The expected equilibrium of the CpG dinucleotide in vertebrate genomes under a mutation model.

The CpG dinucleotide is present at approximately 20% of its expected frequency in vertebrate genomes, a deficiency thought due to a high mutation rate from the methylated form of CpG to TpG and CpA. We examine the hypothesis that the 20% frequency represents an equilibrium between rate of creation of new CpGs and accelerated rate of CpG loss from methylation. Using this model, we calculate the expected reduction in the equilibrium frequency of the CpG dinucleotide and find that the observed CpG deficiency can be explained by mutation from methylated CpG to TpG/CpA at approximately 12 times the normal transition rate, the exact rate depending on the ratio of transitions to transversions. The observed rate of CpG dinucleotide loss in a human alpha-globin nonprocessed pseudogene, psi alpha 1, and the apparent replenishment of the CpG pool in this sequence by new mutations, agree with the above parameters. These calculations indicate that it would take 25 million years or less, a small fraction of the time for vertebrate evolution, for CpG frequency to be reduced from undepleted levels to the current depleted levels. IMAGES: