Approaches for Evaluating Rare Polymorphisms in Genetic Association Studies

Genome-wide associations studies (GWAS) have been very successful in identifying common genetic variation associated to numerous complex diseases [1]. However, most of the identified common genetic variants appear to confer modest risk and few causal alleles have been identified [2]. Furthermore, these associations account for a small portion of the total heritability of inherited disease variation [1]. This has led to the reexamination of the contribution of environment, gene-gene and gene-environment interactions, and rare genetic variants in complex diseases [1, 3, 4]. There is strong evidence that rare variants play an important role in complex disease etiology and may have larger genetic effects than common variants [2]. Currently, much of what we know regarding the contribution of rare genetic variants to disease risk is based on a limited number of phenotypes and candidate genes. However, rapid advancement of second generation sequencing technologies will invariably lead to widespread association studies comparing whole exome and eventually whole genome sequencing of cases and controls. A tremendous challenge for enabling these "next generation" medical genomic studies is developing statistical approaches for correlating rare genetic variants with disease outcome. The analysis of rare variants is challenging since methods used for common variants are woefully underpowered. Therefore, methods that can deal with genetic heterogeneity at the trait-associated locus have been developed to analyze rare variants. These methods instead analyzing individual variants analyze variants within a region/gene as a group and usually rely on collapsing. They can be applied to both in cases vs. controls and quantitative trait studies are needed. The paper of Bansal et al. in this volume describes the application of a number of statistical methods for testing associations between rare variants in two genes to obesity. The authors considered the relative merits of the different methods as well as important implementation details, such as the leveraging of genomic annotations and determining p-values. Knowledge of haplotypes can increase the power of GWAS studies and also highlight associations that are impossible to detect without haplotype phase (e.g. loss of heterozygosity). Even more complicated phase-dependent interactions of variants in linkage equilibrium have also been suggested as possible causes of missing heritability. In their work, Hallsorsson et al. formulate algorithmic strategies for haplotype phasing by multi-assembly of shared haplotypes from next-generation sequencing data. These methods would allow testing haplotypes harboring rare variants for association and potentially increase their explanatory power. Since single SNP tests are often underpowered in rare variant association analysis, Zeggini and Asimit propose a locus-based method that has high power in the presence of rare variants and that incorporate base quality scores available for sequencing data. Their results suggest that this multi-marker approach may be best suited for smaller regions, or after some filtering to reduce the number of SNPs that are jointly tested to reduce loss of power due to multiple-testing adjustments. Finally, the paper of Zhou et al., presents a penalized regression framework for association testing on sequence data, in the presence of both common and rare variants. This method also introduces the use of weights to incorporate available biological information on the variants. Although these tactics improve both false positive and false negative rates, they represent an incremental development and there is still significant room for improvement. With the development of sequencing technologies and methods to detect complex trait rare variant associations many new and exciting discovery are imminent. The analysis of rare variants is still in its infancy and the next few years promises to produce many new methods to meet the special demands of analyzing this type of data. Note from Publisher: This article contains the abstract and references.

