Efficient association study design via power-optimized tag SNP selection

Vitamin D Receptor Polymorphism and Myasthenia Gravis in Chinese Han Population

Myasthenia gravis (MG) is an autoimmune disease in which antibodies bind to acetylcholine receptors (AChR) or other functional molecules in the postsynaptic membrane at the neuromuscular junction. Vitamin D (VD) has a number of pluripotent effects, which include immune-regulation and bone metabolism. The immunomodulatory actions of 1,25(OH)2D3 are mediated by its binding to a vitamin D receptor (VDR). In the study, we undertook a case-control study to explore the association between VDR gene polymorphism and the susceptibility and severity of MG patients. Four hundred and eighty MG patients and 487 healthy controls were included and gene polymorphisms of VDR were determined with improved multiplex ligation detection reaction technique and SNPscan^TM technique. MG patients were classified into subgroups by essential clinical features and by a comprehensive classification. The frequencies of alleles and genotypes were compared between the MG group and the control group, between each MG subgroup and the control group, and between each pair of MG subgroups. There were no significant differences in frequencies of alleles and genotypes between MG patients and healthy controls, between MG subgroups and healthy controls, or between each pair of MG subgroups in the analysis of subgroups classified by essential clinical features (onset age, gender, thymoma, AChRAb positivity, onset involvement) and the maximal severity (modified Oosterhuis score). In the analysis of subgroups with a comprehensive classification, the frequencies of alleles and genotypes in rs731236 showed significant differences between adult non-thymoma AChRAb negative MG subgroup and the control group, as well as the adult non-thymoma AChRAb positive MG group. In the Chinese Han population, rs731236 was found to be possibly associated with adult non-thymoma AChRAb negative MG patients, although this needs further confirmation.

The Role of Osteopontin and Its Gene on Glucocorticoid Response in Myasthenia Gravis

Biomarkers that assess treatment response for patients with the autoimmune disorder, myasthenia gravis (MG), have not been evaluated to a significant extent. We hypothesized the pro-inflammatory cytokine, osteopontin (OPN), may be associated with variability of response to glucocorticoids (GCs) in patients with MG. A cohort of 250 MG patients treated with standardized protocol of GCs was recruited, and plasma OPN and polymorphisms of its gene, secreted phosphoprotein 1 (SPP1), were evaluated. Mean OPN levels were higher in patients compared to healthy controls. Carriers of rs11728697*T allele (allele definition: one of two or more alternative forms of a gene) were more frequent in the poorly GC responsive group compared to the GC responsive group indicating an association of rs11728697*T allele with GC non-responsiveness. One risk haplotype (AGTACT) was identified associated with GC non-responsiveness compared with GC responsive MG group. Genetic variations of SPP1 were found associated with the response to GC among MG patients.

FAPI: Fast and accurate P-value Imputation for genome-wide association study

Imputing individual-level genotypes (or genotype imputation) is now a standard procedure in genome-wide association studies (GWAS) to examine disease associations at untyped common genetic variants. Meta-analysis of publicly available GWAS summary statistics can allow more disease-associated loci to be discovered, but these data are usually provided for various variant sets. Thus imputing these summary statistics of different variant sets into a common reference panel for meta-analyses is impossible using traditional genotype imputation methods. Here we develop a fast and accurate P-value imputation (FAPI) method that utilizes summary statistics of common variants only. Its computational cost is linear with the number of untyped variants and has similar accuracy compared with IMPUTE2 with prephasing, one of the leading methods in genotype imputation. In addition, based on the FAPI idea, we develop a metric to detect abnormal association at a variant and showed that it had a significantly greater power compared with LD-PAC, a method that quantifies the evidence of spurious associations based on likelihood ratio. Our method is implemented in a user-friendly software tool, which is available at http://statgenpro.psychiatry.hku.hk/fapi.

The DYX2 locus and neurochemical signaling genes contribute to speech sound disorder and related neurocognitive domains

