Skip to main content
NCBI home page
Biomarker Insights logoLink to Biomarker Insights
. 2007 Aug 8;2:283–292.

Genome-Wide Association Studies: Progress in Identifying Genetic Biomarkers in Common, Complex Diseases

Stephen F Kingsmore 1,✉, Ingrid E Lindquist 1, Joann Mudge 1, William D Beavis 1
PMCID: PMC2717811  PMID: 19662211

Abstract

Novel, comprehensive approaches for biomarker discovery and validation are urgently needed. One particular area of methodologic need is for discovery of novel genetic biomarkers in complex diseases and traits. Here, we review recent successes in the use of genome wide association (GWA) approaches to identify genetic biomarkers in common human diseases and traits. Such studies are yielding initial insights into the allelic architecture of complex traits. In general, it appears that complex diseases are associated with many common polymorphisms, implying profound genetic heterogeneity between affected individuals.

Keywords: Genome-wide association studies, Complex diseases, Complex traits, Genetic biomarkers, Population genetics

Introduction

Genetic biomarkers are uniquely important since they are unambiguously associated with causality of diseases, traits or phenotypes. In common with other types of biomarkers, they are useful for diagnosis, patient stratification and prognostic or therapeutic categorization. Distinctively, however, they are frequently useful for provision of novel insights into disease pathogenesis and, thereby, novel therapeutic targets and strategies. Furthermore, inherited genetic biomarkers are present at birth, enabling institution of timely preventative or ameliorative measures.

During the past 25 years, genetic linkage studies have been exceptionally effective in identifying genetic biomarkers in Mendelian (simple or single gene) disorders. These studies have identified causal gene variants in more than 1300 Mendelian diseases (Botstein and Risch, 2003). Most common diseases, traits and phenotypes, however, do not exhibit Mendelian patterns of inheritance, but rather exhibit complex, multifactorial expression and inheritance. None-the-less, linkage based methods, especially the transmission disequilibrium test (Spielman et al. 1993) were developed and applied to a number of complex traits in the 1990s. These linkage studies met with little success in the identification of the allelic determinants of these common, complex disorders or traits (Freimer and Sabatti, 2004). In particular, there has been a lack of replication among studies, whereby an initial study will identify a genotype with large estimated genetic effects (relative risk) but subsequent studies will not corroborate the results (Lander and Kruglyak, 1995; Göring et al. 2001). In part, this reflects the dependence of linkage studies on unusually informative pedigrees (with multiple affected and unaffected individuals), which induce a bias toward rare, semi-Mendelian disease subsets in sub-populations. Tremendous excitement, therefore, has met recent reports of successful identification of genetic biomarkers in complex traits using an approach that does not have this limitation—genome wide association (GWA) studies.

Genome wide association studies are predicated on two observations:

  1. The brief history of most human populations precludes sufficient number of generations (or meioses) to create recombination events (or mutations) between closely linked markers; and

  2. Suppression of meiotic recombination (coldspots) occurs very frequently in mammalian genomes.

Thus, approximately 80% of the human genome is comprised of ~10 kilobase regions (haplotype blocks, HB) that do not show recombination in human populations (International HapMap Consortium, 2005). Genetic variants within HB are in linkage disequilibrium (LD). This phenomenon enables much of the recombination history in a population to be ascertained by genotyping a large set of randomly spaced tags throughout the genome, especially if these tags are located within HB. During the last ten years, more than ten million single nucleotide polymorphisms (SNPs) have been aggregated in a public database (Sherry et al. 2001). Furthermore, the International HapMap project has genotyped over three million of these SNPs that are common (occur with a minor allele frequency of >5%) in human populations and have assembled these genotypes computationally into a genome-wide map of SNP-tagged HB (International HapMap Consortium, 2005). These resources, together with array technologies for massively parallel SNP genotyping, have made GWA studies feasible. Further, the well established epidemiological case-control experimental design is directly applicable to GWA studies.

Initially, association genetic studies (focused on candidate genetic loci) also exhibited a lack of replication among studies (Ionnidis et al. 2001; Hirschhorn et al. 2002). The potential reasons for inconsistent results consist of unobserved, confounding biological sources of heterogeneity including inconsistent or poorly defined measurements of the phenotype, heterogeneous genetic sources for the phenotype, population stratification, population-specific LD, heterogeneous genetic and epigenetic backgrounds, or heterogeneous environmental influences. In addition to these biological explanations, there are statistical reasons for inconsistent results including failure to control the rate of false discoveries, lack of power, model misspecification and heterogeneous bias in estimated effects among studies (Cardon and Bell, 2001; Cardon and Palmer, 2003; Redden and Allison, 2003; Sillanpää and Auranen, 2004). Among these, the most likely source of non-replication is the lack of power due to the limited number of individuals genotyped and phenotyped in these experiments (Risch, 2000; Lohmueller et al. 2003). In the past year, however, a significant number of studies have been published that have used SNP genotyping arrays in large cohorts of individuals resulting in replicated associations between individual SNP-tagged HB and common, complex traits, phenotypes or diseases. In common with other biomarker development approaches, GWA study designs entail discovery, validation and replication stages (Kingsmore, 2006). This design is critical for detection of meaningful, hypothesis-generating genotype-phenotype associations given the large number of comparisons involved, prior probability estimates of association, sample sizes, resampling procedures, and statistical significance thresholds.

