NetPhosYeast: Prediction of protein phosphorylation sites in yeast.
Christian Ravnsborg Ingrell1, Martin Lee Miller2, Ole Nørregaard Jensen1 and Nikolaj Blom2.
Accepted for publication in Bioinformatics, 2007.

1University of Southern Denmark, Campusvej 55, DK-5230, Odense M, Denmark
2Center for Biological Sequence Analysis, BioCentrum-DTU, The Technical University of Denmark, DK-2800 Lyngby, Denmark


We here present a neural network based method for the prediction of protein phosphorylation sites in yeast - an important model organism for basic research. Existing protein phosphorylation site predictors are primarily based on mammalian data and show reduced sensitivity on yeast phosphorylation sites compared to those in humans, suggesting the need for a yeast-specific phosphorylation site predictor. NetPhosYeast achieves a correlation coefficient close to 0.75 with a sensitivity of 0.84 and specificity of 0.90 and outperforms existing predictors in the identification of phosphorylation sites in yeast.