Crossassociation: a method of comparing protein sequences

Crossassociation is a computer method of comparing protein sequences. It can help detect amino acid matches, deletions, insertions, and other similarities which would be hard to detect by eye. The method is to "slide" the sequences past each other one step at a time and to count the number of amino acids that match. At each overlap position, the program prints the percentage match and statistical significance measures of the matching. The null hypothesis for significance is the random arrangement of amino acids in the proportions found in the sequences under study. For most protein pairs, the expected proportion of matches is about 1/14. The method includes computation of three overall similarity measures between sequences which should have use in both evolutionary and taxonomic studies. The use of the method has been tested with actual and hypothetical sequences. Problems of recovering evolutionary relationships by this and related methods are discussed.