A neutrality test for detecting selection on DNA methylation using single methylation polymorphism frequency spectrum.

Inheritable epigenetic mutations (epimutations) can contribute to transmittable phenotypic variation. Thus, epimutations can be subject to natural selection and impact the fitness and evolution of organisms. Based on the framework of the modified Tajima's D test for DNA mutations, we developed a neutrality test with the statistic "D(m)" to detect selection forces on DNA methylation mutations using single methylation polymorphisms. With computer simulation and empirical data analysis, we compared the D(m) test with the original and modified Tajima's D tests and demonstrated that the D(m) test is suitable for detecting selection on epimutations and outperforms original/modified Tajima's D tests. Due to the higher resetting rate of epimutations, the interpretation of D(m) on epimutations and Tajima's D test on DNA mutations could be different in inferring natural selection. Analyses using simulated and empirical genome-wide polymorphism data suggested that genes under genetic and epigenetic selections behaved differently. We applied the D(m) test to recently originated Arabidopsis and human genes, and showed that newly evolved genes contain higher level of rare epialleles, suggesting that epimutation may play a role in origination and evolution of genes and genomes. Overall, we demonstrate the utility of the D(m) test to detect whether the loci are under selection regarding DNA methylation. Our analytical metrics and methodology could contribute to our understanding of evolutionary processes of genes and genomes in the field of epigenetics. The Perl script for the "D(m)" test is available at http://fanlab.wayne.edu/ (last accessed December 18, 2014).