For decades, the heritable contribution of cancer has been the subject of intense study, first pursued in families with multiple affected individuals and more recently investigated in unrelated population-based studies. For example, family studies have yielded important genes for breast cancer (BRCA1 and BRCA2), colon cancer (MSH2 and MLH1), and families with a spectrum of less common cancers (TP53). The pursuit of candidate genes yielded only a handful of conclusive associations, such as NAT2 and GSTM1 in bladder cancer or alcohol dehydrogenase genes (ADH1B and ADH7) in aerodigestive cancers. New genotyping technologies together with a comprehensive map of human haplotypes (International HapMap Project) have enabled investigators to scan across the genome in a sufficiently large set of cases and controls without a set of prior hypotheses in search of common susceptibility alleles with low effect sizes. Genome-wide association studies (GWAS) have emerged as an important tool for discovering regions of the genome using hundreds of thousands of common single nucleotide polymorphism (SNP) markers. This ‘agnostic’ approach has conclusively discovered more than 125 regions for nearly two dozen types of cancer. At least eight regions have been associated with more than one distinct cancer type. For instance, the region flanking the MYC oncogene on 8q24 harbors at least five independent loci associated with prostate cancer as well as loci associated with cancers of the breast, colon, bladder, ovaries, and chronic lymphocytic leukemia. On 5p15.33, there are at least seven cancers, namely, basal cell carcinoma, bladder, brain, lung (adenocarcinoma), melanoma, pancreas, and testicular that map to the TERT-CLPTM1L locus. This locus is notable for the telomerase gene (TERT) in which rare mutations have been associated with dyskeratosis congenital (an inherited bone marrow failure syndrome), idiopathic pulmonary fibrosis, acute myelogenous leukemia, and chronic lymphocytic leukemia. The number of variants identified by GWAS per cancer varies greatly, suggesting different underlying genetic architectures, namely genetic contributions to common and uncommon cancers. Of the 125 regions associated with cancer risk, none of the regions have been conclusively associated with prognosis. In prostate cancer, there are at least 35 distinct loci harboring common susceptibility alleles identified by GWAS yet not a single one clearly distinguishes between aggressive and nonaggressive disease. This observation suggests that there could be regions that distinctly contribute to cancer risk and cancer outcomes (e.g., aggressive or metastatic disease). For lung cancer, a disease strongly driven by exposure to tobacco products, so far, only three or four regions have been conclusively established. Of these regions, the signal on chromosome 15 could be related to smoking. For nearly all common SNP markers (namely those with a minor allele frequency of greater than 5%), the estimated effect sizes are below 1.5. A notable exception is for testicular cancer in which variants on chromosome 12 near the KITLG gene have estimated effect sizes of greater than 2.5 for heterozygotes; this is not surprising since the hereditary component of testicular cancer is very strong. Overall, each region confers a small contribution to the risk for cancer, which suggests that it is daunting to consider any single SNP as a clinical test. GWAS regions are undergoing fine mapping to nominate the optimal candidates for functional studies designed to explain the direct association. Nearly all GWAS signals have mapped to noncoding regions, underscoring the contribution of regulatory mechanisms in neighboring genes (e.g., FGFR2 and breast cancer, MSMB/NCOA4 and prostate cancer). The plausibility of the underlying association can form the foundation for developing new strategies to intervene or diagnose cancer at an earlier stage by providing new insights into biological mechanisms that become targets. The application of GWAS to therapeutic outcomes and toxicities is beginning to discover new regions worthy of follow-up studies. However, for pharmacogenomics to eventually have an impact in the clinic, there will have to be a transition from discovery to characterization of the regions that ultimately yields targets for therapy and chemoprevention. Since we are still early in the discovery of common genetic variants associated with risk for specific cancers, it is not surprising that the utility of applying common SNPs to disease prediction is premature. Common genetic variants represent only a proportion of genetic variants that contribute to disease risk; uncommon, rare and copy number variants will undoubtedly contribute to risk in both familial and unrelated settings, and their relative contributions are expected to vary by cancer site. The addition of common SNP markers to classical epidemiological risk factors for breast cancer has minimally shifted the receiver operator curve (ROC) from 50% to approximately 60%. For prostate cancer, a disease in which epidemiological studies have identified age, ethnic background, and family history as the only conclusive risk factors, it is not surprising that more than 30 common SNPs shift the ROC curve to roughly 65%. Discovery of additional genetic variants should improve the risk prediction but each will be limited by the overall heredity of the cancer. New approaches towards applying common genetic variants could be beneficial in the reclassification of individuals based on genetic risk profiles or alternatively, the identification of subsets of individuals at high risk who might undergo a diagnostic or therapeutic intervention. In conclusion, the fruits of GWAS have important implications for investigating the etiology of different cancers, particularly as it relates to gene-environment and gene-gene interactions. The complex genomic architecture of disease susceptibility will require further discovery of uncommon and rare variants using next-generation sequence technologies. It will be critical to assess the applicability of genetic tests in specific clinical settings, such as when to perform screening tests with calculable risks (e.g., biopsies or chemoprevention) before incorporating SNPs into clinical practice. To incorporate the fruits of current genomic observations, new studies will need to be designed to validate the utility of known genetic variants in assessing risk for cancer as well as its outcomes. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr PL01-02. doi:10.1158/1538-7445.AM2011-PL01-02

