| One of the main challenges in biological research is to determine protein conformation (3-D structure) and function(s), which in turn has several applications in novel drug discovery and disease treatment. Proteins have a lower level structure, called secondary structure, which contributes to protein conformation. The focus of this research is to predict the content of the secondary structure of protein sequences. Protein secondary structure content prediction can be utilized to predict protein conformation. Current alignment based approaches are applicable to the content prediction problem in case high homology is present between proteins with unknown structure and those with known structure. At the same time, growing number of proteins with unknown structure versus proteins with known structure becomes one of the main challenges facing the content prediction problem. In the course of this thesis, novel approaches with respect to feature space and prediction method are proposed to improve the accuracy of protein secondary structure content prediction for low homology protein sequences. |
