Genome-wide association studies: progress and potential for drug discovery and development

Abstract

Although genetic studies have been critically important for the identification of therapeutic targets in Mendelian disorders, genetic approaches aiming to identify targets for common, complex diseases have traditionally had much more limited success. However, during the past year, a novel genetic approach - genome-wide association (GWA) - has demonstrated its potential to identify common genetic variants associated with complex diseases such as diabetes, inflammatory bowel disease and cancer. Here, we highlight some of these recent successes, and discuss the potential for GWA studies to identify novel therapeutic targets and genetic biomarkers that will be useful for drug discovery, patient selection and stratification in common diseases.

Figure 1. Overview of the general design and workflow of a genome-wide association (GWA) study

The discovery phase entails genotyping many case and control DNA samples and evaluation for significant associations. The replication phase involves fine mapping of association signals and independent confirmation in a second cohort. Biological validation is important for translation of GWA findings into diagnostic or therapeutic discoveries.

Figure 2. Schematic view of genetic linkage, GWA results, fine mapping and linkage disequilibrium structure in a region of chromosome 8q24.21 that demonstrates an association of rs1447295 and rs16901979 with prostate cancer

a | Previously reported genetic linkage scan results for chromosome 8, centiMorgans (cM) 100–170 (that is, 8q) from 871 Icelandic individuals with prostate cancer in 323 extended families. A quantitative trait locus (QTL) for prostate cancer susceptibility with log of the odds (lod) score of ~2 is shown. The interval between the two dashed horizontal lines corresponds to a previously reported admixture signal that is associated with prostate cancer. b | Genome-wide association (GWA) results for 1,660 single nucleotide polymorphisms (SNPs) mapping to chromosome 8 Mb 125–135 in 1,453 Icelandic individuals with prostate cancer and 3,064 controls. Association testing P values smaller than 0.1, corrected for relatedness and population stratification, are shown for single SNPs (blue circles), two SNPs (red circles) and linkage disequilibrium (LD)-block haplotypes (green circles). Four SNPs (including rs1447295) and three haplotype blocks (including Hap C, defined by 14 SNPs) show significant association signals (P <1.58 × 10⁻⁷). Single SNP association two-sided P values were derived using Fisher’s exact test and were unadjusted for multiple comparisons. Association testing of haplotype block P values were carried out using the expectation-maximization (EM) algorithm directly for the observed data. c | Association results from b, shown in greater detail, for a 1.4 Mb interval on 8q24.21. Filled black circles represent 225 SNPs and the orange boxes represent recombination hotspots (calculated from the HapMap using the likelihood ratio test). d | LD between SNPs, measured by the square of the correlation coefficient calculated for each pairwise comparison of SNPs (r²) from the Centre d’Etude du Polymorphisme Humain from Utah (CEU) HapMap population for the 225 SNPs in c; the blue boxes at the bottom indicate the location of the FAM84B, AF268618 and MYC genes and the AW183883 expressed sequence tag. Figure modified, with permission, from Nature Genetics REF. © 2007 Macmillan Publishers Ltd.