Protein phosphorylation is a crucial regulatory mechanism in various organisms. With recent improvements in mass spectrometry, phosphorylation site data are rapidly accumulating.  We developed P3DB (http://www.p3db.org/), a comprehensive resource of protein phosphorylation data from multiple plants. With a web-based user interface, the database is browsable, downloadable and searchable by protein accession number, description and sequence. Despite this wealth of data, computational prediction of phosphorylation sites remains a challenging task. This is particularly true in plants, due to the limited information on substrate specificities of protein kinases in plants and the fact that current phosphorylation prediction tools are mostly trained with kinase-specific phosphorylation data from non-plant organisms.  To address these issues, we proposed a new machine learning approach for phosphorylation site prediction. We incorporate features of amino acid frequencies, protein disorder information, k-nearest neighbors (KNN) using support vector machines (SVM) with supervised or semi-supervised learning for predicting phosphorylation sites. Test results on the PhosPhAt dataset of phosphoserines in Arabidopsis and the TAIR7 non-redundant protein database show good performance of our proposed phosphorylation site prediction method. Our method combined both KNN to take advantage of potential similarities to known phosphopeptides and SVM to account for generic sequence features. As more phosphorylation sites are experimentally identified, the accuracy of our method is expected to increase automatically.

Bio:
Dong Xu is James C. Dowell Professor and Chair of Computer Science Department, with appointments in the Christopher S. Bond Life Sciences Center and the Informatics Institute at the University of Missouri. He obtained his Ph.D. from the University of Illinois, Urbana-Champaign in 1995 and did two-year postdoctoral work at the US National Cancer Institute. He was a Staff Scientist at Oak Ridge National Laboratory until 2003 before joining University of Missouri. His research includes protein structure prediction, high-throughput biological data
analyses, in silico studies of plants, microbes, and cancers. He has published more than 150 papers. He is a recipient of 2001 R&D 100 Award and 2003 Federal Laboratory Consortium’s Award of Excellence in Technology Transfer. He is a member of the Editorial Board for “Current Protein and Peptide Science” and “Applied and Environmental Microbiology”. He is a standing member of the NIH Biodata Management and Analysis Panel.