The importance of computer simulation and modeling in structural and molecular biology has been recognized since the beginning of the computer science era. Indeed, simulations are a bridge between the theories underlying (sub)molecular interactions and experiments, able to capture their behavior up to the macroscopic and statistical level. Therefore, the last six-decade challenge for theoretical and numerical chemists and physicists has been to elongate the side of the bridge from the atomistic to the macroscopic scales, exploiting the ever increasing computer capabilities.
It soon appeared clear that multi-scale approaches were a winning strategy in this challenge. Reducing the resolution of a system limits the cost of the simulation, allowing the representation of larger systems for longer times, and simplifies the analysis of results, forcing to keep only the relevant degrees of freedom. On the other hand, the capability of keeping or getting back to the atomistic view whenever needed brings unprecedented levels of insight at any scale. The power of this extremely flexible, yet accurate, approach was recognized already from the early studies on bio-molecules by pioneers such as Karplus, Berendsen, McCammon, Warshel and Levitt, tracing back to the seventies, and culminated in 2013 with the Nobel Prize in Chemistry, assigned for “the development of multiscale models for complex chemical systems”. While the current high performance computation resources allow simulations of entire viruses or portion of the cytoplasm even at the atomistic level, the advantages and challenges posed by multi-scale approaches are not exhausted.
Inspired by the workshop
Multiscale Modeling from Macromolecules to Cell (CECAM Lausanne, Feb 4-6 2019), this paper collection aims at including an exhaustive discussion on old and new issues emerging in multi-scale modeling, especially as the method is pushed towards the macroscopic scales. The inaccuracies of the standard atomistic force fields on the large time scales call for their complete rethinking towards a new class parameterized by combining bottom-up and top-down approaches. Related to this, the issue emerges about the compatibility between resolutions, from atomistic to different levels of coarse graining, up to the meso-scale, calling into play the comparison between diverse parameterization strategies (force or potential matching, relative entropy minimization, Boltzmann inversion and others) and different methods for the phase space exploration and evaluation of thermodynamic and statistical properties (e.g. Brownian dynamics, meta or accelerated dynamics, parallel tempering, path sampling methods etc). At the border with bio-engineering and bio-informatics, the largely unexplored domain of integrative approaches, ranging from the more familiar “embedding” (implicit solvent, Poisson-Boltzmann method…) to the combination of finite-difference based continuum representations with particle-based ones, is expected to give innovative contributions. Last, but not least, methods and algorithms to improve the scaling of calculations must be adapted to the increasing complexity of parallelism in computation resources.
Contributions illustrating any or more of these aspects will be welcome in this collection, even and especially if applied to real cases of biological interest. The contributors may choose any of the allowed article types in Frontiers in Molecular Biosciences – Biological Modeling and Simulation.