ComPaSS-GWAS: A method to reduce type I error in genome-wide association studies when replication data are not available.

Results from association studies are traditionally corroborated by replicating the findings in an independent data set. Although replication studies may be comparable for the main trait or phenotype of interest, it is unlikely that secondary phenotypes will be comparable across studies, making replication problematic. Alternatively, there may simply not be a replication sample available because of the nature or frequency of the phenotype. In these situations, an approach based on complementary pairs stability selection for genome-wide association study (ComPaSS-GWAS), is proposed as an ad-hoc alternative to replication. In this method, the sample is randomly split into two conditionally independent halves multiple times (resamples) and a GWAS is performed on each half in each resample. Similar in spirit to testing for association with independent discovery and replication samples, a marker is corroborated if its p-value is significant in both halves of the resample. Simulation experiments were performed for both nongenetic and genetic models. The type I error rate and power of ComPaSS-GWAS were determined and compared to the statistical properties of a traditional GWAS. Simulation results show that the type I error rate decreased as the number of resamples increased with only a small reduction in power and that these results were comparable with those from a traditional GWAS. Blood levels of vitamin pyridoxal 5'-phosphate from the Trinity Student Study (TSS) were used to validate this approach. The results from the validation study were compared to, and were consistent with, those obtained from previously published independent replication data and functional studies.

A genome-wide association study implicates the BMP7 locus as a risk factor for nonsyndromic metopic craniosynostosis.

Our previous genome-wide association study (GWAS) for sagittal nonsyndromic craniosynostosis (sNCS) provided important insights into the genetics of midline CS. In this study, we performed a GWAS for a second midline NCS, metopic NCS (mNCS), using 215 non-Hispanic white case-parent triads. We identified six variants with genome-wide significance (P ≤ 5 × 10-8): rs781716 (P = 4.71 × 10-9; odds ratio [OR] = 2.44) intronic to SPRY3; rs6127972 (P = 4.41 × 10-8; OR = 2.17) intronic to BMP7; rs62590971 (P = 6.22 × 10-9; OR = 0.34), located ~ 155 kb upstream from TGIF2LX; and rs2522623, rs2573826, and rs2754857, all intronic to PCDH11X (P = 1.76 × 10-8, OR = 0.45; P = 3.31 × 10-8, OR = 0.45; P = 1.09 × 10-8, OR = 0.44, respectively). We performed a replication study of these variants using an independent non-Hispanic white sample of 194 unrelated mNCS cases and 333 unaffected controls; only the association for rs6127972 (P = 0.004, OR = 1.45; meta-analysis P = 1.27 × 10-8, OR = 1.74) was replicated. Our meta-analysis examining single nucleotide polymorphisms common to both our mNCS and sNCS studies showed the strongest association for rs6127972 (P = 1.16 × 10-6). Our imputation analysis identified a linkage disequilibrium block encompassing rs6127972, which contained an enhancer overlapping a CTCF transcription factor binding site (chr20:55,798,821-55,798,917) that was significantly hypomethylated in mesenchymal stem cells derived from fused metopic compared to open sutures from the same probands. This study provides additional insights into genetic factors in midline CS.

Genetic associations with childhood brain growth, defined in two longitudinal cohorts.

Genome-wide association studies (GWASs) are unraveling the genetics of adult brain neuroanatomy as measured by cross-sectional anatomic magnetic resonance imaging (aMRI). However, the genetic mechanisms that shape childhood brain development are, as yet, largely unexplored. In this study we identify common genetic variants associated with childhood brain development as defined by longitudinal aMRI. Genome-wide single nucleotide polymorphism (SNP) data were determined in two cohorts: one enriched for attention-deficit/hyperactivity disorder (ADHD) (LONG cohort: 458 participants; 119 with ADHD) and the other from a population-based cohort (Generation R: 257 participants). The growth of the brain's major regions (cerebral cortex, white matter, basal ganglia, and cerebellum) and one region of interest (the right lateral prefrontal cortex) were defined on all individuals from two aMRIs, and a GWAS and a pathway analysis were performed. In addition, association between polygenic risk for ADHD and brain growth was determined for the LONG cohort. For white matter growth, GWAS meta-analysis identified a genome-wide significant intergenic SNP (rs12386571, P = 9.09 × 10-9 ), near AKR1B10. This gene is part of the aldo-keto reductase superfamily and shows neural expression. No enrichment of neural pathways was detected and polygenic risk for ADHD was not associated with the brain growth phenotypes in the LONG cohort that was enriched for the diagnosis of ADHD. The study illustrates the use of a novel brain growth phenotype defined in vivo for further study.

Fine-mapping additive and dominant SNP effects using group-LASSO and fractional resample model averaging.

