Submitted on May 9, 2007
Revised on September 27, 2007
Accepted on October 13, 2007
Phosphoblast: A computational tool for comparing phosphoprotein signatures among large datasets
Yingchun Wang and Richard L. Klemke
Department of Pathology and Moores Cancer Center, University of California, San Diego, La Jolla, CA 92093
Corresponding Author: rklemke{at}ucsd.edu
Identification of specific protein phosphorylation sites provides predicative signatures of cellular activity and specific disease states such as cancer, diabetes, Alzheimer’s, and rheumatoid arthritis. Recent progress in phosphopeptide isolation technology and tandem mass spectrometry has provided the means to identify thousands of phosphorylation sites from a single biological sample. These advances now make it possible to profile global changes in the phosphoproteome at an unprecedented level. However, while this technology is generating a wealth of information, there is currently no efficient means to identify phosphoprotein signatures shared among large phosphoprotein databases. Identification of common phosphoprotein signatures found in biological relevant systems and their conservation throughout evolution would provide valuable insight into mechanisms of signal transduction and cell function. Here we describe the development of a computational program (PhosphoBlast) that can rapidly match thousands of phosphopeptides that share phosphorylation sites within and across species. PhosphoBlast analysis of several large phosphoprotein datasets from the literature revealed common phosphorylation signatures shared across diverse experimental platforms and species. Moreover, PhosphoBlast is a powerful analysis tool to identify specific phosphosite mutations. Comparison of the mouse and human phosphoproteomes revealed more than 130 specific phosphoamino acid mutations, some of which are predicted to alter protein function. Further analysis revealed that known phosphorylated amino acids are more evolutionally conserved than the S/T/Y amino acids not known to be phosphorylated. Together our results demonstrate that PhosphoBlast is a versatile mining tool capable of identifying related phosphorylation signatures and phosphoamino acid mutations among complex proteomic datasets in a highly efficient and accurate manner. PhosphoBlast will aid in the informatic analysis of the phosphoproteome and the identification of phosphoprotein biomarkers of disease.
