The role that modelling and computation has played in our understanding of how proteins and biomolecules work has been steadily increasing in the past 40 years. CECAM has played a crucial role in the establishment of a scientific community and the development of a steadily broadening field starting from a landmark workshop held in Orsay from May 24 to July 17, 1976. That workshop, entitled “Models for Protein Dynamics” organised by Herman Berendsen and with a format inspired by CECAM’s founder and then director Carl Moser, gathered a number of mostly young scientists who went on to become leaders in an area of research that has been growing ever since.
The increase in computer power since the 1976 workshop, although predicted, has been extraordinary. Yet computer power alone has not brought about the revolution that many predicted. In the 300 page workshop report are described a 8.8 ps simulation of a 58 amino-acid protein in vacuo (McCammon), and a 1 ps simulation in a crystal (Rahman & Hermans). Interestingly, the methodology has changed very little since, and several of the present-day challenges had already been clearly stated then. While enormous progress has been made on some issues, remarkably little has been made on others. Thanks to immense advances on the experimental techniques available to investigate biological molecules, and with the crucial contribution of computational methods in the analysis of experimental data, we now have a much broader, and deeper, fundamental understanding of how biological macromolecules behave. Biomolecular simulation has been instrumental in revealing the role of dynamics in biological function, and in directing the development of experimental techniques more suited to understand biological function, such as single molecule ones.
Despite 40 years of remarkable progress, applications of biomolecular simulation still lag behind. The role of simulation in the development of new drugs, for example, is modest relative to what may have been predicted, say, twenty years ago. Simulation has not replaced experimental structure determination of proteins; conversely, the importance of dynamics highlighted by the first simulations  is now broadly recognised as crucial to all properties of proteins and the structure–function paradigm, and provides a spotlight on functional disordered protein states, allostery and aggregation. Methods to treat quantum mechanical degrees of freedom or inclusion of polarisability in classical macromolecular models are still evolving but are currently computationally inefficient, while classical potentials, not dissimilar from those used 40 years ago, turn out to be surprisingly accurate and, with a few tweaks they allow ab initio folding of a few proteins to be observed during millisecond long simulations .
While accuracy in estimating atomic interactions and the extension of timescales through custom-built supercomputers are seen by some as the most promising avenues, others focus on the development of coarse-grained models  and approximate solutions [4, 5] as a way to elucidate general questions on the function of broad classes of biological questions, such as the function of molecular machines , molecular interactions in the crowded cellular environment , enzyme catalysis , etc. While some see experiment as a validation tool for molecular models and simulation as an alternative to experiments, others focus on the use of simulation as a tool to interpret experiments and bridge resolution gaps between different experimental techniques.
The scope of the proposed workshop/conference is to gather those who participated to the 1976 workshop “Models for Protein Dynamics” and current key players in the field of biomolecular simulation to discuss successes, failures and future challenges. What will computational studies of biomolecules deliver in the next 40 years? And what should we focus on now for that progress to occur faster?
This workshop will be the occasion for leaders in the field to present their vision, suggest goals that are within reach and those that require collective effort. In the spirit of the original CECAM workshop, even if their duration is something that cannot be replicated, a half-day session will be dedicated to small group discussions on specific challenges in biomolecular simulation.
It will also be an opportunity to discuss how to rigorously assess the quality of force-fields , coarse-grain methods [3, 10] and establish new ways of assess and compare simulation trajectories, as well as methods to determine free energy landscapes from trajectories.
The workshop will also provide an occasion to remember the contribution of some key figures who passed away, such as Carl Moser, CECAM’s founder and director in 1976, Anees Rahman and Peter Kollman.