Results of Initial GWA Studies

The first GWA study, published in 2002, examined myocardial infarction (Ozaki et al. 2002). The discovery phase comprised genotyping of 65,671 informative, coding domain SNPs (cSNPs) in 752 cases and control individuals (Table 1). Gene-tagging SNPs are more informative than random SNPs, since the vast majority of true-positive associations will be with genes (Botstein and Risch, 2003). The validation phase featured 26 SNPs and 2137 individuals and confirmed an association with a 50 kb HB containing the lymphotoxin-α (LTA), nuclear factor of kappa light polypeptide gene enhancer in B cells, inhibitor-like 1 (NFKBIL1) and HLA-B associated transcript 1 (BAT1) genes (Table 2). Replication studies have subsequently been undertaken, some of which have confirmed association of this region to myocardial infarction-related phenotypes, and, in particular, to a nonsynonymous SNP (nsSNP) in LTA (Yamada et al. 2004; Laxton et al. 2005; Mizuno et al. 2006; Clarke et al. 2006; Kimura et al. 2007; Sedlacek et al. 2007; Koch et al. 2007). As with most GWA studies, the association of LTA with myocardial infarction was an unexpected finding and suggests novel therapeutic approaches.

Table 1.

Discovery and validation designs of recent GWA studies.

Disease Discovery Phase
Validation Phase
Reference
Cohort size # SNPs Population Cohort size # SNPs validated/tested Population
AMD 146 105,980 Caucasian 96 5/5 Same Klein et al. 2005
Bipolar Disorder 1024, pooled 555,235 Western European 1648 88/1877 Same Baum et al. 2007
Breast Cancer 754 227,876 European 45,426 7/30 same Easton et al. 2007
Breast Cancer 2287 528,173 European 3848 2/8 same Hunter et al. 2007
Breast Cancer 13,163 311,524 Icelandic 7968 2/9 various Stacey et al. 2007
Crohn’s disease 1923 304,413 European 2150 8/20 Same Rioux et al. 2007
Crohn’s disease 1103 16,360 Non-synonymous German 2670 3/72 Same Hampe et al. 2007
Crohn’s disease 1475 317,497 Belgian 2236 10/10 Same Libioulle et al. 2007
IBD 1095 + 834 308,332 European 2885 10/27 Same Duerr et al. 2006
Late-Onset 1086 502,627 Caucasian N.D. — — Coon et al. 2007
Alzheimer’s Memory 341 502,627 Swiss 680 1/2 Several Papassotiropoulos et al. 2006
Myocardial Infarction 752 65,671 cSNPs Japanese 2137 4/26 same Ozaki et al. 202
Obesity 4862 490,032 British/Irish 29,596 1/1 same Frayling et al. 2007
Prostate Cancer 4517 316,515 & 243,957 haplotypes Icelandic 3655 2/2 various Gudmundsson et al. 2007
Prostate Cancer 2339 550,000 European 6266 2/2 Yeager et al. 2007
T2DM 2335 315,635 Finnish 2473 10/80 same Scott et al. 2007
T2DM 16,179 393,453 European 9103 10/77 same Zeggini et al. 2007
T2DM 7805 313,179 & 339,846 haplotypes Icelandic/Danish 3382 2/47 same Steinthorsdottir et al. 2007
T2DM 1316 392,935 French 5511 8/57 same Sladek et al. 2007
T2DM & Triglyeride levels 2931 386,731 & 284,968 haplotypes Finnish/Swedish 10,850 3/107 several Diabetes Genetics Initiative, Broad Inst. et al. 2007
Open in a new tab

Abbreviations: AMD, age-related macular degeneration; IBD, inflammatory bowel disease; T2DM, type II diabetes mellitus; ND, not determined.

Table 2.

Most significant associated Haplotype Blocks (HB) in GWA studies.