Schizophrenia is a complex and often chronic psychiatric disorder with high heritability. Diagnosis of schizophrenia is still made clinically based on psychiatric symptoms; no diagnostic tests or biomarkers are available. Pathophysiology-based diagnostic scheme and treatments are also not available. Elucidation of the pathogenesis is needed for development of pathology-based diagnostics and treatments. In the past few decades, genetic research has made substantial advances in our understanding of the genetic architecture of schizophrenia. Rare copy number variations (CNVs) and rare single-nucleotide variants (SNVs) detected by whole-genome CNV analysis and whole-genome/-exome sequencing analysis have provided the great advances. Common single-nucleotide polymorphisms (SNPs) detected by large-scale genome-wide association studies have also provided important information. Large-scale genetic studies have been revealed that both rare and common genetic variants play crucial roles in this disorder. In this review, we focused on CNVs, SNVs, and SNPs, and discuss the latest research findings on the pathogenesis of schizophrenia based on these genetic variants. Rare variants with large effect sizes can provide mechanistic hypotheses. CRISPR-based genetics approaches and induced pluripotent stem cell technology can facilitate the functional analysis of these variants detected in patients with schizophrenia. Recent advances in long-read sequence technology are expected to detect variants that cannot be detected by short-read sequence technology. Various studies that bring together data from common variant and transcriptomic datasets provide biological insight. These new approaches will provide additional insight into the pathophysiology of schizophrenia and facilitate the development of pathology-based therapeutics.

IL28B single nucleotide polymorphisms in the treatment of hepatitis C

Previously described methods for the combined analysis of common and rare variants have disadvantages such as requiring an arbitrary classification of variants or permutation testing to assess statistical significance. Here we propose a novel method which implements a weighting scheme based on allele frequencies observed in both cases and controls. Because the test is unbiased, scores can be analyzed with a standard t-test. To test its validity we applied it to data for common, rare, and very rare variants simulated under the null hypothesis. To test its power we applied it to simulated data in which association was present, including data using the observed allele frequencies of common and rare variants in NOD2 previously reported in cases of Crohn’s disease and controls. The method produced results that conformed well to those expected under the null hypothesis. It demonstrated more power to detect association when rare and common variants were analyzed jointly, the power further increasing when rare variants were assigned higher weights. 20,000 analyses of a gene containing 62 variants could be performed in 80 minutes on a laptop. This approach shows promise for the analysis of data currently emerging from genome wide sequencing studies.

Genetic Association Studies Identify Unanticipated Gene Pathways Influencing Sepsis Outcome

