Skip to main content
Fig. 4 | Genome Biology

Fig. 4

From: Trans effects of chromosome aneuploidies on DNA methylation patterns in human Down syndrome and mouse models

Fig. 4

Parallels between DS-DM in humans and DM in chromosome-engineered mouse models of DS. a In the mouse WGBS data each 1-kb genomic window corresponding to all evaluable regions queried by the human 450K arrays was evaluated for DM using stringent criteria employing Fisher exact tests at the individual CpG level (see "Materials and methods"). Using these sets of DM loci, we asked whether the human DS-DM genes identified in our comparisons between DS and control human brain cells (neurons and/or glia) were statistically enriched among the mouse DM genes. Although the number of DM genes decreases with increasing stringency, there is a progressive relative enrichment for human DS-DM gene orthologues in the mouse DM gene sets as the stringency requirements for DM in the mouse data are increased. The p values for enrichment using the combined Dp(10)1Yey and Dp(16)1Yey data at the indicated stringency levels are from hypergeometric tests. Highly significant enrichment for human DS-DM genes was also seen separately in both the Dp(10) and Dp(16) DM gene sets (respectively, p = 1.5 × 10−27 and 2.5 × 10−27 at 0.05 stringency level and p = 7.1 × 10−9 and p = 3.2 × 10−5 at the 0.005 stringency level). b The Venn diagram on the left shows that a substantial group of human DS-DM genes (31 %) show DM at orthologous genomic locations in the mouse models, using a Fisher exact test p value of 0.05 as the criterion for DM of individual CpGs in the mouse data, and that there is only partial overlap between the DM gene sets found in Dp(16) and Dp(10). The Venn diagram on the right shows the genes that overlap when a more stringent p-value is applied to the WGBS data. c Stacked bar graphs showing the relative frequencies of gains (dark grey) and losses (light grey) of methylation in the two mouse models, compared with wt, for CpGs corresponding to the regions covered by the human 450K arrays. Paralleling the situation in human brains, both mouse models show preferential gains of methylation (above the dashed line), with this effect being stronger in Dp(10) than in Dp(16)

Back to article page