Predicting Posttranslational Lysine Acetylation Using Support Vector Machines

Bioinformatics Advance Access published May 26, 2010

Predicting Posttranslational Lysine Acetylation Using Support Vector Machines

Florian Gnad 1,2, Shubin Ren 1, Chunaram Choudhary 1,3, Jürgen Cox 1 and Matthias Mann 1*

1. Department of Proteomics and Signal Transduction, Max-Planck-Institute of Biochemistry, Am Klopferspitz 18, D-82152 Martinsried, Germany
2. Department of Systems Biology, Harvard Medical School, 200 Longwood Avenue, Boston, MA 02115, USA
3. The Novo Nordisk Foundation Center for Protein Research, Faculty of Health Sciences, University of Copenhagen, Bleg-damsvej 3, DK-2200 Copenhagen, Denmark
* To whom correspondence should be addressed. Prof. Matthias Mann

Motivation: Lysine acetylation is a posttranslational protein modification and a primary regulatory mechanism that controls many cell signaling processes. Lysine acetylation sites are recognized by acetyltransferases and deacetylases through sequence patterns (motifs). Recently we used high-resolution mass spectrometry to identify 3600 lysine acetylation sites on 1750 human proteins covering most of the previously annotated sites and providing the most comprehensive acetylome so far. This data set should provide an excellent source to train support vector machines allowing the high accuracy in silico prediction of acetylated lysine residues.

Results: We developed a support vector machine to predict acetylated residues. The precision of our acetylation site predictor is 78% at 78% recall on input data containing equal numbers of modified and non-modified residues.

Availability: The online predictor is available at http://www.phosida.com