Two genome-wide association studies of prostate cancer identified single nucleotide polymorphism (SNP) markers in a region of chromosome 11q13. As part of the Cancer Genetic Markers of Susceptibility (CGEMS) Initiative the region flanking the most significant marker, rs10896449, was fine mapped using common single nucleotide polymorphisms (SNPs) selected from a two-staged tagging strategy. Of the 120 SNPs analyzed in 10,272 cases and 9,123 controls of European origin, single locus analysis identified 18 SNPs below genome-wide significance (P < 10−8); rs10896449, initially reported remained most significant (P = 7.94 × 10−19). Multi-locus models that included the 18 significant SNPs sequentially identified a second association at rs12793759 (OR = 1.14 P = 4.76 × 10−5), independent of rs10896449 that remained significant after adjustment for multiple testing within the region. A third signal, rs10896438 (OR = 1.07, P = 5.92 × 10−3), independent of both rs10896449 and rs12793759 was detected. To comprehensively catalog genetic variants in strong LD with the above SNPs as well as to determine the region's LD pattern, next-generation sequence analysis of 123 kb (chr11: 68,642,755 - 68,765,690) was performed in 75 individuals of European origin. 447 SNPs with minor allele frequency (MAF) > 1% were identified, which included 180 novel SNPs. Based on r2 > 0.8, 105 SNPs are needed to monitor SNPs with MAF > 1%, whereas 62 SNPs are required for MAF > 5%. 51 SNPs with MAF > 5% are not monitored with r2 > 0.8; 24 singletons (most of which were monitored with r2 > 0.6) and 27 SNPs in 9 correlation bins (r2 > 0.8). For the strongest hits, rs10896449, rs12793759 and rs10896438, 26, 6 and 18 SNPs are correlated with r2 > 0.8, respectively. Our results illustrate the complex architecture of the common genetic variants conferring prostate cancer risk on chromosome 11q13. The fine mapping and sequence analysis point towards a subset of SNPs worthy of follow-up studies designed to understand the molecular basis of the susceptibility alleles. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 3863.

PTEN (phosphatase and tensin homolog deleted on chromosome 10) functions as a major tumor suppressor gene and is frequently deleted in different types of tumors including prostate cancer (PCa). It was hypothesized that germ-line genetic changes of PTEN affect susceptibility to PCa. Both common (with a minor allele frequency 5%) and rare (with a minor allele frequency <5%) germ-line variants of PTEN were comprehensively evaluated. A total of 15 germ-line variants were identified by re-sequencing the PTEN gene, including 5' untranslated region, all nine exons, exon-intron junctions and 3' untranslated region, in 188 probands of hereditary prostate cancer (HPC) families recruited from Johns Hopkins Hospital. Two microsatellite markers surrounding PTEN were used to test the co-segregation of 10 rare variants, which may give rise to highly penetrance in HPC. Two common single nucleotide polymorphisms (SNPs) were evaluated in the 188 HPC families using a family-based association study approach. To study low penetrant SNPs in PCa susceptibility, 33 SNPs covering PTEN were selected from the whole genome-wide association studies (GWAS) from our available case-control studies in Sweden (Cancer of the Prostate in Sweden (CAPS)) and the publicly available cancer genetic markers of susceptibility (CGEMS) study. Germ-line copy-number variations (CNVs) in PTEN were assessed in CAPS. Co-segregation of germ-line variants and PCa was not observed among HPC families and no significant differences in the allele frequencies were observed in sporadic cases and controls, aggressive and non-aggressive PCa (P>0.05). These results suggest that germ-line variants in PTEN do not have an important role in PCa susceptibility.

Doxorubicin and the ERBB2 targeted therapy, trastuzumab, are routinely used in the treatment of HER2+ breast cancer. In mouse models, doxorubicin is known to cause cardiomyopathy and conditional cardiac knock out of Erbb2 results in dilated cardiomyopathy and increased sensitivity to doxorubicin-induced cell death. In humans, these drugs also result in cardiac phenotypes, but severity and reversibility is highly variable. We examined the association of decline in left ventricular ejection fraction (LVEF) at 15,204 single nucleotide polymorphisms (SNPs) spanning 72 cardiomyopathy genes, in 800 breast cancer patients who received doxorubicin and trastuzumab. For 7033 common SNPs (minor allele frequency (MAF) > 0.01) we performed single marker linear regression. For all SNPs, we performed gene-based testing with SNP-set (Sequence) Kernel Association Tests: SKAT, SKAT-O and SKAT-common/rare under rare variant non-burden; rare variant optimized burden and non-burden tests; and a combination of rare and common variants respectively. Single marker analyses identified seven missense variants in OBSCN (p = 0.0045–0.0009, MAF = 0.18–0.50) and two in TTN (both p = 0.04, MAF = 0.22). Gene-based rare variant analyses, SKAT and SKAT-O, performed very similarly (ILK, TCAP, DSC2, VCL, FXN, DSP and KCNQ1, p = 0.042–0.006). Gene-based tests of rare/common variants were significant at the nominal 5% level for OBSCN as well as TCAP, DSC2, VCL, NEXN, KCNJ2 and DMD (p = 0.044–0.008). Our results suggest that rare and common variants in OBSCN, as well as in other genes, could have modifying effects in cardiomyopathy.

