Figure 8

Schematic representation of an evolutionary landscape of the segments. The contour map in a sequence space represents compatibility with the structure-based profile of the β-hairpin-like structure we identified. Points closer to the highest point (open circle) on the map are more compatible with the structure-based profile of cluster #159. Crosses and inclined crosses represent segments in presently existing proteins, which are classified respectively into two families, PF00014 and PF02910. Squares indicate ancestral sequences of each family, so that the map involves evolutionary directions of present segments from their ancestors. The sequence space in the map is defined by the PCA axes (PC1 and PC2). These axes were determined using principal component analysis (PCA) of sequences of all segments, in which the Hamming distance was used as a dissimilarity parameter between the two sequences. Contour levels shown with color scaling were drawn by the interpolation algorithm embedded in IGOR using the compatibility values S of both existing and virtual sequences.