Predictions from the regularities of the primary structure of proteins

Statistical analysis of large data sets indicates the existence of short-range regularities in the primary structure of proteins. In this paper the range and measure of these short-range regularities and various prediction methods based on them are discussed. The methods include predictions for domain boundaries of multidomain proteins; sizes of low-energy building blocks used in the calculation of protein structure as an assembly of stable overlapping segments; replaceability of amino acids; cis and trans conformation of proline residues; disulfide-forming Cys residues; surface exposure of amino acids; and tyrosine sulfation sites in proteins.