Abstract
The extent to which epigenetic variation affects complex traits in natural populations is not
known. We addressed this question using transcriptome and DNA methylation data from a
sample of 135 sequenced A. thaliana accessions. Across individuals, expression was significantly
associated with cis-methylation for hundreds of genes, and many of these associations
remained significant after taking SNP effects into account. The pattern of correlations
differed markedly between gene body methylation and transposable element methylation.
The former was usually positively correlated with expression, and the latter usually negatively
correlated, although exceptions were found in both cases. Finally, we developed
graphical models of causality that adapt to a sample with heavy population structure, and
used them to show that while methylation appears to affect gene expression more often
than expression affects methylation, there is also strong support for both being independently
controlled. In conclusion, although we find clear evidence for epigenetic regulation,
both the number of loci affected and the magnitude of the effects appear to be small compared
to the effect of SNPs.
…
July 11, 2016,
978-3-7001-8040-1
18 Seiten,
Research Article