Figure 2.

The map of the CATH hierarchy. For visual clarity we have omitted the CATH chains with “0” name (about half the chains). The full map is similar, but more cluttered. The figure is a projection from ℝ³⁰ to ℝ² along the directions of largest variation. The structures shown are representatives of their respective classes. The rectangle in the Upper Left contains all the chains in CATH, colored according to their class membership: α, β, αβ, and little secondary structure. Note that the αβ members are located between α and β islands, a consequence of the triangle inequality. Next, in the Upper Right, we enlarge CATH class 1, the helical proteins, and its five constituent architectures. In the Lower Right we display in larger details topologies 1.10.(220–250). Finally, in the Lower Left we show the H categories of topology 1.10.238. Although it is difficult to reproduce the cluster separation adequately in two dimensions, when one descends to lower levels of the hierarchy and the topological diversity, the clusters corresponding to different folds become more distinct.