Discovery and fine-mapping of height loci via high-density imputation of GWASs in individuals of African ancestry

Although many loci have been associated with height in European ancestry populations, very few have been identified in African ancestry individuals. Furthermore, many of the known loci have yet to be generalized to and fine-mapped within a large-scale African ancestry sample. We performed sex-combined and sex-stratified meta-analyses in up to 52,764 individuals with height and genome-wide genotyping data from the African Ancestry Anthropometry Genetics Consortium (AAAGC). We additionally combined our African ancestry meta-analysis results with published European genome-wide association study (GWAS) data. In the African ancestry analyses, we identified three novel loci (SLC4A3, NCOA2, ECD/FAM149B1) in sex-combined results and two loci (CRB1, KLF6) in women only. In the African plus European sex-combined GWAS, we identified an additional three novel loci (RCCD1, G6PC3, CEP95) which were equally driven by AAAGC and European results. Among 39 genome-wide significant signals at known loci, conditioning index SNPs from European studies identified 20 secondary signals. Two of the 20 new secondary signals and none of the 8 novel loci had minor allele frequencies (MAF) < 5%. Of 802 known European height signals, 643 displayed directionally consistent associations with height, of which 205 were nominally significant (p < 0.05) in the African ancestry sex-combined sample. Furthermore, 148 of 241 loci contained ≤20 variants in the credible sets that jointly account for 99% of the posterior probability of driving the associations. In summary, trans-ethnic meta-analyses revealed novel signals and further improved fine-mapping of putative causal variants in loci shared between African and European ancestry populations.

Meta-analysis of up to 622,409 individuals identifies 40 novel smoking behaviour associated genetic loci

Smoking is a major heritable and modifiable risk factor for many diseases, including cancer, common respiratory disorders and cardiovascular diseases. Fourteen genetic loci have previously been associated with smoking behaviour-related traits. We tested up to 235,116 single nucleotide variants (SNVs) on the exome-array for association with smoking initiation, cigarettes per day, pack-years, and smoking cessation in a fixed effects meta-analysis of up to 61 studies (up to 346,813 participants). In a subset of 112,811 participants, a further one million SNVs were also genotyped and tested for association with the four smoking behaviour traits. SNV-trait associations with P < 5 × 10-8 in either analysis were taken forward for replication in up to 275,596 independent participants from UK Biobank. Lastly, a meta-analysis of the discovery and replication studies was performed. Sixteen SNVs were associated with at least one of the smoking behaviour traits (P < 5 × 10-8) in the discovery samples. Ten novel SNVs, including rs12616219 near TMEM182, were followed-up and five of them (rs462779 in REV3L, rs12780116 in CNNM2, rs1190736 in GPR101, rs11539157 in PJA1, and rs12616219 near TMEM182) replicated at a Bonferroni significance threshold (P < 4.5 × 10-3) with consistent direction of effect. A further 35 SNVs were associated with smoking behaviour traits in the discovery plus replication meta-analysis (up to 622,409 participants) including a rare SNV, rs150493199, in CCDC141 and two low-frequency SNVs in CEP350 and HDGFRP2. Functional follow-up implied that decreased expression of REV3L may lower the probability of smoking initiation. The novel loci will facilitate understanding the genetic aetiology of smoking behaviour and may lead to the identification of potential drug targets for smoking prevention and/or cessation.

Abstract 057: Epigenome Wide Association Study Identifies DNA Methylation Sites Associated With Target Organ Damage in Elderly African Americans

Introduction: Epigenetic mechanisms, including DNA methylation, influence gene expression and mediate responses to the environment over the lifecourse. Previous research has identified a number of DNA methylation sites (CpGs) associated with arteriosclerosis and its risk factors. Given the underlying role of immunological and inflammatory processes in arteriosclerosis and the potential of pleiotropic effects of individual CpG sites across organ systems influenced by arteriosclerosis, we investigated the DNA methylation profiles of peripheral blood leukocytes from a cohort of African American (AA) hypertensive sibships to explore their association with measures of target organ damage (TOD), analyzed using both single and multi-trait (multivariate) models.

Methods: Using the Genome-wide Efficient Mixed Model Association algorithm (GEMMA), we assessed the association between ~790,000 CpGs across the genome and 4 TOD traits: Estimated glomerular filtration rate (eGFR), urinary albumin-creatinine ratio (UACR), left ventricular mass index (LVMI) and relative wall thickness (RWT) in 961 subjects with mean baseline age of 57.5 years (SD: 10.3). DNA methylation was measured at baseline using the Infinium MethylationEPIC BeadChip and TOD was assessed 5 years later.

Results: Single trait models adjusting for age, sex, blood cell proportions, genetic principal components, and time difference between the measures identified a single CpG site associated with UACR (cg04816311), LVMI (cg21134922) and RWT (cg03042953, all FDR < 0.1). Pleiotropic (multivariate) models identified 8 CpGs associated with at least one of the traits in the CATSPERD, C7orf50, ALDH1L1, IFT43 and OAT genes. After adjusting for SBP, DBP and antihypertensive medication use, 6 CpGs remained significant and 1 additional CpG (cg21447579 in SNORD116 gene cluster) was identified. After further adjustment for BMI, T2DM and smoking, two CpGs (cg02264946 in CATSPERD and cg12661888 in IFT43) remained significant.

Conclusions: The study highlights the role of specific epigenetic sites in risk of TOD in an African American population. Additionally, it highlights the utility of using a pleiotropy-based analysis to identify CpG sites by leveraging information across related traits.