Gene-guided therapy for catheter-ablation of atrial fibrillation: are we there yet?

A sizeable fraction of multiple myeloma (MM) is expected to be explained by heritable factors. Genome-wide association studies (GWAS) have successfully identified a number of common single-nucleotide polymorphisms (SNPs) influencing MM risk. While these SNPs only explain a small proportion of the genetic risk it is unclear how much is left to be detected by other, yet to be identified, common SNPs. Therefore, we applied Genome-Wide Complex Trait Analysis (GCTA) to 2,282 cases and 5,197 controls individuals to estimate the heritability of MM. We estimated that the heritability explained by known common MM risk SNPs identified in GWAS was 2.9% (±2.4%), whereas the heritability explained by all common SNPs was 15.2% (±2.8%). Comparing the heritability explained by the common variants with that from family studies, a fraction of the heritability may be explained by other genetic variants, such as rare variants. In summary, our results suggest that known MM SNPs only explain a small proportion of the heritability and more common SNPs remain to be identified.

The authors conducted a review of the history and empirical basis of genomewide association studies (GWAS), the rationale for GWAS of psychiatric disorders, results to date, limitations, and plans for GWAS meta-analyses. A literature review was carried out, power and other issues discussed, and planned studies assessed. Most of the genomic DNA sequence differences between any two people are common (frequency >5%) single nucleotide polymorphisms (SNPs). Because of localized patterns of correlation (linkage disequilibrium), 500,000 to 1,000,000 of these SNPs can test the hypothesis that one or more common variants explain part of the genetic risk for a disease. GWAS technologies can also detect some of the copy number variants (deletions and duplications) in the genome. Systematic study of rare variants will require large-scale resequencing analyses. GWAS methods have detected a remarkable number of robust genetic associations for dozens of common diseases and traits, leading to new pathophysiological hypotheses, although only small proportions of genetic variance have been explained thus far and therapeutic applications will require substantial further effort. Study design issues, power, and limitations are discussed. For psychiatric disorders, there are initial significant findings for common SNPs and for rare copy number variants, and many other studies are in progress. GWAS of large samples have detected associations of common SNPs and of rare copy number variants with psychiatric disorders. More findings are likely, since larger GWAS samples detect larger numbers of common susceptibility variants, with smaller effects. The Psychiatric GWAS Consortium is conducting GWAS meta-analyses for schizophrenia, bipolar disorder, major depressive disorder, autism, and attention deficit hyperactivity disorder. Based on results for other diseases, larger samples will be required. The contribution of GWAS will depend on the true genetic architecture of each disorder.

