Proteins are by no means static and often undergo large-amplitude structural changes to perform their function. Thus, defining both the three dimensional molecular structure and its dynamical behavior is often essential for understanding how a protein functions. In many cases, high resolution structural information can be routinely obtained by X-ray crystallography or Nuclear Magnetic Resonance (NMR) spectroscopy. However, these experiments usually can not provide information about low-populated intermediate states and hence about the overall conformational dynamics.

Nowadays, due to the impressive progress in experimental techniques, such as pump-probe spectroscopy, advanced NMR experiments, neutron scattering and cryo-electron microscopy we have access to detailed information about the protein dynamics.

However, translating the experimental data into a detailed atomic view is generally non-trivial. In most cases, the experiments provide time- or ensemble-averaged data whose interpretation requires molecular modeling. A particularly challenging case is the study of proteins that exhibit a significant conformational flexibility which cannot be described by simple fluctuations around a single well-defined average structures [1].

Molecular dynamics (MD) simulations are a valuable tool for studying complex bio-molecular systems, providing an atomistic description of their structure and dynamics. Unfortunately, its predictive power has been limited by the complexity of the biomolecule free energy landscape, which prevented exhaustive sampling by means of standard MD. Recently the algorithmic advances in molecular dynamics codes [2,3] the use of specialized hardware [4] and distributed computing platforms[5] as well as the development of advanced sampling algorithms [6–9], have significantly alleviated the timescale limitation. Therefore, fully atomistic MD simulations can now approach timescales comparable to those characteristic of many experimental techniques, providing a powerful tool for interpreting and complementing the experiments. At this juncture, much is to be gained by combining MD simulations with experiments even further.  This can be achieved either by restraining the MD simulations to reproduce the experimental data or by calculating the experimental observables from extensive MD sampling. This trend has also been helped by the increase in time-resolution of the experiments [10]. Studies based on the combination of experiments and simulations to characterize large scale protein dynamics, generally happening in the time range of ms to s. are starting to appear in the literature (see State of the Art section). Moreover, quantitative comparison of MD studies with experiments allows not only to gain a deeper insight into protein biophysics but to devise more accurate force fields [11].

However, the communication between the experimental and computational community has been so far limited. There are very few occasions for the two communities to discuss about the common issues. Indeed, despite the progress, several issues are still open and obtaining a detailed understanding of large-scale protein dynamics is, in many cases, still a challenging task. Indeed, even if much attention has been devoted to the problem of conformational averaging, devising effective simulation algorithms or models that can circumvent this problem is far from trivial even for single-molecule experiments. Moreover, much work has still to be done in order to extend the timescales of both experiments and simulations in order to shed light into many relevant biological processes.

The uniqueness of this workshop is in bringing together theoreticians and experimentalists working on large scale conformational dynamics of bio-molecules. This will give them the opportunity to discuss on open issues and find possible solutions in novel experimental or computational techniques or in the combination of the two approaches.  This will surely enhance collaboration across different disciplines. We envision a cross fertilization that will  foster new collaborations among participants and lead to significant scientific advances.

Recently, a few seminal papers combining experiments and simulations to describe large-scale protein dynamics appeared in the literature. In those studies, the combination of MD simulations and advanced experiments was used to attain an unprecedented combination of high resolution in time and space. This concept was clearly illustrated in the case of adenylate kinase (AK), an enzyme that catalyzes the conversion of an ATP and an AMP molecule into two ADP molecules. Comparison of X-ray structures of substrate-bound and free AK suggested a ligand-induced conformational change. However, the combination of molecular simulations with crystallography, NMR spectroscopy, and single-molecule FRET experiments have shown a fundamentally different picture. Large-scale transitions between the fully open and fully closed states were detected for the free enzyme, demonstrating that the ligand has at most a "conformational selection" effect on the enzyme [10,12]. A similar conclusion was reached also in the case of ubiquitin. Using residual dipolar couplings and molecular dynamics simulations, it was shown that the native protein has considerable conformational heterogeneity [13]. Conformational selection, rather than induced-fit motion, suffices to fully explain the molecular recognition dynamics of ubiquitin[14]. In a separate study, enhanced sampling from biased potential molecular dynamics simulation, combined with NMR dipolar and scalar coupling data was used to define a self-consistent ensemble of conformations for solvated ubiquitin [15]. Regarding conformationally heterogeneous systems, in a recent paper Nettels et. al.[16] combined single-molecule FRET and other biophysical techniques with replica-exchange MD simulations to investigate the effect of the temperature on the dimensions of the unfolded state of a small cold shock protein under a wide range of different conditions. Their results remarkably showed a compaction of the unfolded state with increasing temperature. The explicit-solvent MD simulations allowed to shed further light on the microscopic origin of such surprising behavior, suggesting a crucial contribution of intra-molecular hydrogen bonding to the temperature-induced collapse. The potential of combining the atomic-level resolution of MD simulation with low-resolution experimental techniques has been clearly demonstrated also by Trabuco et. al.[17,18] who recently developed a novel MD-based method to fit atomic structures into electron microscopy (EM) maps. Techniques as cryo-EM indeed can image systems captured in different functional states, thus providing crucial information regarding the large macromolecular assemblies. The gap between the poor resolution of this approach and the atomic resolution of standard X-ray crystallography can be bridged by means of biased MD simulations. Such method, dubbed molecular dynamics flexible fitting (MDFF), was successfully applied to obtain high resolution structures of the E.Coli ribosome, paving the way for the characterization of complex cellular processes.