Disease Most significant HB OR of Most significant HB Cumulative OR # Alleles in cumulative OR Reference
AMD CFH 7.4 N.D. — Klein et al. 2005
Bipolar disorder DGKH 1.59 3.8 19 Baum et al. 2007
Breast Cancer FGFR2 1.26 3.6% of variance 5 Easton et al. 2007
Breast Cancer FGFR2 1.2 N.D. — Hunter et al. 2007
Breast Cancer BC029912 1.28 PAR 25% 2 Stacey et al. 2007
Crohn’s disease ATG16L1 1.45 N.D. — Rioux et al. 2007
Crohn’s disease ATG16L1 1.45 1.45 28 Hampe et al. 2007
Crohn’s disease IL23R 2.92 N.D. — Libioulle et al. 2007
IBD IL23R 1.56/0.26 N.D. — Duerr et al. 2006
Late-Onset APOE ∈ 4 4.01 N.D. — Coon et al. 2007
Alzheimer’s Memory KIBRA 1.24 N.D. — Papassotiropoulos et al. 2006
Myocardial Infarction LTA 1.69 N.D. — Ozaki et al. 2002
Obesity FTO 1.31 N.D. — Frayling et al. 2007
Prostate Cancer 8q24, locus 1 1.71 PAR 13% 2 Gudmundsson et al. 2007
Prostate Cancer 8q24, locus 2 1.26 Yeager et al. 2007
T2DM TCF7L2 1.37 ~20% incidence 10 Scott et al. 2007
T2DM TCF7L2 1.37 N.D. — Zeggini et al. 2007
T2DM CDKAL1 1.20 N.D. — Steinthorsdottir et al. 2007
T2DM SLC30A8 1.65 heterozygous relative risk PAR 70% 5 Sladek et al. 2007
T2DM KCNJ11, PPARG, TCF7L2 1.14–1.48 5.71 6 WTCCC 2007
T2DM & Triglyeride levels Non-coding near CDKN2A/B 1.20 2.3% of variance 8 Diabetes Genetics Initiative, Broad Inst. et. al. 2007
Open in a new tab

Abbreviations: PAR, Population attributable risk; AMD, age-related macular degeneration; IBD, inflammatory bowel disease; T2DM, type II diabetes mellitus; HB, haplotype block; OR, odds ratio; ND, not determined.

A second, pioneering GWA study examined age-related macular degeneration (AMD) (Klein et al. 2005, Table 1). The discovery phase genotyped 105,980 SNPs in only 146 cases and control individuals. SNPs in the complement factor H (CFH) gene, including an nsSNP, showed significant association with AMD. A validation phase was not performed, but numerous subsequent studies have replicated associations of CFH variants with protection and predisposition to AMD (Zareparsi et al. 2005; Hageman et al. 2005; Souied et al. 2005; Magnusson et al. 2006). Of all complex traits examined by GWA to date, AMD is unique in that a single, novel HB explained 61% of the genetic variance, conferring an odds ratio (OR) of 7.4 (Table 2). To put this in perspective, this OR is of similar magnitude to the classic associations of HLA-B27 with anterior uveitis/ankylosing spondylitis and HLA alleles with type 1 diabetes mellitus. Complement pathway dysregulation was a novel, unexpected association with AMD. Excitingly, this defect is likely to be therapeutically tractable.

During 2006, considerable technical and cost-effectiveness challenges were overcome, resulting in broad adoption and numerous GWA study publications.

Five large GWA studies have examined Crohn’s disease and ulcerative colitis, the two most common types of inflammatory bowel disease (IBD) (Table 1). Three of the studies used microarrays featuring ~300,000 random SNPs (Duerr et al. 2006; Rioux et al. 2007; Libioulle et al. 2007), while the fourth study used custom chips featuring ~16,000 non-synonymous SNPs (Hampe et al. 2007). The behemoth fifth study used 469,557 random SNPs in 14,000 individuals with seven common diseases but lacked replication (WTCCC 2007). Of 14 novel HB associations, 10 were unique to a single study. Three HB were concordant in four of the five studies (representing the genes CARD15, IL23R and ATG16L1). One HB was identified in two of the five studies (PTGER4). Estimated genetic effects, i.e. relative risks, of associated loci were small; the cumulative odds ratio associated with 28 risk alleles was only 1.45. Several studies sought evidence for epistatic interactions between IBD-associated HB. Two studies found suggestive evidence for epistasis involving two different pairs of HB. As yet, IBD candidate genes do not appear to be coalescing into biologic networks or pathways.

Six large GWA studies have examined type II (or adult onset) diabetes mellitus (T2DM), providing the best example to date of capabilities and limitations of GWA studies (Diabetes Genetics Initiative, Broad Inst. et al. 2007; Scott et al. 2007; Sladek et al. 2007; Steinthorsdottir et al. 2007; Zeggini et al. 2007; WTCCC 2007; Table 1). The discovery phase of these studies comprised genotyping of 313,179–469,557 SNPs in 2,335–16,179 individuals. Several studies sought association both with SNPs and with haplotype blocks in the discovery phase. Many of these studies were very large. Case-control and family-based association statistics were employed in most of the studies. Two employed over 9,000 individuals in the validation phase while one lacked replication.

Concordance of associated genes between T2DM GWA studies was striking; of 10 novel HB associations, only two were unique to a single study. Discordance of associations partly reflected different coverage of specific HB by the two microarray platforms used for genotyping. In common with IBD, T2DM-associated HB exhibited small estimated genotypic effect sizes. In contrast to IBD, however, many of the T2DM-associated candidate genes coalesced into biologic pathways, such as pancreatic islet beta cell function, including insulin biosynthesis.

Three studies performed initial modeling of how loci combine to affect susceptibility to T2DM. One study found suggestive evidence for epistatic interactions between two HB. Otherwise, it appeared that T2DM fits a polygenic threshold model with additive/multiplicative effects of individual loci.

