Andrew D. Skol

Joint analysis is more efficient than replication-based analysis for two-stage genome-wide association studies

Genome-wide association is a promising approach to identify common genetic variants that predispose to human disease. Because of the high cost of genotyping hundreds of thousands of markers on thousands of subjects, genome-wide association studies often follow a staged design in which a proportion (πsamples) of the available samples are genotyped on a large number of markers in stage 1, and a proportion (πsamples) of these markers are later followed up by genotyping them on the remaining samples in stage 2. The standard strategy for analyzing such two-stage data is to view stage 2 as a replication study and focus on findings that reach statistical significance when stage 2 data are considered alone. We demonstrate that the alternative strategy of jointly analyzing the data from both stages almost always results in increased power to detect genetic association, despite the need to use more stringent significance levels, even when effect sizes differ between the two stages. We recommend joint analysis for all two-stage genome-wide association studies, especially when a relatively large proportion of the samples are genotyped in stage 1 (πsamples ≥ 0.30), and a relatively large proportion of markers are selected for follow-up in stage 2 (πmarkers ≥ 0.01).

Probability of Detection of Genotyping Errors and Mutations as Inheritance Inconsistencies in Nuclear-Family Data

Gene-mapping studies routinely rely on checking for Mendelian transmission of marker alleles in a pedigree, as a means of screening for genotyping errors and mutations, with the implicit assumption that, if a pedigree is consistent with Mendels laws of inheritance, then there are no genotyping errors. However, the occurrence of inheritance inconsistencies alone is an inadequate measure of the number of genotyping errors, since the rate of occurrence depends on the number and relationships of genotyped pedigree members, the type of errors, and the distribution of marker-allele frequencies. In this article, we calculate the expected probability of detection of a genotyping error or mutation as an inheritance inconsistency in nuclear-family data, as a function of both the number of genotyped parents and offspring and the marker-allele frequency distribution. Through computer simulation, we explore the sensitivity of our analytic calculations to the underlying error model. Under a random-allele-error model, we find that detection rates are 51%-77% for multiallelic markers and 13%-75% for biallelic markers; detection rates are generally lower when the error occurs in a parent than in an offspring, unless a large number of offspring are genotyped. Errors are especially difficult to detect for biallelic markers with equally frequent alleles, even when both parents are genotyped; in this case, the maximum detection rate is 34% for four-person nuclear families. Error detection in families in which parents are not genotyped is limited, even with multiallelic markers. Given these results, we recommend that additional error checking (e.g., on the basis of multipoint analysis) be performed, beyond routine checking for Mendelian consistency. Furthermore, our results permit assessment of the plausibility of an observed number of inheritance inconsistencies for a family, allowing the detection of likely pedigree-rather than genotyping-errors in the early stages of a genome scan. Such early assessments are valuable in either the targeting of families for resampling or discontinued genotyping.

Interaction of prenatal exposure to cigarettes and MAOA genotype in pathways to youth antisocial behavior

Genetic susceptibility to antisocial behavior may increase fetal sensitivity to prenatal exposure to cigarette smoke. Testing putative gene × exposure mechanisms requires precise measurement of exposure and outcomes. We tested whether a functional polymorphism in the gene encoding the enzyme monoamine oxidase A (MAOA) interacts with exposure to predict pathways to adolescent antisocial behavior. We assessed both clinical and information-processing outcomes. One hundred seventy-six adolescents and their mothers participated in a follow-up of a pregnancy cohort with well-characterized exposure. A sex-specific pattern of gene × exposure interaction was detected. Exposed boys with the low-activity MAOA 5′ uVNTR (untranslated region variable number of tandem repeats) genotype were at increased risk for conduct disorder (CD) symptoms. In contrast, exposed girls with the high-activity MAOA uVNTR genotype were at increased risk for both CD symptoms and hostile attribution bias on a face-processing task. There was no evidence of a gene–environment correlation (rGE). Findings suggest that the MAOA uVNTR genotype, prenatal exposure to cigarettes and sex interact to predict antisocial behavior and related information-processing patterns. Future research to replicate and extend these findings should focus on elucidating how gene × exposure interactions may shape behavior through associated changes in brain function.

Variants at 6q21 implicate PRDM1 in the etiology of therapy-induced second malignancies after Hodgkin's lymphoma.

Survivors of pediatric Hodgkins lymphoma are at risk for radiation therapy–induced second malignant neoplasms (SMNs). We identified two variants at chromosome 6q21 associated with SMNs in survivors of Hodgkins lymphoma treated with radiation therapy as children but not as adults. The variants comprise a risk locus associated with decreased basal expression of PRDM1 (encoding PR domain containing 1, with ZNF domain) and impaired induction of the PRDM1 protein after radiation exposure. These data suggest a new gene-exposure interaction that may implicate PRDM1 in the etiology of radiation therapy-induced SMNs.

