Figure 1:

Clustering of an amalgamated trajectory of five anisotropic network models (ANMs) of a 12 bead system. (A) Schematic of the five topologies considered including left-handed helix, right-handed helix, beta-sheet like, and partially unfolded (PUBS), and linear. (B) Log likelihood per frame as a function of number of clusters for uniform shape-GMM and weighted shape-GMM. Each model has two corresponding curves: the log likelihood per frame of the training set and the log likelihood per frame from the cross-validation (CV) set. Error bars are the standard deviation of sampling ten different training sets for each model. Shaded regions denote the 90% confidence interval.