Estimation of evolutionary distance between distantly related sequences of amino acids, taking account of patterns of amino acid replacement

Abstract

When amino acid sequences are distantly related-for instance, when their identity is <0.30-it is difficult to estimate their evolutionary distance. A method called the 'similarity distance method' (SD method) was developed to obtain maximum-likelihood estimates of evolutionary distance between amino acid sequences, on the basis of a given pattern of amino acid replacement. Computer simulation revealed that, by using the new method, evolutionary distance can be estimated efficiently even when the expected identity between the sequences is as low as 0.14 and the length of the sequences is only 50 amino acid residues. The new method is useful in phylogenetic analyses of distantly related and/or short sequences of amino acids, such as structural or functional domains commonly found in several proteins.