A major milestone of child development is the acquisition and use of speech and language. Communication disorders, including speech sound disorder (SSD), can impair a child's academic, social and behavioral development. Speech sound disorder is a complex, polygenic trait with a substantial genetic component. However, specific genes that contribute to SSD remain largely unknown. To identify associated genes, we assessed the association of the DYX2 dyslexia risk locus and markers in neurochemical signaling genes (e.g., nicotinic and dopaminergic) with SSD and related endophenotypes. We first performed separate primary associations in two independent samples - Cleveland SSD (210 affected and 257 unaffected individuals in 127 families) and Denver SSD (113 affected individuals and 106 unaffected individuals in 85 families) - and then combined results by meta-analysis. DYX2 markers, specifically those in the 3' untranslated region of DCDC2 (P = 1.43 × 10(-4) ), showed the strongest associations with phonological awareness. We also observed suggestive associations of dopaminergic-related genes ANKK1 (P = 1.02 × 10(-2) ) and DRD2 (P = 9.22 × 10(-3) ) and nicotinic-related genes CHRNA3 (P = 2.51 × 10(-3) ) and BDNF (P = 8.14 × 10(-3) ) with case-control status and articulation. Our results further implicate variation in putative regulatory regions in the DYX2 locus, particularly in DCDC2, influencing language and cognitive traits. The results also support previous studies implicating variation in dopaminergic and nicotinic neural signaling influencing human communication and cognitive development. Our findings expand the literature showing genetic factors (e.g., DYX2) contributing to multiple related, yet distinct neurocognitive domains (e.g., dyslexia, language impairment, and SSD). How these factors interactively yield different neurocognitive and language-related outcomes remains to be elucidated.

Mutations in cardiovascular connexin genes

Connexins (Cxs) form a family of transmembrane proteins comprising 21 members in humans. Cxs differ in their expression patterns, biophysical properties and ability to combine into homomeric or heteromeric gap junction channels between neighbouring cells. The permeation of ions and small metabolites through gap junction channels or hemichannels confers a crucial role to these proteins in intercellular communication and in maintaining tissue homeostasis. Among others, Cx37, Cx40, Cx43, Cx45 and Cx47 are found in heart, blood and lymphatic vessels. Mutations or polymorphisms in the genes coding for these Cxs have not only been implicated in cardiovascular pathologies but also in a variety of other disorders. While mutations in Cx43 are mostly linked to oculodentodigital dysplasia, Cx47 mutations are associated with Pelizaeus-Merzbacher-like disease and lymphoedema. Cx40 mutations are principally linked to atrial fibrillation. Mutations in Cx37 have not yet been described, but polymorphisms in the Cx37 gene have been implicated in the development of arterial disease. This review addresses current knowledge on gene mutations in cardiovascular Cxs systematically and links them to alterations in channel properties and disease.

Association of CDH11 with non-syndromic ASD

We report a sporadic patient with Autism Spectrum Disorder (ASD), mild intellectual disability and attention deficit hyperactivity disorder (ADHD) with a de novo partial deletion of CADHERIN 11 (CDH11). The deletion is associated with one of the breakpoints of a de novo complex chromosomal rearrangement 46,XY,t(3;16;5)(q29;q22;q15)inv4(p14;q21)ins(4;5)(q21;q14.3q15). Cadherins are cell adhesion molecules involved in synaptic plasticity. Since genetic evidence points towards a role for cadherins in ASD, we studied the possible contribution of CDH11 to ASD. A case-control association study for 14 SNP variants in 519 ASD cases and 1,192 controls showed significant overrepresentation of rs7187376C/C genotypes in the patient group [P = 0.0049 (Chi-square = 7.90 1 df) and O.R. 3.88 C.I. = 1.403-10.733]. There was no association for C/T versus T/T [P = 0.6772 (Chi-square = 0.17 1 df)] nor was there association at the allelic level [P = 0.4373 (Chi-square = 0.6 1 df)]. In addition to the association of common variants in CDH11 with ASD, we studied the possible contribution of rare variants by sequencing CDH11 in 247 patients, and found three novel variants in the coding region of CDH1, of which two variants were unlikely to be causal. Targeted CNV screening in these 247 patients did not reveal copy number variation in CDH11. In conclusion, the data provide evidence for the involvement of CDH11 in ASD which is consistent with the association of other cadherins with ASD and neuropsychiatric diseases.

Characterization of the DYX2 locus on chromosome 6p22 with reading disability, language impairment, and IQ

Reading disability (RD) and language impairment (LI) are common neurodevelopmental disorders with moderately strong genetic components and lifelong implications. RD and LI are marked by unexpected difficulty acquiring and processing written and verbal language, respectively, despite adequate opportunity and instruction. RD and LI—and their associated deficits—are complex, multifactorial, and often comorbid. Genetic studies have repeatedly implicated the DYX2 locus, specifically the genes DCDC2 and KIAA0319, in RD, with recent studies suggesting they also influence LI, verbal language, and cognition. Here, we characterize the relationship of the DYX2 locus with RD, LI, and IQ. To accomplish this, we developed a marker panel densely covering the 1.4 Mb DYX2 locus and assessed association with reading, language, and IQ measures in subjects from the Avon Longitudinal Study of Parents and Children. We then replicated associations in three independent, disorder-selected cohorts. As expected, there were associations with known RD risk genes KIAA0319 and DCDC2. In addition, we implicated markers in or near other DYX2 genes, including TDP2, ACOT13, C6orf62, FAM65B, and CMAHP. However, the LD structure of the locus suggests that associations within TDP2, ACOT13, and C6orf62 are capturing a previously reported risk variant in KIAA0319. Our results further substantiate the candidacy of KIAA0319 and DCDC2 as major effector genes in DYX2, while proposing FAM65B and CMAHP as new DYX2 candidate genes. Association of DYX2 with multiple neurobehavioral traits suggests risk variants have functional consequences affecting multiple neurological processes. Future studies should dissect these functional, possibly interactive relationships of DYX2 candidate genes.

