Lay summary
Different molecular processes can occur over extremely broad time-scales (from femtosecond to millisecond, seconds and minutes), and may involve very different numbers of atoms. Therefore, in some cases, large system sizes and long time periods place severe restrictions on the nature and type of simulations that can be carried out. In particular, biological systems fall into this category. In the first place, biomolecules such as proteins may contain up to several hundreds of amino acids, i.e., thousands of atoms. Secondly, biological processes may span characteristic time-scales from milliseconds to many minutes, hours or even days. In addition, these processes can occur in solution, and moreover solvent molecules can have an active role and need to be explicitly considered. Finally, dynamical effects both on short and long time scales are extremely important and must be taken into account. Thus, the predictive capability of computational models for biological systems is limited by the overall accuracy to which relevant phase-space regions are sampled. From this standpoint, a direct effort in exploration of novel techniques aimed at improving efficiency of phase-space sampling is needed to successfully develop and apply methods to soft-matter systems.
To bridge the gap between time scales of feasible simulations and those needed for the description of biologically relevant events, coarser description than all-atom Hamiltonians for biopolymers have been proposed. Such schemes are able to efficiently span the phase-space, but they lose the chemical details of a protein sequence. Therefore, they may not be used to investigate all those processes involving molecular recognition. Such processes, however, are usually highly localised, and involve portions of the protein only. Current trends in molecular simulations are driving efforts of different researchers into development of novel multi-scale techniques, able to combine coarse-grained approaches to atomistic details.
In the present proposal, I will complete and extend multiscale methodologies I have been developing in the course of the first years of my work. Such task will be achieved starting from the seeding works as well as ongoing developments achieved in the past years. In particular, I will complete development of a new CG force field for proteins able to reliably span the conformational space, and couple it to hybrid MM/CG methods. Applications to different systems of relevance in biophysics and biochemistry are presented in the second part of the research plan.