ABSTRACT:
In this manuscript, I am proposing an approach for identification of correlated exchange in proteins via analysis of the NMR relaxation dispersion data. For a set of spins experiencing exchange, every relaxation dispersion datasets is fit individually and then—globally while paired with every other dataset. The corrected Akaike’ s Information Criteria (AICc) for individual and global fits are used to evaluate the likelihood of two spins to report on the same dynamic event. Application of hierarchical cluster analysis reveals correlated spin groups using the difference in AICcs as a measure of similarity within the pairs. This approach to detection of correlated dynamics is independent of accuracy of best-fit parameters rendering it less sensitive to experimental noise. High throughput and the absence of the operator bias might make it applicable to a relatively lower quality NMR relaxation dispersion data from large and poorly soluble systems.
