O-glycosylation sites prediction using the meta-prediction approach

Glycosylation is one of important post-translational modifications. It is the chemical modification of a protein after its translation. It is a common biological mechanism for regulating protein localization, function, cellular communication, and turnover. The function of a modified protein is often strongly affected by these modifications. Thus increased knowledge about the glycosylation of a target protein may increase our understanding of the molecular processes in these proteins. Many methods for predicting glycosylation sites in protein sequences have been developed. Finding effective meta-prediction strategies that integrate different kinds of glycosylation sites predictors to achieve higher prediction performance is highly required. In this paper, we use the framework of multiplicative update algorithms in on-line decision problem to obtain more accurate meta-predictors. The experimental results show that performances of our meta-predictor are better than Oglyc, DictyOGlyc, YinOYan, NetOGlyc, GlycoEP in O-glycosylation sites prediction.