A coarse-grained model for β-d-glucose based on force matching

Cellulosic ethanol production is a two-stage process that involves the hydrolysis of cellulose to form simple sugars and the fermentation of these sugars to ethanol. Hydrolysis of cellulose is the rate-limiting step, and there is a great need to characterize the process with numerical simulations to better understand the complex mechanisms involved. The ultimate goal is to generate accurate coarse-grained molecular models that are capable of predicting the structure of lignocellulose before and after pretreatment so that subsequent ab initio calculations can be performed to probe the degradation pathways. As a first step toward that goal, the force-matching method is used to derive coarse-grained models for β-d-glucose molecules in aqueous solution. Using the same reference, an all-atom molecular dynamics simulation trajectory, two sets of three- and six-site coarse-grained models of β-d-glucose are developed using two definitions of the coarse-grained center site location: center of mass (CG-CM) and geometric center (CG-GC). The performance of these coarse-grained models is evaluated by comparing the coarse-grained predictions for bond-length distributions and radial distribution functions to those obtained from the all-atom reference simulation. The six-site coarse-grained models retain more structural details than the three-site coarse-grained models. Comparison between center site definitions shows that CG-CM models generally predict local ordering better, while CG-GC models predict long-range structure better.