Genomewide association analysis in diverse inbred mice: power and population structure

Abstract

The discovery of quantitative trait loci (QTL) in model organisms has relied heavily on the ability to perform controlled breeding to generate genotypic and phenotypic diversity. Recently, we and others have demonstrated the use of an existing set of diverse inbred mice (referred to here as the mouse diversity panel, MDP) as a QTL mapping population. The use of the MDP population has many advantages relative to traditional F(2) mapping populations, including increased phenotypic diversity, a higher recombination frequency, and the ability to collect genotype and phenotype data in community databases. However, these methods are complicated by population structure inherent in the MDP and the lack of an analytical framework to assess statistical power. To address these issues, we measured gene expression levels in hypothalamus across the MDP. We then mapped these phenotypes as quantitative traits with our association algorithm, resulting in a large set of expression QTL (eQTL). We utilized these eQTL, and specifically cis-eQTL, to develop a novel nonparametric method for association analysis in structured populations like the MDP. These eQTL data confirmed that the MDP is a suitable mapping population for QTL discovery and that eQTL results can serve as a gold standard for relative measures of statistical power.

Figure 1.—

eQTL plots for hypothalamus. eQTL plots were generated using three haplotype association mapping (HAM) methods: (A) parametric analysis, (B) nonparametric analysis, and (C) weighted nonparametric analysis. In all plots, the x-axis represents the genomic SNP axis and the y-axis represents the genomic probe set axis. Each spot represents an association between the expression of a gene and the strain distribution pattern at a SNP location. Alternating colors indicate chromosome boundaries on the x-axis. In each plot, the top 10,000 eQTL associations are shown.

Figure 2.—

Comparison of parametric and nonparametric HAM methods. The chart displays the cis-eQTL enrichment as a function of eQTL rank for each of three different HAM variants.

Figure 3.—

Clustering of gene expression data. The clustering dendrogram displays the relationship of global gene expression patterns between strains. Coloring of the strain names reflects clusters derived from clustering of genotype data. The clear relationship between global gene expression patterns and genomewide genetic similarity underscores the need to account for population structure in association analyses.

Figure 4.—

Weight exponent analysis. To optimize the choice of the weight exponent, we calculated cis-eQTL enrichment using a range of weight powers and eQTL ranks. We chose a weight exponent of three for all subsequent studies.

Figure 5.—

Two-factor association analysis. Strain distribution pattern and sex were treated as independent factors in a two-factor HAM analysis. An eQTL analysis was performed using hypothalamus over 12 strains. For comparison, the cis-eQTL enrichment ratios are shown for the corresponding one-factor analyses.