Challenges and opportunities in connecting simulations with experiments via molecular dynamics of cellular environments

Abstract

Computer simulations are widely used to study molecular systems, especially in biology. As simulations have greatly increased in scale reaching cellular levels there are now significant challenges in managing, analyzing, and interpreting such data in comparison with experiments that are being discussed. Management challenges revolve around storing and sharing terabyte to petabyte scale data sets whereas the analysis of simulations of highly complex systems will increasingly require automated machine learning and artificial intelligence approaches. The comparison between simulations and experiments is furthermore complicated not just by the complexity of the data but also by difficulties in interpreting experiments for highly heterogeneous systems. As an example, the interpretation of NMR relaxation measurements and comparison with simulations for highly crowded systems is discussed.

Large-Scale Conformational Changes and Protein Function: Breaking the in silico Barrier

“…We have revisited the many approaches available to explore the FEL of proteins, optimizing hardware, software and algorithms pursuing the dream of the seconds-long sampling. From a completely different standpoint, simulations in crowded cell-like soups of multiple copies of the same protein, although still in the ns-scale, are already a reality that holds promise to reveal dynamical complexity in local microenvironments, providing yet another approach to the sampling problem (Yu et al, 2016; Feig et al, 2018). We have also briefly mentioned machine learning algorithms, paradigmatic of a series of novel fast-developing non-physically based strategies which are gaining ground to study transitions, either alone or in combination with MD or CG-methods: from co-evolution analysis (Morcos et al, 2013; Sutto et al, 2015; Sfriso et al, 2016) to cross-correlation, network and community approaches (Potestio et al, 2009; Morra et al, 2012; Rivalta et al, 2012; Papaleo, 2015; Negre et al, 2018), neural networks and deep learning (Ung et al, 2018; Degiacomi, 2019), or integrative sequence and structural analysis (Flock et al, 2015).…”

An Overview of Molecular Modeling for Drug Discovery with Specific Illustrative Examples of Applications

“…The large amounts of data coupled with the high degree of complexity in many systems presents formidable challenges in managing, analyzing, and interpreting such big data in comparison with experiments that are being discussed. As traditional approaches to the analysis of simulations do not scale well to highly complex systems of macromolecules, a greater emphasis on automated machine learning and artificial intelligence will be required in the future [56,57]. On the other hand, the increased capabilities and flexibility of recent modern graphics processing units (GPUs) hardware combined with high level GPU programming languages such as CUDA and OpenCL has made computational power accessible to computational community.…”

ProLint: a web-based framework for the automated data analysis and visualization of lipid–protein interactions

“…Effective communication of scientific results is contingent on the careful analysis of the collected data and its display in a condensed but still information rich format. This is particularly the case in computational studies, such as Molecular Dynamics (MD) simulations, where even using conservative output-control parameters tera-bytes of data are continuously generated ( 1 , 2 ). When studying interactions between membrane proteins with their surrounding lipid environment, MD simulations have proven to be an excellent tool to probe many details of their interplay that are inaccessible to experiments.…”