while still capturing the fundamental characteristics of the system. Using CG models, molecular systems with a wide range of spatial and temporal scales can be explored.

Table 1: (a) Schematics of a coarse-grained polystyrene chain for the illustration of separations in bonded interactions: bond length L, bond angle φ, and dihedral angle φ; (b)Coarse-grained mapping between all-atom model and coarse-grained model of a ten-monomer polystyrene chain.

“Recently, we present a versatile method, namely, differentiable coarse-graining (DiffCG), which combines multi-objective optimization and differentiable simulation. The DiffCG approach is capable of constructing robust CG models by iteratively optimizing effective potentials to simultaneously match multiple target properties.”

Table 2: A general workflow for the differentiable coarse-graining: mapping between fine-grained (FG) and coarse-grained (CG) systems; simulation sampling; potential energy functions U (Θ); multiple observables (e.g., radial distribution functions, bond length distributions, pressure, etc.).

Dr. Wu and his team demonstrate their approach by concurrently optimizing bonded and non-bonded potentials of a CG model of polystyrene (PS) melts. The resulting CG-PS model effectively reproduces both the structural characteristics, such as the equilibrium probability distribution of microscopic degrees of freedom, and the thermodynamic pressure of the AA counterpart. More importantly, leveraging the multi-objective optimization capability, they develop a precise and efficient CG model for PS melts that is transferable across a wide range of temperatures, i.e., from 400 to 600 K. It is achieved via optimizing a pairwise potential with non-linear temperature dependence in the CG model to simultaneously match target data from AA-MD simulations at multiple thermodynamic states. The temperature transferable CG-PS model demonstrates its ability to accurately predict the radial distribution functions and density at different temperatures, including those that are not included in the target thermodynamic states. Their work opens up a promising route for developing accurate and transferable CG models of complex soft-matter systems through multi-objective optimization with differentiable simulation.

Materials and proofreading: Dr. Zhenghao Wu