A series of reports in 2007 established that single nucleotide polymorphisms (SNPs) in the 9p21.3 locus are associated with coronary heart disease (CHD), both in patients with acute myocardial infarction and chronic atherosclerosis.1,–,4 Subsequently, many additional studies used a candidate locus approach to replicate the 9p21.3-CHD association. Intriguingly, other phenotypes have also been found to be associated with this genomic region, including type II diabetes, stroke, malignant melanoma, aortic aneurysms, cerebral aneurysms, and periodontitis. Unfortunately, neither the causative variants nor genes have been established. Although genetic associations may permit better disease risk stratification, without the knowledge of the genes and variants we lose opportunities to improve our basic understanding of the disease process and hence the potential for targeted therapies. Article see p 445 Many cellular pathways in multiple tissues contribute to the pathogenic processes resulting in CHD. There is value in using intermediate phenotypes as outcomes in genetic association studies because there is enhanced power to detect gene associations when the number of genes potentially responsible for the phenotype is reduced, thereby increasing the fraction of the variance explained by any single factor or gene. In addition, intermediate traits are usually easier to define (have less heterogeneity) than clinical disease. Somewhat surprisingly, 9p21.3 has not been associated with measures believed to represent the chronic process of atherosclerosis or endothelial cell reactivity,5 suggesting that this locus may contribute to CHD by an alternate pathophysiologic mechanism. There is abundant pathological and clinical evidence demonstrating the vital role played by blood platelets in acute coronary syndromes, which result from the formation of occlusive platelet thrombi in coronary arteries at the sites of ruptured atherosclerotic plaques,6,7 and antiplatelet agents have become a mainstay of therapy in patients with acute coronary syndromes. In addition to …

Genome-wide association studies (GWAS) have identified hundreds of common variants associated with alcohol consumption. In contrast, genetic studies of alcohol consumption that use rare variants are still in their early stages. No prior studies of alcohol consumption have examined whether common and rare variants implicate the same genes and molecular networks, leaving open the possibility that the two approaches might identify distinct biology. To address this knowledge gap, we used publicly available alcohol consumption GWAS summary statistics (GSCAN, N = 666,978) and whole exome sequencing data (Genebass, N = 393,099) to identify a set of common and rare variants for alcohol consumption. We used gene-based analysis to implicate genes from common and rare variant analyses, which we then propagated onto a shared molecular network using a network colocalization procedure. Gene-based analysis of each dataset implicated 294 (common variants) and 35 (rare variants) genes, including ethanol metabolizing genes ADH1B and ADH1C, which were identified by both analyses, and ANKRD12, GIGYF1, KIF21B, and STK31, which were identified in only the rare variant analysis, but have been associated with other neuropsychiatric traits. Network colocalization revealed significant network overlap between the genes identified via common and rare variants. The shared network identified gene families that function in alcohol metabolism, including ADH, ALDH, CYP, and UGT. Seventy-one of the genes in the shared network were previously implicated in neuropsychiatric or substance use disorders but not alcohol-related behaviors (e.g. EXOC2, EPM2A, and CACNG4). Differential gene expression analysis showed enrichment in the liver and several brain regions. Genes implicated by network colocalization identify shared biology relevant to alcohol consumption, which also underlie neuropsychiatric traits and substance use disorders that are comorbid with alcohol use, providing a more holistic understanding of two disparate sources of genetic information.

Aggregating information across multiple variants in a gene or region can improve power for rare variant association testing. Power is maximized when the aggregation region contains many causal variants and few neutral variants. In this paper, we present a method for the localization of the association signal in a region using a sliding-window based approach to rare variant association testing in a region. We first introduce a novel method for analysis of rare variants, the Difference in Minor Allele Frequency test (DMAF), which allows combined analysis of common and rare variants, and makes no assumptions about the direction of effects. In whole-region analyses of simulated data with risk and protective variants, DMAF and other methods which pool data across individuals were found to outperform methods which pool data across variants. We then implement a sliding-window version of DMAF, using a step-down permutation approach to control type I error with the testing of multiple windows. In simulations, the sliding-window DMAF improved power to detect a causal sub-region, compared to applying DMAF to the whole region. Sliding-window DMAF was also effective in localizing the causal sub-region. We also applied the DMAF sliding-window approach to test for an association between response to the drug gemcitabine and variants in the gene FKBP5 sequenced in 91 lymphoblastoid cell lines derived from white non-Hispanic individuals. The application of the sliding-window test procedure detected an association in a sub-region spanning an exon and two introns, when rare and common variants were analyzed together.