Fig. 3: Methylome divergence is associated with differential transcriptional activity in Lake Malawi cichlids.
From: Mapping epigenetic divergence in the massive radiation of Lake Malawi cichlid fishes
a Heatmap and unsupervised hierarchical clustering of gene expression values (Z-score) of all differentially expressed genes (DEGs) found among livers of four Lake Malawi cichlid species (Wald tests corrected for multiple testing using false discovery rate FDR < 1%). GO enrichment analysis for three DEG clusters are shown in Supplementary Fig. 9c. b Significant overlap between DEG and differentially expressed regions (DMRs; p < 0.05) linked to a gene (exact hypergeometric test, p = 4.71 × 10−5), highlighting putative functional DMRs (pfDMRs). c Bar plot showing the percentage of pfDMRs localised in either promoters, intergenic regions (0.5−4kbp away from genes), or in gene bodies, with the proportion of TE content for each group. d Heatmap representing significant GO terms for DEGs associated with pfDMRs for each genomic feature. GO categories: BP, Biological Process; MF, Molecular Function. Only GO terms with Benjamini−Hochberg FDR-corrected p-values < 0.05 are shown. Examples of pfDMRs significantly associated with species-specific liver transcriptional changes for the genes thiosulfate:glutathione sulfurtransferase tstd1-like (LOC101468457; q = 6.82 × 10−16) (e), carbonyl reductase [NADPH]-1 cbr1-like (LOC101465189; MZ vs DL, q = 0.002; MZ vs RL, q = 1.18 × 10−7) (f) and perforin-1 prf1-like (LOC101465185; MZ vs DL, q = 3.68 × 10−19; MZ vs RL, q = 0.00034) (g). Liver and muscle methylome profiles in green and purple, respectively (averaged mCG/CG levels [%] in 50 bp bins; n = 3 biological replicates for liver DL, PG, and MZ; n = 2 biological replicates for liver RL, AS, and AC, and muscle DL, RL, and PG). h−j Boxplots showing gene expression values (transcript per million) for the genes in (e−g). in livers (green) and muscle (pink). n = 3 biological replicates for liver DL, MZ, PG; n = 2 biological replicates for liver RL and muscle DL, MZ, PG, and RL. Two-sided q values for Wald tests corrected for multiple testing (Benjamini−Hochberg FDR) are shown in graphs. Box plots indicate median (middle line), 25th, 75th percentile (box), and 5th and 95th percentile (whiskers) as well as outliers (single points). CGI, CpG islands; Repeats, transposons and repetitive regions.
