Practical Issues in Imputation-Based Association Mapping

Imputation-based association methods provide a powerful framework for testing untyped variants for association with phenotypes and for combining results from multiple studies that use different genotyping platforms. Here, we consider several issues that arise when applying these methods in practice, including: (i) factors affecting imputation accuracy, including choice of reference panel; (ii) the effects of imputation accuracy on power to detect associations; (iii) the relative merits of Bayesian and frequentist approaches to testing imputed genotypes for association with phenotype; and (iv) how to quickly and accurately compute Bayes factors for testing imputed SNPs. We find that imputation-based methods can be robust to imputation accuracy and can improve power to detect associations, even when average imputation accuracy is poor. We explain how ranking SNPs for association by a standard likelihood ratio test gives the same results as a Bayesian procedure that uses an unnatural prior assumption—specifically, that difficult-to-impute SNPs tend to have larger effects—and assess the power gained from using a Bayesian approach that does not make this assumption. Within the Bayesian framework, we find that good approximations to a full analysis can be achieved by simply replacing unknown genotypes with a point estimate—their posterior mean. This approximation considerably reduces computational expense compared with published sampling-based approaches, and the methods we present are practical on a genome-wide scale with very modest computational resources (e.g., a single desktop computer). The approximation also facilitates combining information across studies, using only summary data for each SNP. Methods discussed here are implemented in the software package BIMBAM, which is available from http://stephenslab.uchicago.edu/software.html.     

Additive Antiatherogenic Effects of CETP rs708272 on Serum LDL Subfraction Levels in Patients with CHD Under Statin Therapy

Recently, subfraction analysis of serum low density lipoprotein (LDL) is considered to be a better predictor of the risk of coronary heart disease (CHD) compared to the other lipid parameters. The aim of this study was to examine the effects of the HDL-associated Taq1B (rs708272) SNP of cholesterol ester transfer protein (CETP) gene on serum LDL subfractions in patients with CHD. Serum lipid levels were measured enzymatically and LDL subfraction analysis was carried out by the Lipoprint System (Quantimetrix, CA, USA). The CETP rs708272 SNP was studied in 66 healthy controls and 79 patients with CHD receiving statin therapy by the PCR–RFLP technique. The CHD patients had elevated antiatherogenic LDL-1 subfraction (p = 0.042), decreased atherogenic IDL-C subfraction (p = 0.023), and total IDL (p = 0.030) levels compared to the healthy controls. The CETP rs708272 Taq1B minor B2 allele was associated with increased levels of antiatherogenic LDL-1 (B2: 0.40 ± 0.20 vs. B1B1: 0.25 ± 0.08, p = 0.004) and large-LDL (LDL 1–2) subfractions in the CHD group (B2 allele: 0.68 ± 0.41 vs. B1B1: 0.42 ± 0.20; p < 0.05), while it was associated with reduced levels of the large-LDL subfraction in healthy subjects (B2 allele: 0.29 ± 0.14 vs. B1B1: 0.54 ± 0.24; p = 0.017). However, there was no statistically significant association between the CETP rs708272 SNP and small dense LDL subfraction (LDL 3–7) and lipoprotein levels (p > 0.05). Our findings have indicated that the CETP rs708272 SNP together with statin therapy may show a favorable effect on antiatherogenic LDL-1 and large-LDL subfractions in CHD patients with an atherogenic effect on large-LDL subfraction in healthy subjects. Based on these results, it can be concluded that the effects of the CETP variation on LDL subfraction could change in cardiometabolic events such as CHD and statin therapy.     

Genome-wide association study of genetic determinants of LDL-c response to atorvastatin therapy: importance of Lp(a)

