DNA extraction, RAD-seq and Reduced-Representation Bisulfite Sequencing
We used QIAGEN DNeasy blood and tissue kits to extract DNA from blood samples, following the provided instructions for nucleated blood samples. DNA was quantified using a NanoDrop ND8000 spectrophotometer and a Qubit 2.0 fluorometer with the DNA HS assay kit (Life Technologies). DNA quality was examined on agarose gels. We then performed RAD-sequencing and RRBS-sequencing using standard protocols. For RAD-sequencing (restriction-site-associated DNA sequencing, Baird et al., 2008), the library preparation was done by the Montpellier GenomiX (MGX) platform (CNRS, Montpellier), using the enzyme SbfI. Each individual was identified using a unique six nucleotides tag, individuals were randomly multiplexed in equimolar proportions by libraries of 37 individuals. Each library was sequenced on a lane of an Illumina HiSeq 2500. Paired-end sequencing was used to produce 150 bp reads. This generated an average of 4.9M reads per individual. The DNA of the 60 individuals were processed twice to test for reliability of the genotyping process. The RRBS-sequencing started with DNA digestion using MspI restriction enzymes, which cuts CCGG sites and targets regions that are CG rich, permitting to have a high proportion sequences in promoter regions. Individuals were randomly multiplexed in equimolar proportions by libraries of 10 individuals. Bisulfite treatment converted unmethylated cytosines into uracil, then converted to thymine after PCR amplification. Each library was then sequenced on a lane of an Illumina HiSeq 2500. Paired-end sequencing was used to produce 50 bp reads. This generated an average of 19.3M reads per individual.