A flexible and accurate genotype imputation method for the next generation of genome-wide association studies

Abstract

Genotype imputation methods are now being widely used in the analysis of genome-wide association studies. Most imputation analyses to date have used the HapMap as a reference dataset, but new reference panels (such as controls genotyped on multiple SNP chips and densely typed samples from the 1,000 Genomes Project) will soon allow a broader range of SNPs to be imputed with higher accuracy, thereby increasing power. We describe a genotype imputation method (IMPUTE version 2) that is designed to address the challenges presented by these new datasets. The main innovation of our approach is a flexible modelling framework that increases accuracy and combines information across multiple reference panels while remaining computationally feasible. We find that IMPUTE v2 attains higher accuracy than other methods when the HapMap provides the sole reference panel, but that the size of the panel constrains the improvements that can be made. We also find that imputation accuracy can be greatly enhanced by expanding the reference panel to contain thousands of chromosomes and that IMPUTE v2 outperforms other methods in this setting at both rare and common SNPs, with overall error rates that are 15%-20% lower than those of the closest competing method. One particularly challenging aspect of next-generation association studies is to integrate information across multiple reference panels genotyped on different sets of SNPs; we show that our approach to this problem has practical advantages over other suggested solutions.

Figure 1. Schematic drawing of imputation Scenario A.

In this drawing, haplotypes are represented as horizontal boxes containing 0's and 1's (for alternate SNP alleles), and unphased genotypes are represented as rows of 0's, 1's, 2's, and ?'s (where ‘1’ is the heterozygous state and ‘?’ denotes a missing genotype). The SNPs (columns) in the dataset can be partitioned into two disjoint sets: a set T (blue) that is genotyped in all individuals and a set U (green) that is genotyped only in the haploid reference panel. The goal of imputation in this scenario is to estimate the genotypes of SNPs in set U in the study sample.

Figure 2. Schematic drawing of imputation Scenario B.

In this drawing, haplotypes are represented as horizontal boxes containing 0's and 1's (for alternate SNP alleles), and unphased genotypes are represented as rows of 0's, 1's, 2's, and ?'s (where ‘1’ is the heterozygous state and ‘?’ denotes a missing genotype). The SNPs (columns) in the dataset can be partitioned into three disjoint sets: a set T (blue) that is genotyped in all individuals, a set U₂ (yellow) that is genotyped in both the haploid and diploid reference panels but not the study sample, and a set U₁ (green) that is genotyped only in the haploid reference panel. The goal of imputation in this scenario is to estimate the genotypes of SNPs in set U₂ in the study sample and SNPs in the set U₁ in both the study sample and, if desired, the diploid reference panel.

Figure 3. Percentage discordance versus percentage missing genotypes for Scenario A dataset.

(A) Full range of results, corresponding to calling thresholds from 0.33 to 0.99. (B) Magnified results for calling thresholds near 0.99. (C) Magnified results for calling thresholds near 0.33.

Figure 4. Percentage discordance versus percentage missing genotypes for restricted Scenario B dataset.

(A) Results for masked Illumina genotypes imputed from Affymetrix genotypes in the study sample. (B) Results for masked Affymetrix genotypes imputed from Illumina genotypes in the study sample. (C) Results for masked Illumina genotypes (SNPs with MAF<5% only) imputed from Affymetrix genotypes in the study sample. (D) Results for masked Affymetrix genotypes (SNPs with MAF<5% only) imputed from Illumina genotypes in the study sample.

Figure 5. Percentage discordance versus percentage missing genotypes for full Scenario B dataset.

(A) Results for masked Illumina genotypes imputed from Affymetrix genotypes in the study sample. (B) Results for masked Affymetrix genotypes imputed from Illumina genotypes in the study sample. Solid lines were obtained from the restricted Scenario B dataset (Figure 4) and are shown for reference; dashed lines were obtained from the full Scenario B dataset.