We carried out a genome-wide association study (GWAS) of LDL-c response to statin using data from participants in the Collaborative Atorvastatin Diabetes Study (CARDS; n = 1,156), the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT; n = 895), and the observational phase of ASCOT (n = 651), all of whom were prescribed atorvastatin 10 mg. Following genome-wide imputation, we combined data from the three studies in a meta-analysis. We found associations of LDL-c response to atorvastatin that reached genome-wide significance at rs10455872 (P = 6.13 × 10(-9)) within the LPA gene and at two single nucleotide polymorphisms (SNP) within the APOE region (rs445925; P = 2.22 × 10(-16) and rs4420638; P = 1.01 × 10(-11)) that are proxies for the ε2 and ε4 variants, respectively, in APOE. The novel association with the LPA SNP was replicated in the PROspective Study of Pravastatin in the Elderly at Risk (PROSPER) trial (P = 0.009). Using CARDS data, we further showed that atorvastatin therapy did not alter lipoprotein(a) [Lp(a)] and that Lp(a) levels accounted for all of the associations of SNPs in the LPA gene and the apparent LDL-c response levels. However, statin therapy had a similar effect in reducing cardiovascular disease (CVD) in patients in the top quartile for serum Lp(a) levels (HR = 0.60) compared with those in the lower three quartiles (HR = 0.66; P = 0.8 for interaction). The data emphasize that high Lp(a) levels affect the measurement of LDL-c and the clinical estimation of LDL-c response. Therefore, an apparently lower LDL-c response to statin therapy may indicate a need for measurement of Lp(a). However, statin therapy seems beneficial even in those with high Lp(a).     

Imputation-based analysis of association studies: candidate regions and quantitative traits

We introduce a new framework for the analysis of association studies, designed to allow untyped variants to be more effectively and directly tested for association with a phenotype. The idea is to combine knowledge on patterns of correlation among SNPs (e.g., from the International HapMap project or resequencing data in a candidate region of interest) with genotype data at tag SNPs collected on a phenotyped study sample, to estimate ("impute") unmeasured genotypes, and then assess association between the phenotype and these estimated genotypes. Compared with standard single-SNP tests, this approach results in increased power to detect association, even in cases in which the causal variant is typed, with the greatest gain occurring when multiple causal variants are present. It also provides more interpretable explanations for observed associations, including assessing, for each SNP, the strength of the evidence that it (rather than another correlated SNP) is causal. Although we focus on association studies with quantitative phenotype and a relatively restricted region (e.g., a candidate gene), the framework is applicable and computationally practical for whole genome association studies. Methods described here are implemented in a software package, Bim-Bam, available from the Stephens Lab website http://stephenslab.uchicago.edu/software.html.     

Association of CETP TaqI and APOE polymorphisms with type II diabetes mellitus in North Indians: a case control study

Background

Genetic variants of proteins involved in lipid metabolism may play an important role in determining the susceptibility for complications associated with type II diabetes mellitus (T2DM). Goal of the present study was to determine the association of cholesteryl ester transfer protein TaqI B, D442G, and APOE Hha I polymorphisms with T2DM and its complications.

Methods

Study subjects were 136 patients and 264 healthy controls. All polymorphisms were detected using PCR-RFLP and statistical analysis done with χ² test and ANOVA.

Results

Although CETP TaqI B polymorphism was not associated with the T2DM, yet B1B2 genotype was significantly (p = 0.028) associated with high risk of hypertension in diabetic patients (OR = 3.068, 95% CI 1.183–7.958). In North Indians D442G variation in CETP gene was found to be absent. Frequency of APOE HhaI polymorphism was also not different between patients and controls. In diabetic patients having neuropathy and retinopathy significantly different levels of total-cholesterol [(p = 0.001) and (p = 0.029) respectively] and LDL-cholesterol [(p = 0.001) and (p = 0.001) respectively] were observed when compared to patients with T2DM only. However, lipid levels did not show any correlation with the CETP TaqI B and APOE Hha I genetic polymorphisms.

Conclusion

CETP TaqI B and APOE HhaI polymorphism may not be associated with type II diabetes mellitus in North Indian population, however CETP TaqI B polymorphism may be associated with hypertension along with T2DM.     

