An improved method for estimating sequence divergence between related DNAs from changes in restriction endonuclease cleavage sites.

We have developed a theory to estimate the degree of sequence divergence between related DNAs from the comparison of restriction endonuclease recognition sites. Two major improvements have been made upon a similar method reported by Upholt (1977). First, the most probable value is calculated by the collective use of all available data. This reduces intrinsic statistical error and extends the analyzable range of sequence divergence. Second, all variables are redefined so that they have strict mathematical implications. This corrects a serious error arising from the misinterpretation of the meaning of the fraction of conserved cleavage sites. With this refined method, sequence divergence between rat and mouse mitochondrial DNAs (mtDNAs) was calculated to be about 25% substitutions/nucleotide, which is in good agreement with the DNA-DNA hybridization data obtained by Jakovcic et al. (1975). It was also estimated that the three types of rat mtDNAs differ from one another by 0.3 approximately 1% of total base pairs. These values are 2 approximately 5 times smaller than those obtained with the conventional method.