Genome-wide and fine-resolution association analysis of malaria in West Africa

Abstract

We report a genome-wide association (GWA) study of severe malaria in The Gambia. The initial GWA scan included 2,500 children genotyped on the Affymetrix 500K GeneChip, and a replication study included 3,400 children. We used this to examine the performance of GWA methods in Africa. We found considerable population stratification, and also that signals of association at known malaria resistance loci were greatly attenuated owing to weak linkage disequilibrium (LD). To investigate possible solutions to the problem of low LD, we focused on the HbS locus, sequencing this region of the genome in 62 Gambian individuals and then using these data to conduct multipoint imputation in the GWA samples. This increased the signal of association, from P = 4 × 10(-7) to P = 4 × 10(-14), with the peak of the signal located precisely at the HbS causal variant. Our findings provide proof of principle that fine-resolution multipoint imputation, based on population-specific sequencing data, can substantially boost authentic GWA signals and enable fine mapping of causal variants in African populations.

Figure 1

Principal components analysis of population structure within The Gambia. Plots of the first three principal components from EIGENSTRAT using 100,715 SNPs selected to minimize intermarker LD. Each solid circle represents an individual, and the color is assigned according to self-reported ethnicity. (a) Plot of the first two principal components for all Gambian samples. (b) Plot of the second and third principal component for all Gambian samples.

Figure 2

Principal components analysis of population structure for the Gambian study sample in relation to HapMap reference panels. Plots of the first two principal components from EIGENSTRAT using 100,715 SNPs selected to minimize intermarker LD. Each solid circle represents an individual. (a) Plot of the first two principal components for HapMap and Gambian samples. (b) Plot of the first two principal components for HapMap YRI and Gambian samples.

Figure 3

Quantile-quantile plots of association test statistic. (a–c) Quantile-quantile plots of the trend test statistic for the unstratified analysis, which uses all 958 cases and 1,382 controls (a); the ethnic-stratified analysis, which tests 854 cases, 1,195 controls from the four major ethnic groups (b); and the PCA analysis, which corrects for the first three principal components from EIGENSTRAT and uses all 958 cases and 1,382 controls (c). The shaded region in gray represents the lower and upper 95% probability bounds for the expected quantiles.

Figure 4

Genome-wide signals of association with severe malaria. Plot of the −log₁₀ P values for the trend test correcting for the first three principal components from EIGENSTRAT. Each point represents a SNP from the 402,814 remaining after quality control filters were applied. Different bands of blue are used to differentiate SNPs on consecutive autosomal chromosomes. SNPs with P values less than 10⁻⁴ are represented by red points.

Figure 5

Association signal at the HBB locus. The top panel shows the association signals across a 1-Mb region on chromosome 11 centering on rs334, with the vertical axis representing the −log₁₀ P values from the Armitage trend test. Points in black represent SNPs that are found on the Affymetrix array, and points in red represent SNPs imputed with the resequenced Gambian reference panel. The dashed lines in red indicate the start and end of the sequenced region. The bottom panels focus on the 110-kb sequenced region, together with a map of the recombination rates and genes found in the region. Recombination rates and genes were extracted from the HapMap Genome Browser.

Figure 6

Extent of LD surrounding HbS. Each point show r² (top panel) and D' (bottom panel) between the HbS SNP (rs334) and SNPs in the Gambian reference panel. The shaded pink region indicates the boundaries of the resequenced region. The dashed vertical line indicates the position of rs334.