As anticipated, allele frequencies showed considerable variation between ethnic and racial groups. Somewhat surprising, however, was the incidence of conservation of T2DM and IBD associated risk alleles between independent populations.

Two studies extended GWA analysis of T2DM to endophenotypes related to serum triglycerides and obesity (Diabetes Genetics Initiative, Broad Inst. et al. 2007; Frayling et al. 2007; Table 1). For example, one gene associated with T2DM, fat mass and obesity associated gene (FTO) (Scott et al. 2007), also showed an association with obesity-associated quantitative traits in an independent study (Frayling et al. 2007). Frayling et al. examined the association of the FTO variant with body mass index (BMI) in 13 cohorts with 38,759 participants. The association of FTO SNPs with obesity has been independently confirmed in 8000 individuals (Dina et al. 2007).

Cancers are fascinating genetic diseases as they feature the combined effects of germline risk alleles and multiple somatic mutations. Three GWA studies sought inherited haplotype block associations with breast cancer (Easton et al. 2007; Stacey et al. 2007; Hunter et al. 2007). The discovery phase of these studies comprised genotyping of 227,876–528,173 SNPs in 2,287–13,163 individuals. Validation phases varied in size from 3,848–44,438 individuals. Associations were identified in SNPs in FGFR2 (two studies), TNRC9 (two studies), at haplotype blocks rs13387042 and rs3817198 (which do not contain known genes, one study each), MAP3K1 (one study) and LSP1 (one study).

Two GWA studies sought association signals in prostate cancer (Gudmundsson et al. 2007; Yeager et al. 2007). The discovery phase of these studies comprised genotyping of 316,515–550,00 SNPs in 2,339–4,517 individuals. Validation phases varied in size from 3,655–6,266 individuals. An association with a HB at 8q24 that had previously been identified by linkage analysis (Amundadottir et al. 2006) was seen in both studies. In addition, both studies identified a second 8q24 HB, ~300 kb upstream from the first. As yet, the functional basis of these associations is unclear. While individual 8q24 variants showed modest estimated genetic effects, the cumulative effect of several variants fits a multiplicative model that conferred a population attributable risk (PAR), the expected reduction in prostate cancer incidence if the risk alleles did not exist in the population, of up to 68% (Haiman et al. 2007).

The applicability of GWA studies to complex traits has also been demonstrated. One study undertook GWA with numerous quantitative and categorical memory-associated endophenotypes (Papassotiropoulos et al. 2006). Despite the small size of the discovery cohort (341 individuals), associations with one HB (in the KIBRA gene) were replicated in two validation cohorts (totaling 680 individuals). A notable strength of this study was that associations were sought with multiple types of endophenotypes (performance in seven memory-associated tests and functional magnetic resonance image-based measures of the hippocampus during three memory-associated tests).

In addition to the identification of novel HB associations, GWA studies have confirmed several associations of susceptibility genes that were previously established by linkage analysis in large pedigrees. For example, a GWA study of 502,627 SNPs in 1086 cases of late-onset Alzheimer’s disease and controls verified the well established APOE susceptibility gene (Coon et al. 2007).

A remaining problem with large GWA studies is genotyping cost. One innovative study provided evidence that sample pooling strategies may be effective. In a GWA study of bipolar disorder, investigators created 39 pools, each containing equimolar amounts of DNA from 42–80 individuals (Baum et al. 2007). These pools represented a discovery and replication cohort. Pools were individually genotyped for 555,235 SNPs and normalized allele frequencies were inferred from intensity data. Replicates were assayed for each pool. 37 SNPs showing allele frequency differences in both cohorts were individually genotyped and 76% retained significant associations.

Conclusions

During the past year, the utility of GWA studies for identification of novel genomic associations with complex disorders has unambiguously been established. In general these studies have employed large case-control cohorts featuring both familial and sporadic cases, categorical trait definitions and up to half a million commonly polymorphic SNPs. Excitingly, these studies are starting to provide empiric data to resolve decades of debate about the genetic architecture of complex traits. To date, with the exception of CFH in AMD, the estimated genetic effects of replicated associations have been uniformly and surprisingly small (Table 2). Also surprising is the high frequency of many risk alleles, albeit this may reflect an artifact induced by use of genotyping arrays that primarily feature common polymorphisms (Table 3). Informed by studies to date, the picture of the genetic architecture of complex traits that is emerging is immense polygenicity and individual genetic heterogeneity. In general, the data fit additive, threshold models. In a handful of informative studies, little evidence for epistasis has been observed. If confirmed, an implication will be that genetic diagnostics are still a long way off and will certainly not result in the deterministic prognostications portrayed in the 1997 movie Gattaca.

Table 3.

Candidate variant and functional consequence in associated Haplotype Block (HB).

