New generation sequencing technologies have resulted in significant increases in the number of complete genomes. Functional characterization of these genomes, such as by high-throughput proteomics, is an important but challenging task due to the difficulty of scaling up existing experimental techniques. By use of comparative genomics techniques, experimental results can be transferred from one genome to another, while at the same time minimizing errors by requiring discovery in multiple genomes. In this study, protein phosphorylation, an essential component of many cellular processes, is studied using data from large-scale proteomics analyses of the phosphoproteome. Phosphorylation sites from Homo sapiens, Mus musculus and Drosophila melanogaster phosphopeptide data sets were mapped onto conserved domains in NCBI's manually curated portion of Conserved Domain Database (CDD). In this subset, 25 phosphorylation sites are found to be evolutionarily conserved between the three species studied. Transfer of phosphorylation annotation of these conserved sites onto sequences sharing the same conserved domains yield 3253 phosphosite annotations for proteins from coelomata, the taxonomic division that spans H. sapiens, M. musculus and D. melanogaster. The method scales automatically, so as the amount of experimental phosphoproteomics data increases, more conserved phosphorylation sites may be revealed.