ABSTRACT
The development of protein structure prediction still has not reached the point when there can be only small improvements in the quality of the results. However, the methods for secondary structure prediction are much closer to this leveling point than the much more complex field of tertiary (3D) structure prediction. This paper presents a novel approach to the simpler of the two problems: the assignment of secondary structure elements to a sequence of amino acid residues. The proposed method offers high quality prediction in 70% of the tested cases. This, and the novel S-motif approach it uses, make the method a viable part of any consensus prediction method that may be developed in the future.