Letter to the Editor Unequal Base Frequencies and the Estimation of Substitution Rates

Abstract

Recently, Tamura and Nei ( 1993) presented a new method for estimating the number of nucleotide substitutions between two sequences. Their model of substitution allows different rates of puri ne transition, pyrimidine transition, and transversion. Heterogeneity of substitution among sites is accommodated by allowing the rate to vary according to a gamma distribution. The model assumes a stationary process ( i.e., the observed base composition reflects the nucleotide frequencies at equilibrium). In order to maintain the equilibrium composition, substitutions are weighted by the frequency of the mutant base. The motivation for this weighting arises from an analysis of the patterns and relative rates of base substitutions inferred by parsimony from a distance-based tree. We wish to discuss this analysis of substitution patterns, as well as the sensitivity of divergence estimates to the equilibrium nucleotide frequencies which are assumed. Table 1 shows the substitutions inferred by Tamura and Nei from 95 human control region sequences. Surprisingly, this matrix suggests that the base composition of the control region is changing over time. The number of G’s lost by substitution to another nucleotide is 4 1.5, while the number of G’s created by mutation is 68.5. This suggests a net gain of 27 G’s over this phylogeny. A change in base composition seems unlikely for two reasons. First, nucleotide composition is conserved among hominoid mtDNA sequences over a time period considerably greater than that represented by this data (Kondo et al. 1993), Second, the pattern of mtDNA nucleotide composition in primates, and indeed throughout the animal kingdom, is characterized by a low frequency of G on this strand. If the substitution matrix inferred by parsimony is correct, then the base composition of these sequences is not at equilibrium. Evolving according to the inferred matrix, the composition of these sequences would eventually come to equilibrium at 0.26, 0.25, 0.32, and 0.18 for A, T, C,

2 Figures and Tables

Cite this paper

@inproceedings{Perna1998LetterTT, title={Letter to the Editor Unequal Base Frequencies and the Estimation of Substitution Rates}, author={Nicole T. Perna and Thomas D. Kocher-f and Thomas David Kocher}, year={1998} }