Extended Data Fig. 3: Most mthfd1-1 DMRs that are fully suppressed by thfs are located in heterochromatic TE-associated chromatin regions while partially suppressed mthfd1-1 DMRs tend to be located in active gene body-associated chromatin regions.

a, Mean mCG, mCHG, and mCHH ratios in 10 kb windows along chromosomes (Chr) 1 to 5, and signed difference relative to wild type (Δ = mutant - WT). b, Principal component analysis (PCA) of whole-genome bisulfite sequencing (WGBS) data from different genotypes. DNA methylation ratios in CG (left), CHG (middle), and CHH (right panel) were calculated in 100 bp bins across the Arabidopsis genome. Each point represents an individual sample; mthfd1-1 is clearly separated from the other groups along the first principal component. c, Heat maps showing the relative frequency distribution of all merged DMRs from pairwise comparisons of m, t, and mt to wild type over chromatin states (CS) 1-36 (left) and standardized frequencies (z-scores) of DMRs across clusters (A1 to A4, corresponding to Fig. 1d) (middle). Cluster A1 overlaps mostly with CS31-36, whereas cluster A2 overlaps mostly with CS5&6. Chromatin state annotations from the Plant Chromatin State Database (PCSD)⁸⁸ are shown at right. d, Distribution of mCG ratios in gene body-methylated (gbM; n = 2,448) and CMT2-targeted (n = 2,190) regions in rosette leaves of mthfd1-1 (m), mthfd1-1 thfs double mutant (mt), thfs (t), and wild-type (WT) plants grown under LD. Values were computed from per-site C/T counts and pooled across two biological replicates for each genotype. Box plots indicate medians (centre lines), IQR (boxes), and 1.5×IQR (whiskers); violin width reflects kernel density; n = regions. Different lowercase letters indicate significant differences between genotypes (pairwise Wilcoxon rank-sum tests, two-sided; BH-adjusted P < 0.01). See Supplementary Table 11 for P values. e, Volcano plot of differential TE expression between mthfd1-1 thfs (mt) and wild type (n = 3). Subfamily names of most significant DETEs are shown. f, Principal component analysis (PCA) of mRNA-seq data. Each point represents an individual sample; mthfd1-1 is clearly separated from the other groups along the first principal component. g, Proportions and absolute numbers of DETEs belonging to different TE families for each Cluster A5 to A7, as shown in Fig. 1e.