Extended Data Fig. 2: Analysis of global mCA levels and gene dysregulation in Mecp2 KO mice.
a, Gene expression of MeCP2-repressed (MR) and MeCP2-activated (MA) genes in WT and Mecp2 KO L4, L5, SST, and PV subclasses. Central dots represent medians. b, Number of significantly dysregulated genes identified in each subclass plotted vs average mCA level in that subclass, using previously published single-cell methylation data of the 8-week-old mouse frontal cortex.8 c, Mean fold-change of long (greater than 100 kb), highly methylated (top decile of mCA) genes (left) and core MeCP2-repressed genes (right) vs. average mCA level for all genes in each subclass, using the same methylation data in b. Data are presented as mean values +/- SEM. d, Log2 fold-change (Mecp2 KO/WT) of expression for all genes in subclasses. The x-axis is normalized counts of genes from DESeq2. e, Smooth line plots of mCA levels vs log2 fold-change in gene expression (Mecp2 KO/WT) in each neuronal subclass, separated by genes longer than (red) or shorter than (gray) 100 kb. f, Gene body mCA/CA of core MeCP2-repressed genes in each subclass compared to expression-resampled control genes. ****p < 0.0001 two-sided Wilcoxon rank-sum test. g, Left: depiction of the three brain regions analyzed in analysis of global DNA methylation and gene dysregulation in MeCP2 mutants. Right: global mCA/CA of WT and Mecp2 KO cells in each brain region. ns=not significant (p > 0.05, Wilcoxon rank-sum test). h, Mean fold-change in mRNA expression for long (greater than 100 kb), high mCA (top decile of whole gene mCA) genes in the Mecp2 KO vs. WT in brain tissues. The x-axis represents global mCA/CA for each tissue averaged across bioreplicates. Central dot is the mean fold-change of gene expression between Mecp2 KO and WT, and the y-axis error bars represent the standard error of that mean. The x-axis error bars represent the standard error of the mean of global mCA levels of a tissue type averaged over all replicates. i, Left: mCG/CG of genes in each subclass. Grand mean genic mCG/CG across replicates and standard error are shown. Middle: number of significantly dysregulated genes identified in each subclass plotted vs average mCG/CG in that subclass. Right: expression fold-change of core MeCP2-repressed genes vs. average mCG/CG for all genes in each subclass. Mean and standard error are shown. j, Global mCA (left) and mCG (right) levels of WT and Mecp2 KO neuronal subclasses. ns=not significant (p > 0.05 Wilcoxon rank-sum test). k, WT vs Mecp2 KO gene-body mCA/CA in neuronal subclasses. Genes with a minimum length of 5 kb located in chromosomes 1-19 and X were used. ρ = Spearman’s correlation coefficient. For a, d-f, and i-k, data are from cerebral cortex tissue from 8–10-week-old MeCP2 WT and KO mice. n = 4 bioreplicates per genotype for RNA-seq, n = 2-3 bioreplicates per genotype for whole-genome bisulfite sequencing. WT and Mecp2 KO replicates were averaged for their methylation values except where stated otherwise. For e and f, n = 2 bioreplicates for WT and Mecp2 KO cerebellum, striatum, and hypothalamus whole-genome bisulfite sequencing, n = 5 bioreplicates per genotype of microarray data from brain tissue25,26. Boxplots represent data as described in Fig. 2. WT and Mecp2 KO replicates were averaged for INTACT methylation calculations in b, c, e, and f.
