Methods

Molecular dynamics: In molecular dynamics (MD), the temporal evolution of all the atoms of the system of interest is simulated on a computer. The forces are often calculated using empirical force fields, which provide a reasonable balance of accuracy vs computational cost. One of the strong points of atomistic molecular dynamics is that it allows to take into account electrostatic effects, which have a very important role in charged molecules such as RNA. Moreover, effects such as molecular recognition,crucial in biomolecules, are included. In principle, MD is an extremely powerful microscope, which allows to perform in silico "experiments" and to look at the phenomena at atomistic resolution. In practice, its use is limited by two factors: the accuracy of the standard force fields, and the fact that typical biological phenomena happen on time scales which are too long for MD simulations. We work on both these issues, by increasing the accuracy of the force fields available on the shelf, and by tackling the time scale problem as discussed below.

Coarse-grained models: A possible route to tackle the time-scale problem is to work at a lower resolution, using coarser models. In a coarse-grained model, a group of atoms is represented as a single piece, thus reducing the number of degrees of freedom of the system and allowing to simulate it for a longer time. Traditional coarse-grained models are parametrized using information taken from experiment. However, they are not accurate enough to have a predictive power, and, with a few exceptions, they neglect electrostatic effects. This calls for new parametrizations for coarse-grained RNA models, based on atomistic MD simulations, and able to include electrostatic and molecular-recognition effects.

Rare events sampling: Another route to tackle the time-scale problem is to use special methods to accelerate the sampling of rare events. In the last years a lot of work has been done in this field, and a variety of techniques are now available, such as replica exchange, umbrella sampling, metadynamics, etc. Some of these techniques require the introduction of proper reaction coordinates which have to be developed ad hoc for the problem of RNA folding. More details on this point can be found here.

Non-equilibrium modeling: One of the interesting points of single-molecule experiments is that they allow to study non-equilibrium properties. When translated to MD simulations, this means that one is not anymore interested in the behavior of the system averaged over time, but in a dynamical description of its response to external perturbations [for a good introduction to non-equilibrium physics, see Nonequilibrium Statistical Mechanics, R. Zwanzig (2001)]. In particular, we are interested in modeling the effects of mechanical manipulation on RNA folding, working in collaboration with an experimental group.