THE IMPORTANCE OF BASE COMPOSITION IN THE CORRECT ASSESSMENT OF GENETIC-DISTANCE

Through a simple computer simulation, we provide clear evidence that t he base composition of nucleotide sequences should be taken into accou nt to obtain accurate estimates of the genetic distances between homol ogous nucleotide sequences. We tested several stochastic models and de monstrated clearly that the inferred estimates of the number of substi tutions are strongly dependent on the a priori conditions superimposed in the model, i.e., equiprobability in base composition and/or transi tion transversion ratio. In general, the more a priori assumptions the re are, the more inaccurate the results. Whereas, rather accurate esti mates are obtained with stochastic methods, which take into account ba se composition and do not superimpose unverified a priori conditions.