Disease Most significant HB Variant type HB allele frequency Functional consequence Reference
AMD CFH exon 9 Non-synonymous 0.38 Complement dysregulation Klein et al. 2005
Bipolar disorder DGKH, introns 1 & 7 Various 0.22–0.44 Not known Baum et al. 2007
Breast Cancer FGFR2, intron 2 Non-coding 0.48 Unknown Easton et al. 2007
Breast Cancer FGFR2, intron 2 Non-coding 0.49 Unknown Hunter et al. 2007
Breast Cancer BC029912 Synonymous 0.50 Unknown Stacey et al. 2007
Crohn’s Disase IL23R Non-synonymous 0.07 As for IL23R above Libioulle et al. 2007
Crohn’s Disease ATG16L1 Non-synonymous 0.47 Altered autophagy Hampe et al. 2007
IBD IL23R exons 5–11, 3′ intergenic region Non-synonymous, non coding 0.07 T cell mediated inflammation Duerr et al. 2006
Late-Onset APOE ∈4 Non-synonymous 0.14 Unknown Coon et al. 2007
Alzheimer’s Memory KIBRA, intron 9 Non-coding 0.48 Unknown Papassotiropoulos et al. 2006
Myocardial infarction BAT - LTA Various N.K. Not known Ozaki et al. 2002
Obesity FTO, intron 1 & 2, exon 2 Various 0.41 Unknown Frayling et al. 2007
Prostate Cancer 8q24, 300kb Unknown Various Unknown Gudmundsson et al. 2007 & Yeager et al. 2007
T2DM TCF7L2 Intronic 0.28 Transcriptional control of insulin synthesis Scott et al. 2007 & Zeggini et al. 2007
T2DM CDKAL1, intron 5 N.K. (LD block 202 kb) 0.5 unknown Steinthorsdottir et al. 2007
T2DM SLC30A8 Non-synonymous 0.38 Insulin synthesis Sladek et al. 2007
T2DM & triglyceride levels CDKN2A/B Non-coding 0.36 unknown Diabetes Genetics Initiative, Broad Inst. et al. 2007
Open in a new tab

Abbreviations: AMD, age-related macular degernation; IBD, inflammatory bowel disease; T2DM, type II diabetes mellitus; HB, haplotype block; LD, linkage disequilibrium; NK, not known.

Encouragingly, most associated haplotype blocks are small enough to feature a single gene (Table 3). In large measure, this reflects the use of more outbred populations in fine-mapping validation phases. Furthermore, many HB contain a single, unequivocally functional variant. The distribution of variants does not yet show much difference from causal variants identified in Mendelian disorders. Thus, nsSNPs are relatively common and rSNPs are uncommon, a controversial point (Botstein and Risch, 2003; Knight, 2005; Thomas and Kejariwal, 2004). The vast majority of associated genes identified to date were not candidate genes previously, continuing the marvelous tendency of comprehensive, genetic-driven studies to be hypothesis-informing. It is refreshing to see associations with genes that had never previously been considered in a disease or trait. As yet, the confluence of associated genes into biologic networks and pathways has been disappointing. Surprisingly, there appears to be significant conservation of variant associations between human populations (albeit in the setting of frequently different allele frequencies). The emerging, principal benefit of GWA studies may therefore be elucidation of molecular mechanisms underpinning poorly understood diseases and traits.

In the few informative studies reported to date, endophenotypes have been highly instructive in dissecting the network or pathway perturbed by an individual variant to impact a complex trait. It is particularly exciting to see the application of multimode endophenotypes, such as combinations of psychological testing, brain imaging and gene expression in one study (Papassotiropoulos et al. 2006). The very large cohorts needed to discover and validate variants with small effect sizes preclude the collection of rich, accurate metadata. It is likely that future studies will utilize much greater stratification of traits than the phenotypically crude studies reported to date. Recent GWA studies of breast cancer provide a good example of the added genetic complexity that can be revealed by trait stratification (Easton et al. 2007; Hunter et al. 2007; Stacey et al. 2007). In addition, following replication of associations with categorical traits, it is anticipated that targeted genotypic examination of many endophenotypes will be highly instructive in the dissection of disease pathogenesis.

Future Developments and Implications

Several trends observed in GWA studies to date are anticipated to continue. One million SNP chips are about to be launched and genotype accuracies have improved. Cohort sizes are increasing. Combinations of genotype-based and haplotype-based associations are becoming more prevalent. Experimental designs and statistical methods are becoming more uniform, enabling more meaningful meta-analyses. In particular, the emergence of adaptive designs and the use of Bayesian inferential methods produce probabilistic synthesis from combined analysis (Barry, 2005). Importantly, this will provide an intuitive framework for combining information from multiple studies resulting in more effective utilization of patient information from translational research—especially for detection and validation of weak associations.

As noted above, phenotypes remain crude and the use of endophenotypes or component phenotypes is anticipated to increase significantly. In particular, biomarker phenotypes are anticipated to become widely used. These are likely to include gene expression, proteomic, metabolomic and imaging biomarkers. As determinants of complex traits are identified, genetic stratification of cases and controls will be possible, reducing the genetic complexity of the trait and enabling identification of additional association signals. An area of substantial interest for the pharmaceutical industry will be GWA studies of drug response that identify patient stratification markers for clinical trials and guide drug improvement, particularly for avoidance of adverse events.

