Fig. 1

Prediction of phosphorylation hotspots regions for eukaryotic domain families. a Phylogenetic tree of the species from which phosphorylation data has been obtained. The numbers in the left column correspond to the phosphosites per species obtained and the right column the phosphosites found within Pfam domains. b Hotspot regions are defined as those having higher than randomly expected number of phosphorylation. A rolling window is used to count the observed average number of phosphosites in the alignment (black line) and a background expectation is calculated from random sampling (gray line and gray band for standard deviation). A p-value is calculated for the enrichment of phosphorylation at each position and projected onto structural models. c The capacity to discriminate between phosphosites of known function from other phosphosites was tested using a ROC curve. We compared the discrimination power of the hotspot p-value (blue line for ST and yellow line for Y). d Enrichment over random of human phosphosites with known functions for residues predicted as a hotspot region when compared with the rest of the domain (blue for ST and yellow for Y; p-values for Fisher’s exact test)