Doublet frequency analysis of bacterial DNAs

Summary Nearest neighbour base frequency analyses of the DNAs of fifteen bacteria and two blue-green algae are reported. When expressed in terms of deviations from random expectation, the frequencies can be placed in four distinct groups sharing similarities not dependent on the G + C contents of the DNAs. The majority of the groupings found are in agreement with those of conventional taxonomy but several interesting discrepancies are shown to exist, some of which confirm other recent molecular evidence. The frequencies for the algal DNAs closely resemble those of the largest group of bacteria. The results are considered in relation to possible evolutionary pressures on polypeptide-specifying DNA and inferences are made about the relative usage of alternative codons in different species.Member of the scientific staff of the Medical Research Council Virology Unit.