Despite the current euphoria, GWA studies are likely to have significant limitations. Insufficient numbers of cases will be available for GWA of uncommon traits or diseases. Current GWA genotyping methods will not identify associations with rare variants, even with large effect sizes. Approximately twenty percent of the genome represents recombinational hotspots that are not amenable to LD-based approaches (International HapMap Consortium 2005). At recombinational coldspots, haplotype blocks may be too large for unambiguous identification of causal variants. The extent of the effect of copy number variation (CNV) on association signals is not yet clear. For some common traits or diseases, these considerations may reflect a substantial proportion of the genetic variance. The addition of gene expression profiling, CNV estimation and large-scale resequencing technologies to GWA studies should circumvent some of these limitations. Use of adaptive statistical methods and resampling strategies may circumvent the need for thousands of affected individuals in uncommon traits (Berry, 2004).

Clearly a huge amount of genetics, biochemistry and cell biology remains to be done to confirm the biologic relevance of associations and to elucidate the mechanisms of genotype-phenotype associations. For geneticists, a long term goal is to piece together the genetic architecture of complex traits, evaluating with much greater precision the genetic model and contributions of factors such as epistasis, genocopies, phenocopies and penetrance.

Acknowledgements

A Deo lumen, ab amicis auxilium. This work was partially supported by National Institutes of Health grants N01A000064 and U01AI066569, and by National Science Foundation grant 0524775.

