Power of selective genotyping in genome-wide association studies of quantitative traits

Trait-customized sampling of core collections from a winter wheat genebank collection supports association studies

Subsampling a reduced number of accessions from ex situ genebank collections, known as core collections, is a widely applied method for the investigation of stored genetic diversity and for an exploitation by breeding and research. Optimizing core collections for genome-wide association studies could potentially maximize opportunities to discover relevant and rare variation. In the present study, eight strategies to sample core collections were implemented separately for two traits, namely susceptibility to yellow rust and stem lodging, on about 6,300 accessions of winter wheat (Triticum aestivum L.). Each strategy maximized different parameters or emphasized another aspect of the collection; the strategies relied on genomic data, phenotypic data or a combination thereof. The resulting trait-customized core collections of eight different sizes, covering the range between 100 and 800 accession samples, were analyzed based on characteristics such as population stratification, number of duplicate genotypes and genetic diversity. Furthermore, the statistical power for an association study was investigated as a key criterion for comparisons. While sampling extreme phenotypes boosts the power especially for smaller core collections of up to 500 accession samples, maximization of genetic diversity within the core collection minimizes population stratification and avoids the accumulation of less informative duplicate genotypes when increasing the size of a core collection. Advantages and limitations of different strategies to create trait-customized core collections are discussed for different scenarios of the availability of resources and data.

Association of genetic polymorphisms related to Johne's disease with estimated breeding values of Holstein sires for milk ELISA test scores

BACKGROUND

Johne's disease (JD) is a chronic intestinal inflammatory disease caused by Mycobacterium avium subsp. paratuberculosis (MAP) infection in ruminants. Since there are currently no effective vaccine or treatment options available to control JD, genetic selection may be an alternative strategy to enhance JD resistance. Numerous Single Nucleotide Polymorphisms (SNPs) have been reported to be associated with MAP infection status based on published genome-wide association and candidate gene studies. The main objective of this study was to validate these SNPs that were previously identified to be associated with JD by testing their effect on Holstein bulls' estimated breeding values (EBVs) for milk ELISA test scores, an indirect indicator of MAP infection status in cattle.

RESULTS

Three SNPs, rs41810662, rs41617133 and rs110225854, located on Bos taurus autosomes (BTA) 16, 23 and 26, respectively, were confirmed as significantly associated with Holstein bulls' EBVs for milk ELISA test score (FDR < 0.01) based on General Quasi Likelihood Scoring analysis (GQLS) analysis. Single-SNP regression analysis identified four SNPs that were associated with sire EBVs (FDR < 0.05). This includes two SNPs that were common with GQLS (rs41810662 and rs41617133), with the other two SNPs being rs110494981 and rs136182707, located on BTA9 and BTA16, respectively.

CONCLUSIONS

The findings of this study validate the association of SNPs with JD MAP infection status and highlight the need to further investigate the genomic regions harboring these SNPs.

Genetic variants in the upstream region of activin receptor IIA are associated with female fertility in Japanese Black cattle

Background

Female fertility, a fundamental trait required for animal reproduction, has gradually declined in the last 2 decades in Japanese Black cattle. To identify associated genetic variants in Japanese Black cattle, we evaluated female fertility as a metric to describe the average inverse of the number of artificial inseminations required for conception from the first through the fourth parity (ANAI₄) and conducted a genome-wide association study (GWAS) using 430 animals with extreme ANAI₄ values from 10,399 animals.

Results

We found that 2 variants, namely a single-nucleotide polymorphisms (SNP; g.48476925C > T) and a 3-bp indel (g.48476943_48476946insGGC), in the upstream region of the activin receptor IIA gene (ACVR2A) were associated with ANAI₄. ACVR2A transcripts from Japanese Black cattle of the Q haplotype, defined by the SNP and the 3-bp indel, with increased ANAI₄ were 1.29–1.32-fold more abundant than q-derived transcripts. In agreement, reporter assay results revealed that the activity of the ACVR2A promoter was higher in reporter constructs with the Q haplotype than in those with the q haplotype by approximately 1.2 fold. Expression of exogenous ACVR2A induced dose-dependent increases of reporter activity from the follicle-stimulating hormone, beta polypeptide (FSHB) promoter in response to activin A in a pituitary gonadotrophic cell line. The findings suggested that sequence variations in the upstream region of ACVR2A with the Q haplotype increased ACVR2A transcription, which in turn induced FSHB expression. This association was replicated using a sample population size of 1,433 animals; the frequency of the Q haplotype was 0.39, and Q-to-q haplotype substitution resulted in an increase of 0.02 in terms of ANAI₄.

