Sequence-structure alignment using a statistical analysis of core models and dynamic programming

The expanding availability of protein data enforces the application of empirical methods necessary to recognize protein structures. In this paper a sequence-structure alignment method is described and applied to various Ubiquitin-like folded Ras-binding domains. On the basis of two probability functions that evaluate similarities between the occurrence of amino-acids in the primary and secondary protein structure, different versions of simple scoring functions are proposed. The application of the program ’PLACER’ that uses a dynamic programming approach enables the search for an optimal sequence-structure alignment and the prediction of the secondary structure.