Screening of 134 Single Nucleotide Polymorphisms (SNPs) Previously Associated With Type 2 Diabetes Replicates Association With 12 SNPs in Nine Genes

More than 120 published reports have described associations between single nucleotide polymorphisms (SNPs) and type 2 diabetes. However, multiple studies of the same variant have often been discordant. From a literature search, we identified previously reported type 2 diabetes–associated SNPs. We initially genotyped 134 SNPs on 786 index case subjects from type 2 diabetes families and 617 control subjects with normal glucose tolerance from Finland and excluded from analysis 20 SNPs in strong linkage disequilibrium (r2 > 0.8) with another typed SNP. Of the 114 SNPs examined, we followed up the 20 most significant SNPs (P < 0.10) on an additional 384 case subjects and 366 control subjects from a population-based study in Finland. In the combined data, we replicated association (P < 0.05) for 12 SNPs: PPARG Pro12Ala and His447, KCNJ11 Glu23Lys and rs5210, TNF −857, SLC2A2 Ile110Thr, HNF1A/TCF1 rs2701175 and GE117881_360, PCK1 −232, NEUROD1 Thr45Ala, IL6 −598, and ENPP1 Lys121Gln. The replication of 12 SNPs of 114 tested was significantly greater than expected by chance under the null hypothesis of no association (P = 0.012). We observed that SNPs from genes that had three or more previous reports of association were significantly more likely to be replicated in our sample (P = 0.03), although we also replicated 4 of 58 SNPs from genes that had only one previous report of association.

Evaluation of genetic variability in the dopamine receptor D2 in relation to behavioral inhibition and impulsivity/sensation seeking: an exploratory study with d-amphetamine in healthy participants.

The dopamine D2 receptor (DRD2) appears to be involved in impulsive behaviors, and particularly in behavioral inhibition. We sought to determine whether inhibition and impulsivity were related to genetic polymorphisms in the DRD2 gene (DRD2) in healthy volunteers (N = 93). Participants received placebo or d-amphetamine in random order. They performed the stop task, measuring behavioral inhibition, and rated their mood states on each session. They also completed the Zuckerman-Kuhlman Personality Questionnaire, including an Impulsivity subscale. We investigated the association between 12 single nucleotide polymorphisms (SNPs) and haplotypes in DRD2 and stop task performance in the nondrug (i.e., placebo) session and on the personality measure of impulsivity. We secondarily evaluated the DRD2 SNPs in relation to response to d-amphetamine on stop task performance and mood ratings. Mood was not related to genotypes in either the drug free condition or in response to drug. However, 2 SNPs, rs4648317 and rs12364283, and a haplotype block consisting of those SNPs, were associated with better performance on the stop task in the drug free condition and lower scores on the Impulsivity subscale. We also found that rs12364283 was associated with effects of d-amphetamine on stop task performance: d-amphetamine decreased stop reaction time (RT) in the A/A group but increased stop RT in the combined A/G + G/G genotype. Of the SNPs we evaluated, rs12364283, which has been associated with DRD2 expression, was the most significantly associated with inhibition and impulsivity. The significant relationship between DRD2 genotype and both behavioral inhibition and impulsivity suggests a possible common genetic influence on behavioral and self-report measures of impulsivity.

Genome-Wide Association Studies and the Problem of Relatedness Among Advanced Intercross Lines and Other Highly Recombinant Populations

Model organisms offer many advantages for the genetic analysis of complex traits. However, identification of specific genes is often hampered by a lack of recombination between the genomes of inbred progenitors. Recently, genome-wide association studies (GWAS) in humans have offered gene-level mapping resolution that is possible because of the large number of accumulated recombinations among unrelated human subjects. To obtain analogous improvements in mapping resolution in mice, we used a 34th generation advanced intercross line (AIL) derived from two inbred strains (SM/J and LG/J). We used simulations to show that familial relationships among subjects must be accounted for when analyzing these data; we then used a mixed model that included polygenic effects to address this problem in our own analysis. Using a combination of F2 and AIL mice derived from the same inbred progenitors, we identified genome-wide significant, subcentimorgan loci that were associated with methamphetamine sensitivity, (e.g., chromosome 18; LOD = 10.5) and non-drug-induced locomotor activity (e.g., chromosome 8; LOD = 18.9). The 2-LOD support interval for the former locus contains no known genes while the latter contains only one gene (Csmd1). This approach is broadly applicable in terms of phenotypes and model organisms and allows GWAS to be performed in multigenerational crosses between and among inbred strains where familial relatedness is often unavoidable.