Conclusions

This GWAS identified variants in the upstream region of ACVR2A, which were associated with female fertility in Japanese Black cattle. The variants affected the level of ACVR2A mRNA expression, which could lead to an allelic imbalance. This association was replicated with a sample population of 1,433 animals. Thus, the results suggest that the Q haplotype could serve as a useful marker to select Japanese Black cattle with superior female fertility.

Electronic supplementary material

The online version of this article (doi:10.1186/s12863-015-0282-0) contains supplementary material, which is available to authorized users.

Identifying genetic risk variants for coronary heart disease in familial hypercholesterolemia: an extreme genetics approach

Mutations in the low-density lipoprotein receptor (LDLR) gene cause familial hypercholesterolemia (FH), a disorder characterized by coronary heart disease (CHD) at young age. We aimed to apply an extreme sampling method to enhance the statistical power to identify novel genetic risk variants for CHD in individuals with FH. We selected cases and controls with an extreme contrast in CHD risk from 17,000 FH patients from the Netherlands, whose functional LDLR mutation was unequivocally established. The genome-wide association (GWA) study was performed on 249 very young FH cases with CHD and 217 old FH controls without CHD (above 65 years for males and 70 years of age for females) using the Illumina HumanHap550K chip. In the next stage, two independent samples (one from the Netherlands and one from Italy, Norway, Spain, and the United Kingdom) of FH patients were used as replication samples. In the initial GWA analysis, we identified 29 independent single nucleotide polymorphisms (SNPs) with suggestive associations with premature CHD (P<1 × 10(-4)). We examined the association of these SNPs with CHD risk in the replication samples. After Bonferroni correction, none of the SNPs either replicated or reached genome-wide significance after combining the discovery and replication samples. Therefore, we conclude that the genetics of CHD risk in FH is complex and even applying an 'extreme genetics' approach we did not identify new genetic risk variants. Most likely, this method is not as effective in leveraging effect size as anticipated, and may, therefore, not lead to significant gains in statistical power.

Using genetic prediction from known complex disease Loci to guide the design of next-generation sequencing experiments

A central focus of complex disease genetics after genome-wide association studies (GWAS) is to identify low frequency and rare risk variants, which may account for an important fraction of disease heritability unexplained by GWAS. A profusion of studies using next-generation sequencing are seeking such risk alleles. We describe how already-known complex trait loci (largely from GWAS) can be used to guide the design of these new studies by selecting cases, controls, or families who are most likely to harbor undiscovered risk alleles. We show that genetic risk prediction can select unrelated cases from large cohorts who are enriched for unknown risk factors, or multiply-affected families that are more likely to harbor high-penetrance risk alleles. We derive the frequency of an undiscovered risk allele in selected cases and controls, and show how this relates to the variance explained by the risk score, the disease prevalence and the population frequency of the risk allele. We also describe a new method for informing the design of sequencing studies using genetic risk prediction in large partially-genotyped families using an extension of the Inside-Outside algorithm for inference on trees. We explore several study design scenarios using both simulated and real data, and show that in many cases genetic risk prediction can provide significant increases in power to detect low-frequency and rare risk alleles. The same approach can also be used to aid discovery of non-genetic risk factors, suggesting possible future utility of genetic risk prediction in conventional epidemiology. Software implementing the methods in this paper is available in the R package Mangrove.

