Extracting Experimental Measurables from Molecular Dynamics Simulations of Membranes

Atomistic molecular dynamics (MD) simulations are a powerful computational tool for probing the atomic-scale details of phospholipid bilayer structure and dynamics. Careful validation of the simulations is a critical step if the simulator is to establish productive contact with the experimentalist. The most fruitful approach is to directly compare simulated properties that correspond to primary experimental data, rather than to experimentally reported values which themselves are based upon model assumptions. For example, in the case of X-ray scattering, the most common validating criterion is to compare simulated and experimental cross-sectional areas per lipid molecule, even though that parameter is not directly obtained from the primary scattering profile. Instead, one is better off directly comparing the scattering profiles themselves, which are easily obtained from simulation via the Fourier transform of the simulated electron density profile. Here, we describe a number of important instances in the literature in which this preferred approach has been taken. In particular, we outline both the history and algorithms for the extraction of experimentally measured low- and wide-angle X-ray scattering (which reflect the overall bilayer structure and lipid chain packing, respectively) and both ¹³C- and ³¹P-NMR relaxation rates (which reflect the chain and headgroup dynamics). Finally, we discuss the exciting future role of coarse-grained MD simulations in extending the accessible phenomena that can be simulated, with a prescription for how to maintain an intimate connection to experiment by way of reverse coarse-graining algorithms.