Heritability of submaximal exercise heart rate response to exercise training is accounted for by nine SNPs

Endurance training-induced changes in hemodynamic traits are heritable. However, few genes associated with heart rate training responses have been identified. The purpose of our study was to perform a genome-wide association study to uncover DNA sequence variants associated with submaximal exercise heart rate training responses in the HERITAGE Family Study. Heart rate was measured during steady-state exercise at 50 W (HR50) on 2 separate days before and after a 20-wk endurance training program in 483 white subjects from 99 families. Illumina HumanCNV370-Quad v3.0 BeadChips were genotyped using the Illumina BeadStation 500GX platform. After quality control procedures, 320,000 single-nucleotide polymorphisms (SNPs) were available for the genome-wide association study analyses, which were performed using the MERLIN software package (single-SNP analyses and conditional heritability tests) and standard regression models (multivariate analyses). The strongest associations for HR50 training response adjusted for age, sex, body mass index, and baseline HR50 were detected with SNPs at the YWHAQ locus on chromosome 2p25 (P = 8.1 × 10(-7)), the RBPMS locus on chromosome 8p12 (P = 3.8 × 10(-6)), and the CREB1 locus on chromosome 2q34 (P = 1.6 × 10(-5)). In addition, 37 other SNPs showed P values <9.9 × 10(-5). After removal of redundant SNPs, the 10 most significant SNPs explained 35.9% of the ΔHR50 variance in a multivariate regression model. Conditional heritability tests showed that nine of these SNPs (all intragenic) accounted for 100% of the ΔHR50 heritability. Our results indicate that SNPs in nine genes related to cardiomyocyte and neuronal functions, as well as cardiac memory formation, fully account for the heritability of the submaximal heart rate training response.     

MRFalign: Protein Homology Detection through Alignment of Markov Random Fields

Sequence-based protein homology detection has been extensively studied and so far the most sensitive method is based upon comparison of protein sequence profiles, which are derived from multiple sequence alignment (MSA) of sequence homologs in a protein family. A sequence profile is usually represented as a position-specific scoring matrix (PSSM) or an HMM (Hidden Markov Model) and accordingly PSSM-PSSM or HMM-HMM comparison is used for homolog detection. This paper presents a new homology detection method MRFalign, consisting of three key components: 1) a Markov Random Fields (MRF) representation of a protein family; 2) a scoring function measuring similarity of two MRFs; and 3) an efficient ADMM (Alternating Direction Method of Multipliers) algorithm aligning two MRFs. Compared to HMM that can only model very short-range residue correlation, MRFs can model long-range residue interaction pattern and thus, encode information for the global 3D structure of a protein family. Consequently, MRF-MRF comparison for remote homology detection shall be much more sensitive than HMM-HMM or PSSM-PSSM comparison. Experiments confirm that MRFalign outperforms several popular HMM or PSSM-based methods in terms of both alignment accuracy and remote homology detection and that MRFalign works particularly well for mainly beta proteins. For example, tested on the benchmark SCOP40 (8353 proteins) for homology detection, PSSM-PSSM and HMM-HMM succeed on 48% and 52% of proteins, respectively, at superfamily level, and on 15% and 27% of proteins, respectively, at fold level. In contrast, MRFalign succeeds on 57.3% and 42.5% of proteins at superfamily and fold level, respectively. This study implies that long-range residue interaction patterns are very helpful for sequence-based homology detection. The software is available for download at http://raptorx.uchicago.edu/download/. A summary of this paper appears in the proceedings of the RECOMB 2014 conference, April 2–5.     

Electronegative LDLs from familial hypercholesterolemic patients are physicochemically heterogeneous but uniformly proapoptotic

A highly electronegative fraction of human plasma LDLs, designated L5, has distinctive biological activity that includes induction of apoptosis in bovine aortic endothelial cells (BAECs). This study was performed to identify a relationship between LDL density, electronegativity, and biological activity, namely, the induction of apoptosis in BAECs. Plasma LDLs from normolipidemic subjects and homozygotic familial hypercholesterolemia subjects were separated into five subfractions, with increasing electronegativity from L1 to L5, and into seven subfractions according to increasing density, D1 to D7. L1 to L5 were also separated according to density, and D1 to D7 were separated according to charge. The density profiles of L1 to L5 were similar (maximum density = 1.030 +/- 0.002 g/ml). Induction of apoptosis by all seven density subfractions was confined to the highly electronegative fraction, L5, and within each density subfraction the magnitude of apoptosis correlated with the L5 content. Electronegative LDL is heterogeneous with respect to density and composition, and induction of apoptosis is more strongly associated with LDL electronegativity than with LDL size or density.     