Genomic, Transcriptomic, and Lipidomic Profiling Highlights the Role of Inflammation in Individuals With Low High-density Lipoprotein Cholesterol

Objective—

Low high-density lipoprotein cholesterol (HDL-C) is associated with cardiometabolic pathologies. In this study, we investigate the biological pathways and individual genes behind low HDL-C by integrating results from 3 high-throughput data sources: adipose tissue transcriptomics, HDL lipidomics, and dense marker genotypes from Finnish individuals with low or high HDL-C (n=450).

Approach and Results—

In the pathway analysis of genetic data, we demonstrate that genetic variants within inflammatory pathways were enriched among low HDL-C associated single-nucleotide polymorphisms, and the expression of these pathways upregulated in the adipose tissue of low HDL-C subjects. The lipidomic analysis highlighted the change in HDL particle quality toward putatively more inflammatory and less vasoprotective state in subjects with low HDL-C, as evidenced by their decreased antioxidative plasmalogen contents. We show that the focal point of these inflammatory pathways seems to be the HLA region with its low HDL-associated alleles also associating with more abundant local transcript levels in adipose tissue, increased plasma vascular cell adhesion molecule 1 (VCAM1) levels, and decreased HDL particle plasmalogen contents, markers of adipose tissue inflammation, vascular inflammation, and HDL antioxidative potential, respectively. In a population-based look-up of the inflammatory pathway single-nucleotide polymorphisms in a large Finnish cohorts (n=11 211), no association of the HLA region was detected for HDL-C as quantitative trait, but with extreme HDL-C phenotypes, implying the presence of low or high HDL genes in addition to the population-genomewide association studies–identified HDL genes.

Conclusions—

Our study highlights the role of inflammation with a genetic component in subjects with low HDL-C and identifies novel cis -expression quantitative trait loci ( cis -eQTL) variants in HLA region to be associated with low HDL-C.

Joint analysis of binary and quantitative traits with data sharing and outcome-dependent sampling

We study the analysis of a joint association between a genetic marker with both binary (case-control) and quantitative (continuous) traits, where the quantitative trait values are only available for the cases due to data sharing and outcome-dependent sampling. Data sharing becomes common in genetic association studies, and the outcome-dependent sampling is the consequence of data sharing, under which a phenotype of interest is not measured for some subgroup. The trend test (or Pearson's test) and F-test are often, respectively, used to analyze the binary and quantitative traits. Because of the outcome-dependent sampling, the usual F-test can be applied using the subgroup with the observed quantitative traits. We propose a modified F-test by also incorporating the genotype frequencies of the subgroup whose traits are not observed. Further, a combination of this modified F-test and Pearson's test is proposed by Fisher's combination of their P-values as a joint analysis. Because of the correlation of the two analyses, we propose to use a Gamma (scaled chi-squared) distribution to fit the asymptotic null distribution for the joint analysis. The proposed modified F-test and the joint analysis can also be applied to test single trait association (either binary or quantitative trait). Through simulations, we identify the situations under which the proposed tests are more powerful than the existing ones. Application to a real dataset of rheumatoid arthritis is presented.

Using extreme phenotype sampling to identify the rare causal variants of quantitative traits in association studies

Variants identified in recent genome-wide association studies based on the common-disease common-variant hypothesis are far from fully explaining the hereditability of complex traits. Rare variants may, in part, explain some of the missing hereditability. Here, we explored the advantage of the extreme phenotype sampling in rare-variant analysis and refined this design framework for future large-scale association studies on quantitative traits. We first proposed a power calculation approach for a likelihood-based analysis method. We then used this approach to demonstrate the potential advantages of extreme phenotype sampling for rare variants. Next, we discussed how this design can influence future sequencing-based association studies from a cost-efficiency (with the phenotyping cost included) perspective. Moreover, we discussed the potential of a two-stage design with the extreme sample as the first stage and the remaining nonextreme subjects as the second stage. We demonstrated that this two-stage design is a cost-efficient alternative to the one-stage cross-sectional design or traditional two-stage design. We then discussed the analysis strategies for this extreme two-stage design and proposed a corresponding design optimization procedure. To address many practical concerns, for example measurement error or phenotypic heterogeneity at the very extremes, we examined an approach in which individuals with very extreme phenotypes are discarded. We demonstrated that even with a substantial proportion of these extreme individuals discarded, an extreme-based sampling can still be more efficient. Finally, we expanded the current analysis and design framework to accommodate the CMC approach where multiple rare-variants in the same gene region are analyzed jointly.

