The Application Of Hidden Markov Models to Protein Secondary Structure Prediction

The functional properties of proteins depend upon their 3D structures, therefore, it is advantageous to deduce the 3D structure of a protein from its amino acid sequence. This is a difficult task because there are 20 different amino acids that can be combined into “many more different proteins than there are atoms in the known universe” [2]. De novo prediction methods often involve a first step of protein secondary structure prediction. The focus of this thesis is on the application of Hidden Markov Models to the prediction of the three classes of secondary structures: helices, strands and coils. A 3-State and 21-State model have been constructed to illustrate the potential, and difficulties these models have in the area of secondary structure prediction.