Alleles of a polymorphic ETV6 binding site in DCDC2 confer risk of reading and language impairment

Reading disability (RD) and language impairment (LI) are common learning disabilities that make acquisition and utilization of reading and verbal language skills, respectively, difficult for affected individuals. Both disorders have a substantial genetic component with complex inheritance. Despite decades of study, reading and language, like many other complex traits, consistently evade identification of causative and functional variants. We previously identified a putative functional risk variant, named BV677278 for its GenBank accession number, for RD in DCDC2. This variant consists of an intronic microdeletion and a highly polymorphic short tandem repeat (STR) within its breakpoints. We have also shown this STR to bind to an unknown nuclear protein with high specificity. Here, we replicate BV677278's association with RD, expand its association to LI, identify the BV677278-binding protein as the transcription factor ETV6, and provide compelling genetic evidence that BV677278 is a regulatory element that influences reading and language skills. We also provide evidence that BV677278 interacts nonadditively with KIAA0319, an RD-associated gene, to adversely affect several reading and cognitive phenotypes. On the basis of these data, we propose a new name for BV677278: "READ1" or "regulatory element associated with dyslexia 1."

A Collective Ranking Method for Genome-wide Association Studies

Genome-wide association studies (GWAS) analyze genetic variation (SNPs) across the entire human genome, searching for SNPs that are associated with certain phenotypes, most often diseases, such as breast cancer. In GWAS, we seek a ranking of SNPs in terms of their relevance to the given phenotype. However, because certain SNPs are known to be highly correlated with one another across individuals, it can be beneficial to take into account these correlations when ranking. If a SNP appears associated with the phenotype, and we question whether this association is real, the extent to which its neighbors (correlated SNPs) also appear associated can be informative. Therefore, we propose CollectRank, a ranking approach which allows SNPs to reinforce one another via the correlation structure. CollectRank is loosely analogous to the well-known PageRank algorithm. We first evaluate CollectRank on synthetic data generated from a variety of genetic models under different settings. The numerical results suggest CollectRank can significantly outperform common GWAS methods at the cost of a small amount of extra computation. We further evaluate CollectRank on two real-world GWAS on breast cancer and atrial fibrillation/flutter, and CollectRank performs well in both studies. We finally provide a theoretical analysis that also suggests CollectRank's advantages.

Postassociation cleaning using linkage disequilibrium information

In genetic association studies, quality control (QC) filters are applied to remove potentially problematic markers before the markers are tested for statistical associations. However, spurious associations can still occur after QC. We introduce Post-Association Cleaning (PAC) approach that can complement QC by capturing spurious associations using the information in the post-association results. Specifically, we propose a PAC filter based on the linkage disequilibrium (LD) information. The intuition is that if the association is caused by a true genetic effect, neighboring markers in LD should show comparably significant P-values. If not, it may be evidence of spurious association. Previous studies have applied the same idea but only manually without a formal statistical framework. Our proposed method LD-PAC provides a systematic framework to quantitatively measure the evidence of spurious associations based on the likelihood ratio. Simulations show that LD-PAC can detect spurious associations with high detection rate (84%). In addition to detecting spurious associations, our method can also be used to "rescue" candidate associations from the supposedly unclean data such as the markers excluded by QC. Although the additional associations must be treated with care, they can suggest interesting regions. The application of our method to the Wellcome Trust Case Control Consortium (WTCCC) data led to the discovery of an additional candidate association for type 1 diabetes among the QC-excluded markers. This locus turns out to be in a region recently identified as significant by a meta-analysis performed after the WTCCC study was published.

