Inferring Protein Structure from Limited and Sparse Data
In cases where crystallography is impossible and where only limited data is available, it can be quite challenging for experimental scientists to infer structures of proteins. In this video, OpenMM 2013 Visiting Scholar Justin MacCallum describes his project to develop computational tools that can be combined with limited, sparse, and noisy experimental data to produce reliable structural models, such as for membrane proteins. Dr. MacCallum's framework, called Modeling with Limited Data (MeLD), combines the power of molecular dynamics simulations with statistical inference and Monte Carlo simulation with the goal of solving this problem.