Genomewide association studies (GWAS) sometimes identify loci at which both the number and identities of the underlying causal variants are ambiguous. In such cases, statistical methods that model effects of multiple single-nucleotide polymorphisms (SNPs) simultaneously can help disentangle the observed patterns of association and provide information about how those SNPs could be prioritized for follow-up studies. Current multi-SNP methods, however, tend to assume that SNP effects are well captured by additive genetics; yet when genetic dominance is present, this assumption translates to reduced power and faulty prioritizations. We describe a statistical procedure for prioritizing SNPs at GWAS loci that efficiently models both additive and dominance effects. Our method, LLARRMA-dawg, combines a group LASSO procedure for sparse modeling of multiple SNP effects with a resampling procedure based on fractional observation weights. It estimates for each SNP the robustness of association with the phenotype both to sampling variation and to competing explanations from other SNPs. In producing an SNP prioritization that best identifies underlying true signals, we show the following: our method easily outperforms a single-marker analysis; when additive-only signals are present, our joint model for additive and dominance is equivalent to or only slightly less powerful than modeling additive-only effects; and when dominance signals are present, even in combination with substantial additive effects, our joint model is unequivocally more powerful than a model assuming additivity. We also describe how performance can be improved through calibrated randomized penalization, and discuss how dominance in ungenotyped SNPs can be incorporated through either heterozygote dosage or multiple imputation.

Plasma Levels of Soluble Interleukin-2 Receptor α: Associations With Clinical Cardiovascular Events and Genome-Wide Association Scan.

OBJECTIVE: Interleukin (IL) -2 receptor subunit α regulates lymphocyte activation, which plays an important role in atherosclerosis. Associations between soluble IL-2Rα (sIL-2Rα) and cardiovascular disease (CVD) have not been widely studied and little is known about the genetic determinants of sIL-2Rα levels. APPROACH AND RESULTS: We measured baseline levels of sIL-2Rα in 4408 European American (EA) and 766 African American (AA) adults from the Cardiovascular Health Study (CHS) and examined associations with baseline CVD risk factors, subclinical CVD, and incident CVD events. We also performed a genome-wide association study for sIL-2Rα in CHS (2964 EAs and 683 AAs) and further combined CHS EA results with those from two other EA cohorts in a meta-analysis (n=4464 EAs). In age, sex- and race- adjusted models, sIL-2Rα was positively associated with current smoking, type 2 diabetes mellitus, hypertension, insulin, waist circumference, C-reactive protein, IL-6, fibrinogen, internal carotid wall thickness, all-cause mortality, CVD mortality, and incident CVD, stroke, and heart failure. When adjusted for baseline CVD risk factors and subclinical CVD, associations with all-cause mortality, CVD mortality, and heart failure remained significant in both EAs and AAs. In the EA genome-wide association study analysis, we observed 52 single-nucleotide polymorphisms in the chromosome 10p15-14 region, which contains IL2RA, IL15RA, and RMB17, that reached genome-wide significance (P<5×10(-8)). The most significant single-nucleotide polymorphism was rs7911500 (P=1.31×10(-75)). The EA meta-analysis results were highly consistent with CHS-only results. No single-nucleotide polymorphisms reached statistical significance in the AAs. CONCLUSIONS: These results support a role for sIL-2Rα in atherosclerosis and provide evidence for multiple-associated single-nucleotide polymorphisms at chromosome 10p15-14.

Genome- and exome-wide association study of serum lipoprotein (a) in the Jackson Heart Study.

Lipoprotein (a) (Lp(a)) is an independent risk factor for cardiovascular disease. Lp(a) levels in African Americans (AAs) are much higher compared with that in European Americans. We conducted a genome- and an exome-wide association study of Lp(a) among 2895 AAs participating in the Jackson Heart Study. We observed that local ancestry at 6q25.3 was an important risk factor for Lp(a) in AAs, and that multiple single-nucleotide polymorphisms (SNPs) at the well-established LPA locus were significantly associated with Lp(a) (P<5 × 10(-8)) after adjusting for the local ancestry at 6q25.3. Interestingly, before adjusting for local ancestry, we observed significant (P<5 × 10(-8)) associations for hundreds of SNPs spanning ~10 Mb region on 6q surrounding the LPA gene, whereas after adjusting for local ancestry, the region containing significantly associated SNPs got much narrower and was centered over the LPA gene (<1 Mb). We observed a single nonsynonymous SNP in APOE significantly associated with Lp(a) (P<5 × 10(-8)). A high burden of coding variants in LPA and APOE were also associated with higher Lp(a) levels. Our study provides evidence that ancestry-specific causal risk variant(s) resides in or near LPA and that most of the observed associations outside this narrower region are spurious associations.

