On the evolution of the base composition of DNA

Abstract

The difference of the DNA base ratios for different organisms and their homogeneity for the same organism are explained by evolution from a common organism, using the following assumptions: (1) Most base pairs in DNA can undergo changes that have no or only an insignificant selective effect such that organisms experiencing many of these changes can survive as well as others experiencing only few changes. (2) For each DNA species one kind of base pair, e.g. G-C, has been altered more frequently than the other one, resulting in a shift of the base ratio. It can then be shown: (1) the number of base pair changes, required to explain the present disparity in the base ratio of different organisms, is about one or more per base pair. (2) The disparity of the base ratio may have been caused either by differences in the relative rates with which A-T versus G-C pairs have been attacked, or by differences in the number of base pair changes which the different DNA species have undergone. (3) The influence of the initial base frequency distribution on the present distribution must have decreased exponentially with increasing number of base pair changes, but it may still affect the present distribution. Eventually, aftermany base pair changes, the variance of the base frequency distribution about the mean base frequency (or base ratio) would become much smaller than the observed variance, if the constraint of biological information on DNA changes were negligible; this final variance depends only on the present mean base frequency even if that continues to change. Thus it is easy to explain the homogeneity of the present base ratios within one organism. (4) If the larger observed variance would still contain a contribution by the initial base distribution one could determine which base pair changes have been preferred during evolution. In the second part it is shown how accurately determined dinucleotide frequencies may reveal which organisms are evolutionarily more related than others; this cannot be concluded from the base ratio alone. This relationship may disclose which detailed types of base pair changes occurred predominantly during DNA evolution; in addition, it may allow one to decide whether DNA evolution moves towards more or toward less randomness of the base sequences.