Whole Genome Bisulfite Sequencing of Arabidopsis rdm1 Mutant

Genomic DNA from Wassilewskija (WS), rdm1-3 and two independent lines expressing RDM1-3xFLAG was extracted using the DNAeasy Plant Mini Kit (Qiagen). DNA were sheared to 300 bp with a Covaris S2 (Covaris). Libraries, including bisulfite conversion, were made with the NuGEN Ultralow Methyl-seq kit (NuGEN) and EpiTect Bisulfite kit (Qiagen) following manufacturer’s instructions. WGBS analysis was performed similar as before^{13 (link)}. In general, raw reads were aligned to TAIR10 genome using BSMAP^{28 (link)} with –v 2 (allowing maximal two mismatches) and –n 1 (aligning to both strands). Reads were discarded if there were more than three consecutive methylated CHH sites within the reads (for 50 bp long read)^{29 (link)}. Methylation levels at each cytosine were calculated as #C/(#C+#T). DMRs between WS and rdm1 were defined using R package DMRcaller^{30 (link)}. To define CHH DMRs, parameters were applied as before^{31 (link)}. Basically, 100 bp bins with more than four cytosines and each cytosines with more than 4 read coverage as well as more than 0.1 difference between rdm1 and WS were used as cutoff. DMRs within 200 bp of each other were merged for further analysis. Cytosine methylation over rdm1 hypo CHH DMR were extracted and plotted in R. WGBS tracks displayed in Fig. 1c correspond to wild-type Col-0 samples as described in Stroud et al.^{31 (link)}.