Multiparameter analysis of C2H2-ZF PPIs. (A) Estimated number of unique clusters of human C2H2-ZF proteins based on genomic binding sites, motifs, and PPIs. In each panel, the x-axis shows the number of clusters obtained by the PAM algorithm (R Core Team 2013), and the y-axis corresponds to the silhouette value, a measure of consistency of clustering. The blue dashed lines represent the largest number of clusters that result in 95% of the maximum silhouette value (Rousseeuw 1987), providing an estimate of the highest number of unique profiles (de Amorim and Hennig 2015) that retain high intra-cluster similarity and low inter-cluster similarity. (B) Correlation of functional parameters and sequence similarity among non-KRAB C2H2-ZF protein paralogs. For sequence comparison both ZF-only and full-length sequence without ZF were used. The color gradient corresponds to Pearson correlation between similarity measures. Red indicates positive correlation, i.e., when two paralogs are overall similar in one parameter, they are also similar in the other parameter, and when they have diverged in one parameter, they have also diverged in the other parameter. See also Supplemental Table S11.
