Visualization of the OR sequence space by sequential principal components analysis (PCA). Axes correspond to the first three principal components (linear combinations of the similarities of each gene with every other) and are shifted to be optimally visible: their intersection need not represent the cartesian origin (0,0,0). The vertical axis represents the first principal component. Spheres represent olfactory receptors; small spheres denote OR pseudogenes. The ORs are colored by family as in Figure 1. PCA shows the clustering of genes of common family. The first principal component, by maximizing the variation in the set accounted for, consequently separates the set into two groups (e.g., ORs and GPCRs in panel a). To further resolve the clustering of ORs, we remove the smaller group and run PCA on the remaining group. (a) PCA of all ORs including a diverse set of non-OR GPCRs. (b) PCA of all ORs. (c) PCA of Class II ORs. (d) PCA of Class II ORs excluding family 4.
