Molecular dynamics simulations of RNA: providing a computational perspective to augment experimental data while addressing IInherent limitations in the Method

Advances in computer hardware have enabled routine MD simulations of systems with tens of thousands of atoms for up to microseconds (soon milliseconds). The key limiting factor in whether these simulations can advance hypothesis testing in active research is the accuracy of the force fields. In many ways, force fields for RNA are less mature than those for proteins. Yet even the current generation of force fields offers benefits to researchers as we demonstrate with our re-refinement effort on two RNA hairpins. Additionally, our simulation study of the binding of 2-aminobenzimidazole inhibitors to hepatitis C RNA offers a computational perspective on which of the two rather different published structures (one NMR, the other X-ray) is a more reasonable structure for future CADD efforts as well as which free energy methods are suited to these highly charged complexes. Finally, further effort on force field improvement is critical. We demonstrate an effective method to determine quantitative conformational population analysis of small RNAs using enhanced sampling methods. These efforts are allowing us to uncover force field pathologies and quickly test new modifications. In summary, this research serves to strengthen communication between experimental and theoretical methods in order produce mutual benefit.