Determination of order parameters and correlation times in proteins: A comparison between Bayesian, Monte Carlo and simple graphical methods

We describe a novel approach to deducing order parameters and correlation times in proteins using a Bayesian statistical method, and show how likelihood contours, P(τ,S), and confidence levels can be obtained. These results are then compared with those obtained from a simple graphical method, as well as those from Monte Carlo simulations. The Bayes approach has the advantage that it is simple and accurate. Unlike Monte Carlo methods, it gives useful contour plots of probability (also not provided by the simple graphical method), and provides likelihood/confidence information. In addition, the Bayesian approach gives results in very good agreement with those obtained from Monte Carlo simulations, and as such use of Bayesian statistical methods appears to have a promising future for studies of order and dynamics in macromolecules.