To further investigate whether or not the flexibility was continuous, we proceeded as follows. Images from the two classes were pooled together and, using the ensemble map, subjected to a 3D principal component analysis (PCA). The approach we followed is based on Tagare et al. (2015 ▸), with some minor modifications of the method. A detailed explanation of the modifications is given in Section 2. We initialized the first principal component (PC) to the difference between the open and closed conformation, while the remaining PCs were initialized randomly. Upon convergence, the eigenvalue of each PC and the scatter of the images in PC space was calculated. The eigenvalues of the PCs are shown in Fig. 3 ▸(a). Clearly, the first three PCs are significant. The scatter plot of the image data in PC1–PC3 space is shown in Fig. 3 ▸(b). Fig. 3 ▸(b) strongly suggests that there is ‘continuous flexibility’ rather than ‘tightly clustered’ flexibility. Fig. 3 ▸(b) also shows the projection of the maps corresponding to the open and closed conformations on the extremes of the first three PCs. It is clear that the open and closed conformations are aligned mostly along the first PC, suggesting that the open/closed classification captures the most significant changes. Fig. 3 ▸(c) shows side views of a pair of structures (mean ± 2 × std, where std is the square root of the eigenvalue) for each PC. Additional details of these structures are available in Supplementary Figs. S4 and S5. Note that PCs are not to be understood as structural pathways with a biological meaning, but as directions that summarize the variance of a data set. For instance, the fact that the RBD appears and disappears at the two extremes of PC3 indicates that there is an important variability in these voxels, which is probably indicative of the up and down conformations of the RBD [to be understood in the context of the elastic analysis shown in Fig. 2 ▸(b)].