A simple bias correction in linear regression for quantitative trait association under two-tail extreme selection

Selective genotyping can increase power in quantitative trait association. One example of selective genotyping is two-tail extreme selection, but simple linear regression analysis gives a biased genetic effect estimate. Here, we present a simple correction for the bias.

Polymorphisms in Anopheles gambiae immune genes associated with natural resistance to Plasmodium falciparum

Many genes involved in the immune response of Anopheles gambiae, the main malaria vector in Africa, have been identified, but whether naturally occurring polymorphisms in these genes underlie variation in resistance to the human malaria parasite, Plasmodium falciparum, is currently unknown. Here we carried out a candidate gene association study to identify single nucleotide polymorphisms (SNPs) associated with natural resistance to P. falciparum. A. gambiae M form mosquitoes from Cameroon were experimentally challenged with three local wild P. falciparum isolates. Statistical associations were assessed between 157 SNPs selected from a set of 67 A. gambiae immune-related genes and the level of infection. Isolate-specific associations were accounted for by including the effect of the isolate in the analysis. Five SNPs were significantly associated to the infection phenotype, located within or upstream of AgMDL1, CEC1, Sp PPO activate, Sp SNAKElike, and TOLL6. Low overall and local linkage disequilibrium indicated high specificity in the loci found. Association between infection phenotype and two SNPs was isolate-specific, providing the first evidence of vector genotype by parasite isolate interactions at the molecular level. Four SNPs were associated to either oocyst presence or load, indicating that the genetic basis of infection prevalence and intensity may differ. The validity of the approach was verified by confirming the functional role of Sp SNAKElike in gene silencing assays. These results strongly support the role of genetic variation within or near these five A. gambiae immune genes, in concert with other genes, in natural resistance to P. falciparum. They emphasize the need to distinguish between infection prevalence and intensity and to account for the genetic specificity of vector-parasite interactions in dissecting the genetic basis of Anopheles resistance to human malaria.

Anopheles gambiae is the main malaria vector in Africa, transmitting the parasite when it blood feeds on human hosts. The parasite undergoes several developmental stages in the mosquito to complete its life cycle, during which time it is confronted by the mosquito's immune system. The resistance of mosquitoes to malaria infection is highly variable in wild populations and is known to be under strong genetic control, but to date the specific genes responsible for this variation remain to be identified. The present study uncovers variations in A. gambiae immune genes that are associated with natural resistance to Plasmodium falciparum, the deadliest human malaria parasite. The association of some mosquito genetic loci with the level of infection depended on the P. falciparum isolate, suggesting that resistance is determined by interactions between the genome of the mosquito and that of the parasite. This finding highlights the need to account for the natural genetic diversity of malaria parasites in future research on vector-parasite interactions. The loci uncovered in this study are potential targets for developing novel malaria control strategies based on natural mosquito resistance mechanisms.

Genome-wide association analyses of quantitative traits: the GAW16 experience

The group that formed on the theme of genome-wide association analyses of quantitative traits (Group 2) in the Genetic Analysis Workshop 16 comprised eight sets of investigators. Three data sets were available: one on autoantibodies related to rheumatoid arthritis provided by the North American Rheumatoid Arthritis Consortium; the second on anthropometric, lipid, and biochemical measures provided by the Framingham Heart Study (FHS); and the third a simulated data set modeled after FHS. The different investigators in the group addressed a large set of statistical challenges and applied a wide spectrum of association methods in analyzing quantitative traits at the genome-wide level. While some previously reported genes were validated, some novel chromosomal regions provided significant evidence of association in multiple contributions in the group. In this report, we discuss the different strategies explored by the different investigators with the common goal of improving the power to detect association.