Extended Data Fig. 6: GbM quantitatively affects gene expression.

(a and b) Expression in met1 compared to WT of Bonferroni (α = 0.05) eQTL^teM genes retained (R) after accounting for genetic variation in seedlings across 16 accessions⁷⁵ (a) and different tissues of Col-0 (leaf, seedling and inflorescence^54,77; b). (c-e) Expression in met1 compared to Col-0 of Bonferroni (α = 0.05) eQTL^gbM genes retained (R) after accounting for genetic variation in either leaf (c), inflorescence (d) or seedling (e). (f and g) Expression in met1 compared to WT seedlings of FDR 0.01 (f) and FDR 0.05 (g) eQTL^gbM genes across 16 accessions. (h and i) Expression in met1 compared to WT of FDR 0.01 (h) and FDR 0.05 (i) eQTL^gbM genes in different tissues of Col-0 (leaf, seedling and inflorescence). Center lines within box plots represent sample medians, plus signs correspond to means and are noted below the plots. Box limits indicate the 25th and 75th percentiles, whiskers extend to 1.5 times the interquartile range. Numbers of genes within each group are noted above the plots. P, two-tailed Student’s t-test between indicated group and NA (non-associated) genes. (j-m) Relationship between gbM levels of retained +eQTL and –eQTL genes using the FDR 0.01 (j and l) or FDR 0.05 (k and m) groups in WT plants across 16 accessions⁷⁵ and log₂ fold expression change in met1 compared to WT of that accession. Genes with mCG >5% and/or non-CG methylation >1% in the putative promoter (2 kb upstream or up to the nearest upstream gene, whichever is shorter) were excluded in l and m. Genes were grouped by gbM levels. R and P values correspond to Pearson’s correlation.