A permutation approach for selecting the penalty parameter in penalized model selection.

We describe a simple, computationally efficient, permutation-based procedure for selecting the penalty parameter in LASSO-penalized regression. The procedure, permutation selection, is intended for applications where variable selection is the primary focus, and can be applied in a variety of structural settings, including that of generalized linear models. We briefly discuss connections between permutation selection and existing theory for the LASSO. In addition, we present a simulation study and an analysis of real biomedical data sets in which permutation selection is compared with selection based on the following: cross-validation (CV), the Bayesian information criterion (BIC), scaled sparse linear regression, and a selection method based on recently developed testing procedures for the LASSO.

Reprioritizing genetic associations in hit regions using LASSO-based resample model averaging.

Significance testing one SNP at a time has proven useful for identifying genomic regions that harbor variants affecting human disease. But after an initial genome scan has identified a "hit region" of association, single-locus approaches can falter. Local linkage disequilibrium (LD) can make both the number of underlying true signals and their identities ambiguous. Simultaneous modeling of multiple loci should help. However, it is typically applied ad hoc: conditioning on the top SNPs, with limited exploration of the model space and no assessment of how sensitive model choice was to sampling variability. Formal alternatives exist but are seldom used. Bayesian variable selection is coherent but requires specifying a full joint model, including priors on parameters and the model space. Penalized regression methods (e.g., LASSO) appear promising but require calibration, and, once calibrated, lead to a choice of SNPs that can be misleadingly decisive. We present a general method for characterizing uncertainty in model choice that is tailored to reprioritizing SNPs within a hit region under strong LD. Our method, LASSO local automatic regularization resample model averaging (LLARRMA), combines LASSO shrinkage with resample model averaging and multiple imputation, estimating for each SNP the probability that it would be included in a multi-SNP model in alternative realizations of the data. We apply LLARRMA to simulations based on case-control genome-wide association studies data, and find that when there are several causal loci and strong LD, LLARRMA identifies a set of candidates that is enriched for true signals relative to single locus analysis and to the recently proposed method of Stability Selection.

Genome-wide association study of café-au-lait macule number in neurofibromatosis type 1.

BACKGROUND: Neurofibromatosis type 1 (NF1) is a tumor-predisposition disorder that arises due to pathogenic variants in tumor suppressor NF1. NF1 has variable expressivity that may be due, at least in part, from heritable elements such as modifier genes; however, few genetic modifiers have been identified to date. METHODS: In this study, we performed a genome-wide association analysis of the number of café-au-lait macules (CALM) that are considered a tumor-like trait as a clinical phenotype modifying NF1. RESULTS: A borderline genome-wide significant association was identified in the discovery cohort (CALM1, N = 112) between CALM number and rs12190451 (and rs3799603, r2  = 1.0; p = 7.4 × 10-8 ) in the intronic region of RPS6KA2. Although, this association was not replicated in the second cohort (CALM2, N = 59) and a meta-analysis did not show significantly associated variants in this region, a significant corroboration score (0.72) was obtained for the RPS6KA2 signal in the discovery cohort (CALM1) using Complementary Pairs Stability Selection for Genome-Wide Association Studies (ComPaSS-GWAS) analysis, suggesting that the lack of replication may be due to heterogeneity of the cohorts rather than type I error. CONCLUSION: rs12190451 is located in a melanocyte-specific enhancer and may influence RPS6KA2 expression in melanocytes-warranting further functional studies.

Type I error rates of rare single nucleotide variants are inflated in tests of association with non-normally distributed traits using simple linear regression methods.

In this study, the effects of (a) the minor allele frequency of the single nucleotide variant (SNV), (b) the degree of departure from normality of the trait, and (c) the position of the SNVs on type I error rates were investigated in the Genetic Analysis Workshop (GAW) 19 whole exome sequence data. To test the distribution of the type I error rate, 5 simulated traits were considered: standard normal and gamma distributed traits; 2 transformed versions of the gamma trait (log10 and rank-based inverse normal transformations); and trait Q1 provided by GAW 19. Each trait was tested with 313,340 SNVs. Tests of association were performed with simple linear regression and average type I error rates were determined for minor allele frequency classes. Rare SNVs (minor allele frequency < 0.05) showed inflated type I error rates for non-normally distributed traits that increased as the minor allele frequency decreased. The inflation of average type I error rates increased as the significance threshold decreased. Normally distributed traits did not show inflated type I error rates with respect to the minor allele frequency for rare SNVs. There was no consistent effect of transformation on the uniformity of the distribution of the location of SNVs with a type I error.