Variation in histone configurations correlates with gene expression across nine inbred strains of mice

Abstract

The Diversity Outbred (DO) mice and their inbred founders are widely used models of human disease. However, although the genetic diversity of these mice has been well documented, their epigenetic diversity has not. Epigenetic modifications, such as histone modifications and DNA methylation, are important regulators of gene expression and, as such, are a critical mechanistic link between genotype and phenotype. Therefore, creating a map of epigenetic modifications in the DO mice and their founders is an important step toward understanding mechanisms of gene regulation and the link to disease in this widely used resource. To this end, we performed a strain survey of epigenetic modifications in hepatocytes of the DO founders. We surveyed four histone modifications (H3K4me1, H3K4me3, H3K27me3, and H3K27ac), as well as DNA methylation. We used ChromHMM to identify 14 chromatin states, each of which represents a distinct combination of the four histone modifications. We found that the epigenetic landscape is highly variable across the DO founders and is associated with variation in gene expression across strains. We found that epigenetic state imputed into a population of DO mice recapitulated the association with gene expression seen in the founders, suggesting that both histone modifications and DNA methylation are highly heritable mechanisms of gene expression regulation. We illustrate how DO gene expression can be aligned with inbred epigenetic states to identify putative cis-regulatory regions. Finally, we provide a data resource that documents strain-specific variation in the chromatin state and DNA methylation in hepatocytes across nine widely used strains of laboratory mice.

Figure 1.

The first two principal components of each genomic feature across nine inbred mouse strains. In all panels, each point represents an individual mouse, and strain is indicated by color as shown in the legend at the bottom of the figure. Three individuals per strain are shown. Each panel is labeled with the data used to generate the PC plot. (A) Hepatocyte transcriptome: all transcripts expressed in isolated hepatocytes. (B) DNA methylation: the percentage of methylation at all CpG sites shared across all individuals. (C–F) Histone modifications: the peak heights of the indicated histone modification for positions aligned across strains.

Figure 2.

Overview of chromatin state composition, genomic distribution, and association with expression. (A) Emission probabilities for each histone modification in each chromatin state. Blue indicates the absence of the histone modification, and red indicates the presence of the modification. (B) The distribution of each state around functional elements in the genome. Red indicates that the state is enriched near the annotated functional element. Blue indicates that the state is depleted near the annotated functional element. Rows were scaled to run between zero and one for ease of visualization. Abbreviations are as follows: (Enh.) Enhancer, (Tsd) distal to the transcription start site, (Tsp) proximal to the transcription start site, (Hetero.) heterochromatin, (FR) flanking region. (C) The association between chromatin state variation and gene expression. Bars are colored based on the size and direction of the state's association with expression. Red/blue bars show the associations of chromatin state with gene expression across strains. Blue-gray bars show the associations of chromatin state with gene expression across genes. (D) Plausible annotations for each state based on genomic enrichments and association with gene expression. The numbers in parentheses indicate the percentage of the genome that was assigned to each state. (repress.) Repressor.

Figure 3.

Overview of strain-specific CpG sites. (A) Boxes show the proportion of strain-specific CpG sites that is present in each strain. Boxes are colored by official strain colors for ease of visualization. Short names for strains are indicated below each box. (B) The log₁₀ (fold enrichment) of CpG sites shared across all strains (green) and those that are strain specific (purple). (C) A comparison of enrichments between CpG sites that are shared across all strains and those that are strain specific. Bars above one show where strain-specific CpGs were more enriched than shared CpGs. Bars below one indicate where strain-specific CpGs were less enriched than shared CpGs. The vertical line marks where shared and strain-specific CpGs were equally enriched. Abbreviations are as follows: (FR) flanking region, (Tsp) transcription start site proximal, (Tsd) transcription start site distal, (Hetero.) heterochromatin, (Enh.) enhancer.

Figure 4.

Relative abundance of chromatin states and methylated DNA. (A) Each panel shows the abundance of a single chromatin state relative to gene transcription start site (TSS) and transcription end site (TES). The y-axis in each panel is the percentage of genes containing the state. Each panel has an independent y-axis to better show the shape of each curve. The x-axis is the relative gene position. The TSS and TES are marked as vertical gray dotted lines. (B) The same data shown in panel A, but with all states overlaid onto a single set of axes to show the relative abundance of the states. (C) The density of CpG sites relative to the gene body. The y-axis shows the inverse inter-CpG distance in base pairs. The density is highest near the TSS. CpG sites are less dense within the gene body and in the intergenic space. (D) Percentage of methylation relative to the gene body. The y-axis shows the median percentage of methylation at CpG sites, and the x-axis shows relative gene position. CpG sites near the TSS are unmethylated relative to intragenic sites and to sites just upstream of and downstream from the gene bodies. In both C and D, standard error is shown as a blue envelope around the mean; however, the standard error is so small that it is not visible in the figure.

Figure 5.

Associations of chromatin states with gene expression. Each column shows the association of each chromatin state with gene expression in a different experimental context as labeled. Effects shown are β coefficients from Equation 1. The y-axes vary across each row to emphasize the shape of each effect, so y-axis labels indicate only positive and negative effects. Colored areas show the 95% confidence interval around each estimate. The final column shows the annotation of each state for comparison with its association with gene expression. All x-axes show the relative position along the gene body, running from just upstream of the TSS to just downstream from the TES. Vertical gray dotted lines mark the TSS and TES in all panels.

Figure 6.

Association of DNA methylation with gene expression (A) across genes in hepatocytes, (B) across inbred strains, and (C) in the DO population. The dark gray line shows the estimated effect of percentage of DNA methylation on gene expression. The x-axis is normalized position along the gene body running from the TSS to the TES, marked with vertical gray dotted lines. The horizontal solid black line indicates an association of zero. The shaded gray area shows a 95% confidence interval around the model fit.

Figure 7.

Comparison of the variance explained in DO gene expression by four genomic features: haplotype (Hap.), chromatin state (Chrom.), local SNP genotype (SNP), and local imputed DNA methylation status (DM). (A) Distributions of gene expression variance explained by each feature. (B) Direct comparisons of variance explained by local haplotype and each of the other genomic features. Blue lines show y = x. Each point is a single transcript.

Figure 8.

Example of epigenetic states and imputation results for a single gene, Pkd2. The legend for each panel is displayed to its right. (A) The variance in DO gene expression explained at each position along the gene body by each of the imputed genomic features: SNPs, red X's; chromatin state, blue plus signs; and percentage of methylation, green circles. The horizontal dotted line shows the maximum variance explained by any individual haplotype (in this case, CAST). For reference, the arrow below this panel runs from the TSS of Pkd2 (vertical bar) to the TES (arrowhead) and shows the direction of transcription. The gray arrows at the top indicate two regions of interest where chromatin state explains height amounts of variance in gene expression. (B) The chromatin states assigned to each 200-bp window in this gene for each inbred mouse strain. States are colored by their association with gene expression in the inbred mice. Red indicates a positive association with gene expression, and blue indicates a negative association. Each row shows the chromatin states for a single inbred strain, which is indicated by the label on the left. (C) SNPs along the gene body for each inbred strain. The reference genotype is shown in gray. SNPs are colored by genotype as shown in the legend. (D) Percentage of DNA methylation for each inbred strain along the Pkd2 gene body. Percentages are binned into 0% (blue), 50% (yellow), and 100% (red). (E) Association of haplotype with expression of Pkd2 in the DO. Haplotype effects are colored by the allele from which they were derived. (F) Pkd2 expression levels across inbred mouse strains. For ease of comparison, panels B through F are shown in the same order as the haplotype effects.