Exploiting SNP Correlations within Random Forest for Genome-Wide Association Studies

The primary goal of genome-wide association studies (GWAS) is to discover variants that could lead, in isolation or in combination, to a particular trait or disease. Standard approaches to GWAS, however, are usually based on univariate hypothesis tests and therefore can account neither for correlations due to linkage disequilibrium nor for combinations of several markers. To discover and leverage such potential multivariate interactions, we propose in this work an extension of the Random Forest algorithm tailored for structured GWAS data. In terms of risk prediction, we show empirically on several GWAS datasets that the proposed T-Trees method significantly outperforms both the original Random Forest algorithm and standard linear models, thereby suggesting the actual existence of multivariate non-linear effects due to the combinations of several SNPs. We also demonstrate that variable importances as derived from our method can help identify relevant loci. Finally, we highlight the strong impact that quality control procedures may have, both in terms of predictive power and loci identification. Variable importance results and T-Trees source code are all available at www.montefiore.ulg.ac.be/~botta/ttrees/ and github.com/0asa/TTree-source respectively.     

Fine Mapping Causal Variants with an Approximate Bayesian Method Using Marginal Test Statistics

Two recently developed fine-mapping methods, CAVIAR and PAINTOR, demonstrate better performance over other fine-mapping methods. They also have the advantage of using only the marginal test statistics and the correlation among SNPs. Both methods leverage the fact that the marginal test statistics asymptotically follow a multivariate normal distribution and are likelihood based. However, their relationship with Bayesian fine mapping, such as BIMBAM, is not clear. In this study, we first show that CAVIAR and BIMBAM are actually approximately equivalent to each other. This leads to a fine-mapping method using marginal test statistics in the Bayesian framework, which we call CAVIAR Bayes factor (CAVIARBF). Another advantage of the Bayesian framework is that it can answer both association and fine-mapping questions. We also used simulations to compare CAVIARBF with other methods under different numbers of causal variants. The results showed that both CAVIARBF and BIMBAM have better performance than PAINTOR and other methods. Compared to BIMBAM, CAVIARBF has the advantage of using only marginal test statistics and takes about one-quarter to one-fifth of the running time. We applied different methods on two independent cohorts of the same phenotype. Results showed that CAVIARBF, BIMBAM, and PAINTOR selected the same top 3 SNPs; however, CAVIARBF and BIMBAM had better consistency in selecting the top 10 ranked SNPs between the two cohorts. Software is available at https://bitbucket.org/Wenan/caviarbf.     

Cholesterol ester transfer protein gene is associated with high-density lipoprotein cholesterol levels in Korean population

High-density lipoprotein (HDL) cholesterol levels are associated with decreased risk of coronary artery disease. Several genome-wide association studies (GWAS) for HDL cholesterol levels have implicated cholesterol ester transfer protein (CETP) as possibly causal. We tested for the association between single nucleotide polymorphisms (SNPs) in CETP gene and HDL cholesterol levels in Korean population. A total of 979 subjects in Seoul City were genotyped using a genome-wide marker panel for a discovery study. Another 2,277 subjects in Bundang-Gu in Korea were used for a replication study with selected markers. In the discovery phase, the top SNP associated with mean HDL cholesterol levels was rs6499861 in the CETP gene on chromosome 16 (p=1.18×10^?6 in the Seoul City sample, p=8.91×10^?3 in the Bundang-Gu sample). Another SNP (rs6499863) in the CETP gene was also among the top five SNPs associated with HDL cholesterol levels (p=3.83×10^?5 in the Seoul City sample, p=3.29×10^?3 in the Bundang-Gu sample). SNP rs1800775 was also associated with HDL cholesterol levels (p=4.86×10^?4 in meta-analysis results of 3256 samples). This study clearly demonstrates that genetic variants in CETP influence HDL cholesterol levels in Korean adults.     

