a, H2A.Z (green) and DNA methylation (blue) profiles of Arabidopsis chromosome 2. Each vertical bar represents the log₂ signal ratio of the test sample signal divided by the input control signal (log₂ (test/input)). The black circles denote the position of the centromeric sequence gap. b–c, More detailed views of a euchromatic (positions 547,000 – 587,000, b) and a heterochromatic (4,407,000 – 4,463,000, c) genomic region. DNA methylation from aerial tissues and roots is shown in blue; HA2.Z profiles obtained from two independent BLRP-H2A.Z transgenic lines and via immunoprecipitation of endogenous H2A.Z are shown in green. Genes and transposons on the top and the bottom strands are shown above and below the line, respectively. 5′ peaks of H2A.Z in genes are emphasized by boxes in b. Unmethylated transposons with relatively high levels of H2A.Z are emphasized by boxes in c.

a–b, All TAIR 7 annotated sequences (31,762) were aligned at the 5′ end and stacked from the top of chromosome 1 to the bottom of chromosome 5. BLRP-H2A.Z is displayed as a heat map in a; root DNA methylation is displayed in b. Note the high degree of anticorrelation between H2A.Z and methylation. c–d, Unmethylated transposable elements (listed in Supplementary Table 3). BLRP-H2A.Z is displayed as a heat map in c; root DNA methylation is displayed in d. e, All TAIR 7 annotated sequences were k-means clustered (k=3) based on BLRP-H2A.Z patterns, and displayed as a heat map. For comparison, root DNA methylation of the same sequences is shown as a heat map in f.

a–c, Wild type (WT) root DNA methylation (dark blue), met1-6 root DNA methylation (purple), WT H2A.Z (antibody, green), WT H2A.Z profile subtracted from the met1-6 H2A.Z profile (two sets of independent paired experiments, light blue), and met1-6/WT transcription (red) for FWA in a, copia-like transposable element At5g13205 that loses methylation and gains H2A.Z in met1-6 in b, and F-box gene At1g22000 that is hypermethylated and loses H2A.Z in met1-6 in c. The 5′ region of FWA methylated in WT is emphasized by boxes in a. d, All TAIR 7 annotated sequences were aligned at the 5′ end and stacked from the top of chromosome 1 to the bottom of chromosome 5. The WT H2A.Z pattern subtracted from the met1-6 H2A.Z pattern is displayed as a heat map. The same data after k-means clustering (k=3) are shown in e. For comparison, root DNA methylation of sequences arranged as in e is shown as a heat map in f. g, WT methylation levels (left) and met1-6 methylation levels (right) for probes representing a significant decrease of H2A.Z in met1-6 (Supplementary Fig. 12). The histogram is cumulative for three independent methylation datasets. Grey histograms in the background show the signal distribution for all probes. h, Kernel density plot, which has the effect of tracing the frequency distribution, of all probes in the dataset displayed in d (black trace), transposable elements upregulated in met1-6 (red trace), and transposable elements not upregulated in met1-6 (blue trace).

a, All TAIR 7 annotated sequences were aligned at the 5′ end and stacked from the top of chromosome 1 to the bottom of chromosome 5. The WT methylation pattern subtracted from the pie1 methylation pattern is displayed as a heat map. b, Bisulfite sequencing results for five loci. We sequenced 12 clones from each genotype, except for At1g69850 (10 clones in pie1) and At4g38190 (11 clones in pie1). c, PCR products from bisulfite-converted genomic DNA were digested with TaqI, which recognizes TCGA and will cut only if the C is unconverted (and therefore methylated). L = 100 bp ladder, Unc = uncut PCR product, TaqI = PCR product digested with TaqI. Note the greater digestion, which represents greater methylation, in pie1 compared to WT. d, All genes were aligned at the 5′ end and average scores for each 100-bp interval are plotted from 2 kb away from the gene (negative numbers) to 3 kb into the gene (positive numbers). The data were smoothed with a 5-point sliding window. The dashed line represents the point of alignment. e, Genes were grouped into percentiles based on transcription levels. The red line traces the number of genes hypermethylated in pie1 within each percentile (left Y-axis). The black line traces DNA methylation enrichment (all genes) and the green line traces H2A.Z enrichment in unmethylated genes (right Y-axis). The data were smoothed with a 10-point sliding window. The scale of the right Y-axis was set to start at zero to enable comparison between methylation and H2A.Z.