Background: Protein phosphorylation is one of the pervasive and most important protein posttranslational modifications, which regulates the dynamic behavior and decision processes of eukaryotic cells.So far, thousands of phosphorylation sites have been discovered by proteome-wide mapping and other biochemical studies.However, information of specific protein kinase catalyzing these phosphorylation sites still remains very limited.For example, of 35, 187 phosphorylation instances for human deposited in the latest version of Phospho.ELM, one of the most widely used in vitro and in vivo phosphorylation database, only about 9% of them have corresponding kinase information.Methods: To shorten ever-widening gap between experimentally discovered phosphorylation sites and specific protein kinase information, we developed a novel computational method that precisely identify specific protein kinase for phosphorylation sites.By employing sequential information of phosphorylation sites and protein-protein interaction, this method achieves remarkable performance accompanied with extremely high specificity, a vital requirement for systematically linking protein kinase to phosphorylation sites.Results: By applying this efficient method to all available phosphorylation sites in Phospho.ELM, we successfully identified more than 1,900 novel protein kinase-substrate interactions.By integrating with experimentally identified human phosphorylation sites with corresponding kinase information, we finally obtained a human phosphorylation network with 210 protein kinases, 991 phosphorylated proteins and 2,914 kinase-substrate interactions.Further analysis showed this network exhibits distinct topological properties such as scalefree, small-world, etc.Function annotation of these components demonstrates they are significantly associated with complex human diseases such as breast cancer, etc.Conclusions: These results suggest that systematic identification can greatly help to reveal the complex biological mechanism of phosphorylation and study on human phosphorylation network may provide better understanding of the function of protein kinases and their implication in diseases .