References

  1. Amundadottir LT, Sulem P, Gudmundsson J, et al. A common variant associated with prostate cancer in European and African populations. Nat Genet. 2006;38(6):652–8. doi: 10.1038/ng1808. [DOI] [PubMed] [Google Scholar]
  2. Baum AE, Akula N, Cabanero M, et al. A genome-wide association study implicates diacylglycerol kinase eta (DGKH) and several other genes in the etiology of bipolar disorder. Molecular Psychiatry. 2007:1–11. doi: 10.1038/sj.mp.4002012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Berry DA. Bayesian Statistics and the efficiency and ethics of clinical trials. Statistical Science. 2004;19:175–187. [Google Scholar]
  4. Botstein D, Risch N. Discovering genotypes underlying human phenotypes: past successes for Mendelian disease, future approaches for complex disease. Nat Genet. 2003;33:228–37. doi: 10.1038/ng1090. [DOI] [PubMed] [Google Scholar]
  5. Cardon LR, Bell JI. Association study designs for complex diseases. Nat Rev Genet. 2001;2:91–99. doi: 10.1038/35052543. [DOI] [PubMed] [Google Scholar]
  6. Cardon LR, Palmer LJ. Population stratification and spurious allelic association. Lancet. 2003;361:598–604. doi: 10.1016/S0140-6736(03)12520-2. [DOI] [PubMed] [Google Scholar]
  7. Clarke R, Xu P, Bennett D, et al. Lymphotoxin-alpha gene and risk of myocardial infarction in 6,928 cases and 2,712 controls in the ISIS case-control study. PLoS Genet. 2006;27(e107):0990–6. doi: 10.1371/journal.pgen.0020107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Coon KD, Myers AJ, Craig DW, et al. A high-density whole-genome association study reveals that APOE is the major susceptibility gene for sporadic late-onset Alzheimer’s disease. J Psychiat. 2007;68(4):613–18. doi: 10.4088/jcp.v68n0419. [DOI] [PubMed] [Google Scholar]
  9. Diabetes Genetics Initiative of Broad Institute of Harvard M.I.T. Lund University and Novartis Institutes for BioMedical Research. Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels. Science. 2007;316:1331–6. doi: 10.1126/science.1142358. [DOI] [PubMed] [Google Scholar]
  10. Dina C, Meyre D, Gallina S, et al. Variation in FTO contributes to childhood obesity and severe adult obesity. Nat Genet. 2007;39:724–6. doi: 10.1038/ng2048. [DOI] [PubMed] [Google Scholar]
  11. Duerr RH, Taylor KD, Brant SR, et al. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science. 2006;314:1461–3. doi: 10.1126/science.1135245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Easton DF, Pooley KA, Dunning AM, et al. Genome-wide association study identifies novel breast cancer susceptibility loci. Nature. 2007 May 27; doi: 10.1038/nature05887. Adv. Online Pub. (AOP) [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Flotho C, Steinemann D, Mullighan CG, et al. Genome-wide single-nucleotide polymorphism analysis in juvenile myelomonocytic leukemia identifies uniparental disomy surrounding the NF1 locus in cases associated with neurofibromatosis but not in cases with mutant RAS or PTPN11. Oncogene. 2007:1–6. doi: 10.1038/sj.onc.1210361. [DOI] [PubMed] [Google Scholar]
  14. Fraying TM, Timpson NJ, Weedon MN, et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science. 2007;316:889–94. doi: 10.1126/science.1141634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Freimer N, Sabatti C. The use of pedigree, sib-pair and association studies of common diseases for genetic mapping and epidemiology. Nat Genet. 2004;36:1045–51. doi: 10.1038/ng1433. [DOI] [PubMed] [Google Scholar]
  16. Gudmundsson J, Sulem P, Manolescu A, et al. Genome-wide association study identifies a second prostate cancer susceptibility variant at 8q24. Nat Genet. 2007;39:631–7. doi: 10.1038/ng1999. [DOI] [PubMed] [Google Scholar]
  17. Göring HH, Terwilliger JD, Blangero J. Large upward bias in estimation of locus-specific effects from genomewide scans. Am J Hum Genet. 2001;69:1357–1369. doi: 10.1086/324471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hageman GS, Anderson DH, Johnson LV, et al. A common haplotype in the complement regulatory gene factor H (HF1/CFH) predisposes individuals to age-related macular degeneration. Proc Natl Acad Sci, USA. 2005;102:7227–32. doi: 10.1073/pnas.0501536102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions within 8q24 independently affect risk for prostate cancer. Nat Genet. 2007;39(5):638–44. doi: 10.1038/ng2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hampe J, Franke A, Rosenstiel P, et al. A genome-wide association scan of nonsynonymous SNPs identifies a susceptibility variant for Crohn disease in ATG16L1. Nat Genet. 2007;39(2):207–11. doi: 10.1038/ng1954. [DOI] [PubMed] [Google Scholar]
  21. Hirschhorn JN, Lohmueller K, Byrne E, et al. A comprehensive review of genetic association studies. Gen Med. 2002;4(2):45–61. doi: 10.1097/00125817-200203000-00002. [DOI] [PubMed] [Google Scholar]
  22. Hunter DJ, Kraft P, Jacobs KB, et al. A genome-wide association study identifies alleles in FGFR2 associated with risk of sporadic postmenopausal breast cancer. Nat Genet (AOP) 2007 May 27; doi: 10.1038/ng2075. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. International HapMap Consortium. A haplotype map of the human genome. Nature. 2005;437:1299–1320. doi: 10.1038/nature04226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Ionnidis JP, Ntzani EE, Trikalinos TA, et al. Replication validity of genetic association studies. Nat Genet. 2001;29:306–309. doi: 10.1038/ng749. [DOI] [PubMed] [Google Scholar]
  25. Kimura A, Takahashi M, Choi BY, et al. Lack of association between LTA and LGALS2 polymorphisms and myocardial infarction in Japanese and Korean populations. Tissue Antigens. 2007;69:265–9. doi: 10.1111/j.1399-0039.2006.00798.x. [DOI] [PubMed] [Google Scholar]
  26. Kingsmore SF. Multiplexed protein measurement: Technologies and Applications of Antibody Arrays. Nat Rev Drug Discov. 2006;5:310–321. doi: 10.1038/nrd2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Klein RJ, Zeiss C, Chew EY, et al. Complement factor H polymorphism in age-related macular degeneration. Science. 2005;308:385–9. doi: 10.1126/science.1109557. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Knight JC. Regulatory polymorphisms underlying complex disease traits. J Mol Med. 2005;83(2):97–109. doi: 10.1007/s00109-004-0603-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Koch W, Hoppmann P, Michou E, et al. Association of variants in the BAT1-NFKBIL1-LTA genomic region with protection against myocardial infarction in Europeans. Hum Mol Genet. 2007 May 21; doi: 10.1093/hmg/ddm130. [DOI] [PubMed] [Google Scholar]