Gene-environment interactions of CETP gene variation in a high cardiovascular risk Mediterranean population

Genome-wide association studies show that cholesteryl ester transfer protein (CETP) single nucleotide polymorphisms (SNPs) are more strongly associated with HDL cholesterol (HDL-C) concentrations than any other loci across the genome. However, gene-environment interactions for clinical applications are still largely unknown. We studied gene-environment interactions between CETP SNPs and dietary fat intake, adherence to the Mediterranean diet, alcohol consumption, smoking, obesity, and diabetes on HDL-C in 4,210 high cardiovascular risk subjects from a Mediterranean population. We focused on the −4,502C>T and the TaqIB SNPs in partial linkage disequilibrium (D''= 0.88; P < 0.001). They were independently associated with higher HDL-C (P < 0.001); this clinically relevant association was greater when their diplotype was considered (14% higher in TT/B2B2 vs. CC/B1B1). No gene-gene interaction was observed. We also analyzed the association of these SNPs with blood pressure, and no clinically relevant associations were detected. No statistically significant interactions of these SNPs with obesity, diabetes, and smoking in determining HDL-C concentrations were found. Likewise, alcohol, dietary fat, and adherence to the Mediterranean diet did not statistically interact with the CETP variants (independently or as diplotype) in determining HDL-C. In conclusion, the strong association of the CETP SNPs and HDL-C was not statistically modified by diet or by the other environmental factors.     

Statistical Power of Model Selection Strategies for Genome-Wide Association Studies

Genome-wide association studies (GWAS) aim to identify genetic variants related to diseases by examining the associations between phenotypes and hundreds of thousands of genotyped markers. Because many genes are potentially involved in common diseases and a large number of markers are analyzed, it is crucial to devise an effective strategy to identify truly associated variants that have individual and/or interactive effects, while controlling false positives at the desired level. Although a number of model selection methods have been proposed in the literature, including marginal search, exhaustive search, and forward search, their relative performance has only been evaluated through limited simulations due to the lack of an analytical approach to calculating the power of these methods. This article develops a novel statistical approach for power calculation, derives accurate formulas for the power of different model selection strategies, and then uses the formulas to evaluate and compare these strategies in genetic model spaces. In contrast to previous studies, our theoretical framework allows for random genotypes, correlations among test statistics, and a false-positive control based on GWAS practice. After the accuracy of our analytical results is validated through simulations, they are utilized to systematically evaluate and compare the performance of these strategies in a wide class of genetic models. For a specific genetic model, our results clearly reveal how different factors, such as effect size, allele frequency, and interaction, jointly affect the statistical power of each strategy. An example is provided for the application of our approach to empirical research. The statistical approach used in our derivations is general and can be employed to address the model selection problems in other random predictor settings. We have developed an R package markerSearchPower to implement our formulas, which can be downloaded from the Comprehensive R Archive Network (CRAN) or http://bioinformatics.med.yale.edu/group/.     

Linkage disequilibrium-based quality control for large-scale genetic studies

Quality control (QC) is a critical step in large-scale studies of genetic variation. While, on average, high-throughput single nucleotide polymorphism (SNP) genotyping assays are now very accurate, the errors that remain tend to cluster into a small percentage of "problem" SNPs, which exhibit unusually high error rates. Because most large-scale studies of genetic variation are searching for phenomena that are rare (e.g., SNPs associated with a phenotype), even this small percentage of problem SNPs can cause important practical problems. Here we describe and illustrate how patterns of linkage disequilibrium (LD) can be used to improve QC in large-scale, population-based studies. This approach has the advantage over existing filters (e.g., HWE or call rate) that it can actually reduce genotyping error rates by automatically correcting some genotyping errors. Applying this LD-based QC procedure to data from The International HapMap Project, we identify over 1,500 SNPs that likely have high error rates in the CHB and JPT samples and estimate corrected genotypes. Our method is implemented in the software package fastPHASE, available from the Stephens Lab website (http://stephenslab.uchicago.edu/software.html).     