An optimal weighted aggregated association test for identification of rare variants involved in common diseases

The advent of next generation sequencing technologies allows one to discover nearly all rare variants in a genomic region of interest. This technological development increases the need for an effective statistical method for testing the aggregated effect of rare variants in a gene on disease susceptibility. The idea behind this approach is that if a certain gene is involved in a disease, many rare variants within the gene will disrupt the function of the gene and are associated with the disease. In this article, we present the rare variant weighted aggregate statistic (RWAS), a method that groups rare variants and computes a weighted sum of differences between case and control mutation counts. We show that our method outperforms the groupwise association test of Madsen and Browning in the disease-risk model that assumes that each variant makes an equally small contribution to disease risk. In addition, we can incorporate prior information into our method of which variants are likely causal. By using simulated data and real mutation screening data of the susceptibility gene for ataxia telangiectasia, we demonstrate that prior information has a substantial influence on the statistical power of association studies. Our method is publicly available at http://genetics.cs.ucla.edu/rarevariants.

Statistical Optimization of Pharmacogenomics Association Studies: Key Considerations from Study Design to Analysis

Research in human genetics and genetic epidemiology has grown significantly over the previous decade, particularly in the field of pharmacogenomics. Pharmacogenomics presents an opportunity for rapid translation of associated genetic polymorphisms into diagnostic measures or tests to guide therapy as part of a move towards personalized medicine. Expansion in genotyping technology has cleared the way for widespread use of whole-genome genotyping in the effort to identify novel biology and new genetic markers associated with pharmacokinetic and pharmacodynamic endpoints. With new technology and methodology regularly becoming available for use in genetic studies, a discussion on the application of such tools becomes necessary. In particular, quality control criteria have evolved with the use of GWAS as we have come to understand potential systematic errors which can be introduced into the data during genotyping. There have been several replicated pharmacogenomic associations, some of which have moved to the clinic to enact change in treatment decisions. These examples of translation illustrate the strength of evidence necessary to successfully and effectively translate a genetic discovery. In this review, the design of pharmacogenomic association studies is examined with the goal of optimizing the impact and utility of this research. Issues of ascertainment, genotyping, quality control, analysis and interpretation are considered.

A hierarchical Bayesian network approach for linkage disequilibrium modeling and data-dimensionality reduction prior to genome-wide association studies

Background

Discovering the genetic basis of common genetic diseases in the human genome represents a public health issue. However, the dimensionality of the genetic data (up to 1 million genetic markers) and its complexity make the statistical analysis a challenging task.

Results

We present an accurate modeling of dependences between genetic markers, based on a forest of hierarchical latent class models which is a particular class of probabilistic graphical models. This model offers an adapted framework to deal with the fuzzy nature of linkage disequilibrium blocks. In addition, the data dimensionality can be reduced through the latent variables of the model which synthesize the information borne by genetic markers. In order to tackle the learning of both forest structure and probability distributions, a generic algorithm has been proposed. A first implementation of our algorithm has been shown to be tractable on benchmarks describing 10⁵ variables for 2000 individuals.

Conclusions

The forest of hierarchical latent class models offers several advantages for genome-wide association studies: accurate modeling of linkage disequilibrium, flexible data dimensionality reduction and biological meaning borne by latent variables.

Assessment of SNPs associated with the human glucocorticoid receptor in primary open-angle glaucoma and steroid responders

PURPOSE

While chronic glucocorticoid (GC) therapy leads to ocular hypertension in about one third of individuals, almost all primary open-angle glaucoma (POAG) patients show this response and are called "steroid responders." Two differentially spliced isoforms of the glucocorticoid receptor (GR), GRalpha and GRbeta, regulate GC responsiveness in trabecular meshwork (TM) cells. GRbeta acts as a dominant negative regulator of GC activity and is expressed at lower levels in glaucomatous TM cells, making them more sensitive to GCs. Several arginine/serine-rich splicing factor (SR) proteins have been implicated in alternative splicing of the GR. We have previously demonstrated that immunophilins FKBP5 and FKBP4 are required for GRalpha and GRbeta translocation into the nucleus, which is essential for their biologic activity. The purpose of the present study was to use single nucleotide polymorphism (SNP) genotyping to determine whether there are any allele frequency differences in GR, FKBP4/5, or SR genes between normal control, POAG, and steroid responder populations.

METHODS

Clinically characterized individuals (400 normal controls, 197 POAG, and 107 steroid responders) were recruited from the U. Iowa Ophthalmology Clinics after IRB approved consent. Genotyping of DNA samples for 48 SNPs in SFRS3, SFRS5, SFRS9, FKBP4, FKBP5, and NR3C1 was done at GeneSeek using a mass spectroscopy based system.

RESULTS

All 48 SNPs displayed high call rates (99%). There were no significant differences in allele frequencies or genotypes in SNPs for SFRS5, SFRS9, FKBP4, FKBP5, and NR3C1 between the 3 groups. Up to three SNPs in SFRS3 had p-values <0.05 when comparing controls to POAG or steroid responders, but this statistical significance was lost when the p values were adjusted for multiple measures.