  30. Lander E, Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat Genet. 1995;11(3):241–247. doi: 10.1038/ng1195-241. [DOI] [PubMed] [Google Scholar]
  31. Laxton R, Pearce E, Kyriakou T, et al. Association of the lymphotoxin-alpha gene Thr26Asn polymorphism with severity of coronary atherosclerosis. Genes Immun. 2005;6:539–41. doi: 10.1038/sj.gene.6364236. [DOI] [PubMed] [Google Scholar]
  32. Libioulle C, Louis E, Hansoul S, et al. Novel Crohn disease locus identified by genome-wide association maps to a gene desert on 5p13. 1 and modulates expression of PTGER4. PLoS Genet. 2007;34(e58):0538–43. doi: 10.1371/journal.pgen.0030058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lohmueller KE, Pearce CL, Pike M, et al. Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility of common disease. Nat Genet. 2003;33:177–182. doi: 10.1038/ng1071. [DOI] [PubMed] [Google Scholar]
  34. Magnusson KP, Duan S, Sigurdsson H, et al. CFH Y402H confers similar risk of soft drusen and both forms of advanced, AMD. PLoS Med. 2006;31(e5):0109–14. doi: 10.1371/journal.pmed.0030005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Mizuno H, Sato H, Sakata Y, et al. Impact of atherosclerosis-related gene polymorphisms on mortality and recurrent events after myocardial infarction. Atherosclerosis. 2006;185:400–5. doi: 10.1016/j.atherosclerosis.2005.06.020. [DOI] [PubMed] [Google Scholar]
  36. Mullighan CG, Goorha S, Miller CB, et al. Genome-wide analysis of genetic alterations in acute lymphoblastic leukaemia. Nature. 2007;446:758–64. doi: 10.1038/nature05690. [DOI] [PubMed] [Google Scholar]
  37. Ozaki K, Ohnishi Y, Iida A, et al. Functional SNPs in the lymphotoxin-alpha gene that are associated with susceptibility to myocardial infarction. Nat Genet. 2002;32:650–4. doi: 10.1038/ng1047. [DOI] [PubMed] [Google Scholar]
  38. Papassotiropoulos A, Stephan DA, Huentelman MJ, et al. Common KIBRA alleles are associated with human memory performance. Science. 2006;314:475–8. doi: 10.1126/science.1129837. [DOI] [PubMed] [Google Scholar]
  39. Redden DT, Allison DB. Nonreplication in genetic association studies of obesity and diabetes research. J Nutr. 2003;133:3323–3326. doi: 10.1093/jn/133.11.3323. [DOI] [PubMed] [Google Scholar]
  40. Rioux JD, Xavier RJ, Taylor KD, et al. Genome-wide association study identifies new susceptibility loci for Crohn disease and implicates autophagy in disease pathogenesis. Nat Genet. 2007;39(5):596–604. doi: 10.1038/ng2032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Risch NJ. Searching for genetic determinants in the new millennium. Nature. 2000;405:847–856. doi: 10.1038/35015718. [DOI] [PubMed] [Google Scholar]
  42. Scott LJ, Mohlke KL, Bonnycastle LL, et al. A genome-wide association study of type 2 diabetes in Finns detects multiple susceptibility variants. Science. 2007;316:1341–5. doi: 10.1126/science.1142382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sedlacek K, Neureuther K, Mueller JC, et al. Lymphotoxin-alpha and galectin-2 SNPs are not associated with myocardial infarction in two different German populations. J Mol Med. 2007 May 12; doi: 10.1007/s00109-007-0211-4. [DOI] [PubMed] [Google Scholar]
  44. Sherry ST, Ward MH, Kholodov M, et al. dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001;29:308–11. doi: 10.1093/nar/29.1.308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sillanpää MJ, Auranen K. Replication in genetic studies of complex traits. Ann Hum Genet. 2004;68:646–657. doi: 10.1046/j.1529-8817.2004.00122.x. [DOI] [PubMed] [Google Scholar]
  46. Sladek R, Rocheleau G, Rung J, et al. A genome-wide association study identifies novel risk loci for type 2 diabetes. Nature. 2007;445:881–5. doi: 10.1038/nature05616. [DOI] [PubMed] [Google Scholar]
  47. Souied EH, Leveziel N, Richard F, et al. Y402H complement factor H polymorphism associated with exudative age-related macular degeneration in the French population. Mol Vis. 2005;11:1135–40. [PubMed] [Google Scholar]
  48. Spielman RS, McGinnis RE, Ewens WJ. Transmission test for linkage disequilibrium: the insulin gene regions and insulin-dependent diabetes mellitus (IDDM) Am J Hum Genet. 1993;52:506–516. [PMC free article] [PubMed] [Google Scholar]
  49. Stacey SN, Manolescu A, Sulem P, et al. Common variants on chromosomes 2q35 and 16q12 confer susceptibility to estrogen receptor-positive breast cancer. Nat Genet (AOP) 2007 May 27; doi: 10.1038/ng2064. [DOI] [PubMed] [Google Scholar]
  50. Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al. A variant in CDKAL1 influences insulin response and risk of type 2 diabetes. Nat Genet. 2007;39:770–5. doi: 10.1038/ng2043. [DOI] [PubMed] [Google Scholar]
  51. Thomas PD, Kejariwal A. Coding single-nucleotide polymorphisms associated with complex vs. Mendelian disease: evolutionary evidence for differences in molecular effects. Proc Natl Acad Sci, USA. 2004;101:15398–15403. doi: 10.1073/pnas.0404380101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wellcome Trust Case Control Consortium (WTCCC) Genome-wide association study of 14, 000 cases of seven common diseases and 3, 000 shared controls. Nature. 2007;447:661–78. doi: 10.1038/nature05911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Yamada A, Ichihara S, Murase Y, et al. Lack of association of polymorphisms of the lymphotoxin alpha gene with myocardial infarction in Japanese. J Mol Med. 2004;82:477–83. doi: 10.1007/s00109-004-0556-x. [DOI] [PubMed] [Google Scholar]
  54. Yeager M, Orr N, Hayes RB, et al. Genome-wide association study of prostate cancer identifies a second risk locus at 8q24. Nat Genet. 2007;39:645–9. doi: 10.1038/ng2022. [DOI] [PubMed] [Google Scholar]
  55. Zareparsi S, Branham KE, Li M, et al. Strong association of the Y402H variant in complement factor H at 1q32 with susceptibility to age-related macular degeneration. Am J Hum Genet. 2005;77:149–53. doi: 10.1086/431426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zeggini E, Weedon MN, Lindgren CM, et al. Replication of genome-wide association signals in UK samples reveals risk loci for type 2 diabetes. Science. 2007;316:1336–41. doi: 10.1126/science.1142364. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Biomarker Insights are provided here courtesy of SAGE Publications

RESOURCES