Construction and Deciphering of Human Phosphorylation-Mediated Signaling Transduction Networks

Protein phosphorylation is the most
abundant reversible covalent
modification. Human protein kinases participate in almost all biological
pathways, and approximately half of the kinases are associated with
disease. PhoSigNet was designed to store and display human phosphorylation-mediated
signal transduction networks, with additional information related
to cancer. It contains 11 976 experimentally validated directed edges
and 216 871 phosphorylation sites. Moreover, 3491 differentially expressed
proteins in human cancer from dbDEPC, 18 907 human cancer variation
sites from CanProVar, and 388 hyperphosphorylation sites from PhosphoSitePlus
were collected as annotation information. Compared with other phosphorylation-related
databases, PhoSigNet not only takes the kinase–substrate regulatory
relationship pairs into account, but also extends regulatory relationships
up- and downstream (e.g., from ligand to receptor, from G protein
to kinase, and from transcription factor to targets). Furthermore,
PhoSigNet allows the user to investigate the impact of phosphorylation
modifications on cancer. By using one set of in-house time series
phosphoproteomics data, the reconstruction of a conditional and dynamic
phosphorylation-mediated signaling network was exemplified. We expect
PhoSigNet to be a useful database and analysis platform benefiting
both proteomics and cancer studies.