CONCLUSIONS

Although these 6 genes may be involved in the pathogenesis of GC-induced ocular hypertension, it does not appear that major heritable risk alleles in these genes are responsible for the development of GC-induced ocular hypertension or POAG.

Influence of hepatocyte nuclear factor 4alpha (HNF4alpha) gene variants on the risk of type 2 diabetes: a meta-analysis in 49,577 individuals

BACKGROUND

The nuclear receptor hepatocyte nuclear factor 4alpha (HNF4alpha) contributes to the regulation of a large fraction of liver and pancreatic islet transcriptomes.

AIM

To evaluate the influence of HNF4alpha polymorphisms across the entire locus on the occurrence of type 2 diabetes (T2D) by means of a meta-analysis.

METHODS

We evaluated haplotype block structure of HNF4alpha variants owing to linkage disequilibrium (LD). From 1455 reports, we evaluated 21 observational studies.

RESULTS

Six haplotype blocks of LD were constructed with SNPs with r(2)>0.8; there were also 14 unlinked SNPs. Overall, we included 22,920 cases and 26.657 controls. Among 17 heterogeneous studies (21,881 cases and 24,915 controls), including 3 SNPs of P2 promoter region in block 1, we observed a significant association with T2D in fixed (OR 0.94, 95%CI: 0.905-0.975, p=0.001) and random (OR 0.988, 95%CI: 0.880-0.948, p=0.000012) model. Three homogeneous studies were evaluated in block 2 (2684 cases and 2059 controls), and a significant association with T2D was also observed: OR: 1.121, 95%CI 1.013-1.241, p=0.027. Three additional variants were associated with T2D: two intronic SNPs (rs4810424: OR: 1.080, 95%CI: 1.010-1.154, p<0.03 and rs3212183: OR: 0.843, 95%CI: 0.774-0.918, p<0.00009) and one missense variant (rs1800961: OR: 0.770, 95%CI: 0.595-0.995, p<0.05, 6562 cases and 6723 controls).

CONCLUSIONS

In addition to HNF4alpha variants in the promoter region, other SNPs may be involved on the occurrence of T2D.

Rapid and accurate multiple testing correction and power estimation for millions of correlated markers

With the development of high-throughput sequencing and genotyping technologies, the number of markers collected in genetic association studies is growing rapidly, increasing the importance of methods for correcting for multiple hypothesis testing. The permutation test is widely considered the gold standard for accurate multiple testing correction, but it is often computationally impractical for these large datasets. Recently, several studies proposed efficient alternative approaches to the permutation test based on the multivariate normal distribution (MVN). However, they cannot accurately correct for multiple testing in genome-wide association studies for two reasons. First, these methods require partitioning of the genome into many disjoint blocks and ignore all correlations between markers from different blocks. Second, the true null distribution of the test statistic often fails to follow the asymptotic distribution at the tails of the distribution. We propose an accurate and efficient method for multiple testing correction in genome-wide association studies—SLIDE. Our method accounts for all correlation within a sliding window and corrects for the departure of the true null distribution of the statistic from the asymptotic distribution. In simulations using the Wellcome Trust Case Control Consortium data, the error rate of SLIDE's corrected p-values is more than 20 times smaller than the error rate of the previous MVN-based methods' corrected p-values, while SLIDE is orders of magnitude faster than the permutation test and other competing methods. We also extend the MVN framework to the problem of estimating the statistical power of an association study with correlated markers and propose an efficient and accurate power estimation method SLIP. SLIP and SLIDE are available at http://slide.cs.ucla.edu.

In genome-wide association studies, it is important to account for the fact that a large number of genetic variants are tested in order to adequately control for false positives. The simplest way to correct for multiple hypothesis testing is the Bonferroni correction, which multiplies the p-values by the number of markers assuming the markers are independent. Since the markers are correlated due to linkage disequilibrium, this approach leads to a conservative estimate of false positives, thus adversely affecting statistical power. The permutation test is considered the gold standard for accurate multiple testing correction, but is often computationally impractical for large association studies. We propose a method that efficiently and accurately corrects for multiple hypotheses in genome-wide association studies by fully accounting for the local correlation structure between markers. Our method also corrects for the departure of the true distribution of test statistics from the asymptotic distribution, which dramatically improves the accuracy, particularly when many rare variants are included in the tests. Our method shows a near identical accuracy to permutation and shows greater computational efficiency than previously suggested methods. We also provide a method to accurately and efficiently estimate the statistical power of genome-wide association studies.