Trait-stratified genome-wide association study identifies novel and diverse genetic associations with serologic and cytokine phenotypes in systemic lupus erythematosus

IntroductionSystemic lupus erythematosus (SLE) is a highly heterogeneous disorder, characterized by differences in autoantibody profile, serum cytokines, and clinical manifestations. SLE-associated autoantibodies and high serum interferon alpha (IFN-α) are important heritable phenotypes in SLE which are correlated with each other, and play a role in disease pathogenesis. These two heritable risk factors are shared between ancestral backgrounds. The aim of the study was to detect genetic factors associated with autoantibody profiles and serum IFN-α in SLE.MethodsWe undertook a case-case genome-wide association study of SLE patients stratified by ancestry and extremes of phenotype in serology and serum IFN-α. Single nucleotide polymorphisms (SNPs) in seven loci were selected for follow-up in a large independent cohort of 538 SLE patients and 522 controls using a multi-step screening approach based on novel metrics and expert database review. The seven loci were: leucine-rich repeat containing 20 (LRRC20); protein phosphatase 1 H (PPM1H); lysophosphatidic acid receptor 1 (LPAR1); ankyrin repeat and sterile alpha motif domain 1A (ANKS1A); protein tyrosine phosphatase, receptor type M (PTPRM); ephrin A5 (EFNA5); and V-set and immunoglobulin domain containing 2 (VSIG2).ResultsSNPs in the LRRC20, PPM1H, LPAR1, ANKS1A, and VSIG2 loci each demonstrated strong association with a particular serologic profile (all odds ratios > 2.2 and P < 3.5 × 10-4). Each of these serologic profiles was associated with increased serum IFN-α. SNPs in both PTPRM and LRRC20 were associated with increased serum IFN-α independent of serologic profile (P = 2.2 × 10-6 and P = 2.6 × 10-3 respectively). None of the SNPs were strongly associated with SLE in case-control analysis, suggesting that the major impact of these variants will be upon subphenotypes in SLE.ConclusionsThis study demonstrates the power of using serologic and cytokine subphenotypes to elucidate genetic factors involved in complex autoimmune disease. The distinct associations observed emphasize the heterogeneity of molecular pathogenesis in SLE, and the need for stratification by subphenotypes in genetic studies. We hypothesize that these genetic variants play a role in disease manifestations and severity in SLE.

Genetic heterogeneity in colorectal cancer associations between African and European Americans

BACKGROUND & AIMS Genome-wide association studies of colorectal cancer (CRC) have identified risk variants in 10 genomic regions. None of these studies included African Americans, who have the highest incidence and mortality from CRC in the United States. For the 10 genomic regions, we performed an association study of Americans of African and European descent. METHODS We genotyped 22 single nucleotide polymorphisms (SNPs) in DNA samples from 1194 patients with CRC (795 African Americans and 399 European Americans) and 1352 controls (985 African Americans and 367 European Americans). At chromosome 8q24.21 region 3, we analyzed 6 SNPs from 1000 African American cases and 1393 controls. Association testing was done using multivariate logistic regression controlling for ancestry, age, and sex. RESULTS Among African Americans, the SNP rs6983267 at 8q24.21 was not associated with CRC (odds ratio, 1.18; P = .12); instead, the 8q24.21 SNP rs7014346 (odds ratio, 1.15; P = .03) was associated with CRC in this population. At 15q13.3, rs10318 was associated with CRC in both populations. At 11q23.1, rs3802842 was significantly associated with rectal cancer risk only among African Americans (odds ratio, 1.34; P = .01); this observation was made in previous studies. Among European Americans, SNPs at 8q24.21, 11q23.1, and 16q22.1 were significantly associated with CRC, and the odds ratios were of the same magnitude and direction for all SNPs tested, consistent with previously published studies. In contrast, in African Americans, the opposite allele of rs10795668 at 10p14 was associated with colorectal cancer (odds ratio, 1.35; P = .04), and altogether the odds ratios were in the opposite direction for 9 of the 22 SNPs tested. CONCLUSIONS There is genetic heterogeneity in CRC associations in Americans of African versus European descent.

Genome-wide association study to identify novel loci associated with therapy-related myeloid leukemia susceptibility

Therapy-related acute myeloid leukemia (t-AML) is a rare but fatal complication of cytotoxic therapy. Whereas sporadic cancer results from interactions between complex exposures and low-penetrance alleles, t-AML results from an acute exposure to a limited number of potent genotoxins. Consequently, we hypothesized that the effect sizes of variants associated with t-AML would be greater than in sporadic cancer, and, therefore, that these variants could be detected even in a modest-sized cohort. To test this, we undertook an association study in 80 cases and 150 controls using Affymetrix Mapping 10K arrays. Even at nominal significance thresholds, we found a significant excess of associations over chance; for example, although 6 associations were expected at P less than .001, we found 15 (P(enrich) = .002). To replicate our findings, we genotyped the 10 most significantly associated single nucleotide polymorphisms (SNPs) in an independent t-AML cohort (n = 70) and obtained evidence of association with t-AML for 3 SNPs in the subset of patients with loss of chromosomes 5 or 7 or both, acquired abnormalities associated with prior exposure to alkylator chemotherapy. Thus, we conclude that the effect of genetic factors contributing to cancer risk is potentiated and more readily discernable in t-AML compared with sporadic cancer.