Association of cholesteryl ester transfer protein (TaqIB) and apolipoprotein E (HhaI) gene variants with obesity

Background The pathophysiology of obesity is known to be influenced by alterations in lipid levels. We aimed to evaluate association of cholesteryl ester transfer protein (CETP) and apolipoprotein (APO) E gene variants with asymptomatic obesity. Methods A total of 437 subjects, 159 asymptomatic obese (BMI = 29.29 +/- 3.76) and 278 non-obese (BMI = 23.38 +/- 1.71) individuals, were included in this case-control study. Lipid levels were estimated using standard protocols. Analysis of CETP (TaqIB) and APOE (HhaI) gene polymorphisms was done using PCR-RFLP. Results We found significant difference in blood pressure (systolic, P < 0.0001 and diastolic, P < 0.0001), total cholesterol (P < 0.0001), LDL-cholesterol (P < 0.0001), and HDL-cholesterol (P < 0.0001) in obese as compared to non-obese group. Homozygous APO E4E4 genotype was only observed in 5.7% of obese individuals and none in non-obese group. APO E4 allele carriers were also susceptible for obesity (P = 0.016, OR = 1.73; 95% CI = 1.12-2.68) than non-carriers. Higher blood pressure (Systolic, P = 0.001 and Diastolic, P = 0.004) and triglyceride levels (P = 0.029) were observed in obese subjects with APO E4 allele than individuals without APO E4. However, CETP B1 variant allele carriers did not show alteration in blood pressure and lipid profile in asymptomatic obese subjects. Conclusions APO E4 genotype and allele were found to be associated with asymptomatic obesity, whereas CETP Taq1B polymorphism showed no such association in North Indian subjects.     

Cholesteryl ester transfer protein gene haplotypes, plasma high-density lipoprotein levels and the risk of coronary heart disease

High-density lipoprotein cholesterol (HDL-C) is a known inverse predictor of coronary heart disease (CHD) and is thus a potential therapeutic target. Cholesteryl ester transfer protein (CETP) is a key protein in HDL-C metabolism such that elevated CETP activity is associated with lower HDL-C. Currently available HDL-C raising drugs are relatively ineffective and evidence suggesting the role of CETP in HDL-C levels has promoted the development of CETP inhibitors as potential therapeutic agents for CHD. We investigated three SNPs in the CETP gene in two cross-sectional community-based populations (n = 1,574 and 1,109) and a population of 556 CHD patients to determine if reduced CETP activity due to genetic variations in the CETP gene would increase HDL-C levels and reduce the risk of CHD. CETP genotypes and haplotypes were tested for association with lipid levels, CETP activity and risk of CHD. Multivariate analysis showed the common AAB2 haplotype defined by the G-2708A, C-629A and TaqIB polymorphisms, was consistently associated with reduced CETP activity and increased HDL-C levels. A mean increase in HDL-C levels of 0.16–0.24 mmol/l was observed in individuals with two copies of the AAB2 haplotype relative to non AAB2 carriers across all three populations (P < 0.001). A case-control study of males indicated no association between single SNPs or haplotypes and the risk of CHD. These results suggest that raising HDL-C via CETP inhibition may not alter risk of CHD. Randomized control trials are needed to determine whether CETP inhibition will in reality reduce risk of CHD by raising HDL-C. Pamela A. McCaskie and John P. Beilby contributed equally to this work.     

Genetic variants in the KIF6 region and coronary event reduction from statin therapy

A single nucleotide polymorphism (SNP) in KIF6, a member of the KIF9 family of kinesins, is associated with differential coronary event reduction from statin therapy in four randomized controlled trials; this SNP (rs20455) is also associated with the risk for coronary heart disease (CHD) in multiple prospective studies. We investigated whether other common SNPs in the KIF6 region were associated with event reduction from statin therapy. Of the 170 SNPs in the KIF6 region investigated in the Cholesterol and Recurrent Events trial (CARE), 28 were associated with differential event reduction from statin therapy (P (interaction) < 01 in Caucasians, adjusted for age and sex) and were further investigated in the Pravastatin or Atorvastatin Evaluation and Infection Therapy-Thrombolysis In Myocardial Infarction 22 (PROVE IT-TIMI22) and West of Scotland Coronary Prevention Study (WOSCOPS). These analyses revealed that two SNPs (rs9462535 and rs9471077), in addition to rs20455, were associated with event reduction from statin therapy (P (interaction) < 0.1 in each of the three studies). The relative risk reduction ranged from 37 to 50% (P < 0.01) in carriers of the minor alleles of these SNPs and from -4 to 13% (P > 0.4) in non-carriers. These three SNPs are in high linkage disequilibrium with one another (r (2) > 0.84). Functional studies of these variants may help to understand the role of KIF6 in the pathogenesis of CHD and differential response to statin therapy.