Abstract P530: Pleiotropic Effects on Blood Pressure Traits Using Genome-wide Analysis of Gene-alcohol Interactions

We tested for pleiotropy in European ancestry subjects (N>90K) via GWAS of systolic and diastolic blood pressure (BP), mean arterial pressure, and pulse pressure, using gene (G)-alcohol consumption (E) interactions. The approach was a correlated meta-analysis (PMCID-PMC3773990) that combined simultaneously the 4 BP traits genome-wide GxE interactions summary meta-P values. This approach adjusts for correlations among single traits at the genomic level. A variant was considered pleiotropic when the overall correlated meta-analysis yielded P ≤5E-08 and GxE meta- P ≤E-04 for at least two single traits. The novel pleiotropic variants localize in eight loci. TTLL7 (1p31.1) is a tubulin modifier. DYRK3 (1q32.1) is a transcription regulator. MAPKAPK2 (1q32.1) is a stress-activated serine/threonine-protein kinase involved in cytokine production especially for TNF , IL6 and phosphorylates (among others) LSP1 , identified in our GWAS GxE study for individual BP traits. FSTL5 (4q32.2) is annotated as calcium ion binding . A locus at 11q13.1 includes SNX32 , EFEMP2, and FOSL1 . FOSL1 variants may regulate expression of SNX32 . EFEMP2 is implicated in blood coagulation. CATSPER2 (15q15.3) is a cation channel. CCDC151 (19p13.2) is an outer dynein arm assembly. The functions of two other loci (17q22 and 18q22.3) are unknown. We also identified 4 pleiotropic loci ( SGK223 , TNKS , GATA4 , FTO ) that were found significant at our GxE meta-GWAS of single traits in 572K multi-ancestry individuals. In addition, we detected 24 pleiotropic BP-known loci. Some of these genes relate to alcohol consumption (e.g., BLK , GATA4 , FTO ). TNKS , MAPKAPK2 and FSTL5 interact with the Wnt/β-catenin signaling pathway, which contributes to hypertension. Several pleiotropic variants showed features of regulation by locating at promoter and enhancer histone marks, at DNAse, at proteins binding sites and being eQTL. The 36 novel and BP-known loci comprising 86 significant genes were enriched for Hypertension , Cardiac arrhythmias , Myocardial infarction , Atrial fibrillation, and Left ventricular hypertrophy . Our correlated meta-analysis of GxE interaction approach identified novel pleiotropic loci and validated known BP loci, thus providing insights into the mechanisms of hypertension.

Genome-wide association study of cardiac structure and systolic function in African Americans: the Candidate Gene Association Resource (CARe) study

BACKGROUND

Using data from 4 community-based cohorts of African Americans, we tested the association between genome-wide markers (single-nucleotide polymorphisms) and cardiac phenotypes in the Candidate-gene Association Resource study.

METHODS AND RESULTS

Among 6765 African Americans, we related age, sex, height, and weight-adjusted residuals for 9 cardiac phenotypes (assessed by echocardiogram or magnetic resonance imaging) to 2.5 million single-nucleotide polymorphisms genotyped using Genome-wide Affymetrix Human SNP Array 6.0 (Affy6.0) and the remainder imputed. Within the cohort, genome-wide association analysis was conducted, followed by meta-analysis across cohorts using inverse variance weights (genome-wide significance threshold=4.0 ×10(-7)). Supplementary pathway analysis was performed. We attempted replication in 3 smaller cohorts of African ancestry and tested lookups in 1 consortium of European ancestry (EchoGEN). Across the 9 phenotypes, variants in 4 genetic loci reached genome-wide significance: rs4552931 in UBE2V2 (P=1.43×10(-7)) for left ventricular mass, rs7213314 in WIPI1 (P=1.68×10(-7)) for left ventricular internal diastolic diameter, rs1571099 in PPAPDC1A (P=2.57×10(-8)) for interventricular septal wall thickness, and rs9530176 in KLF5 (P=4.02×10(-7)) for ejection fraction. Associated variants were enriched in 3 signaling pathways involved in cardiac remodeling. None of the 4 loci replicated in cohorts of African ancestry was confirmed in lookups in EchoGEN.

CONCLUSIONS

In the largest genome-wide association study of cardiac structure and function to date in African Americans, we identified 4 genetic loci related to left ventricular mass, interventricular septal wall thickness, left ventricular internal diastolic diameter, and ejection fraction, which reached genome-wide significance. Replication results suggest that these loci may be unique to individuals of African ancestry. Additional large-scale studies are warranted for these complex phenotypes.

Smoking and genetic risk variation across populations of European, Asian, and African American ancestry--a meta-analysis of chromosome 15q25

Recent meta-analyses of European ancestry subjects show strong evidence for association between smoking quantity and multiple genetic variants on chromosome 15q25. This meta-analysis extends the examination of association between distinct genes in the CHRNA5-CHRNA3-CHRNB4 region and smoking quantity to Asian and African American populations to confirm and refine specific reported associations. Association results for a dichotomized cigarettes smoked per day phenotype in 27 datasets (European ancestry (N = 14,786), Asian (N = 6,889), and African American (N = 10,912) for a total of 32,587 smokers) were meta-analyzed by population and results were compared across all three populations. We demonstrate association between smoking quantity and markers in the chromosome 15q25 region across all three populations, and narrow the region of association. Of the variants tested, only rs16969968 is associated with smoking (P < 0.01) in each of these three populations (odds ratio [OR] = 1.33, 95% CI = 1.25-1.42, P = 1.1 × 10(-17) in meta-analysis across all population samples). Additional variants displayed a consistent signal in both European ancestry and Asian datasets, but not in African Americans. The observed consistent association of rs16969968 with heavy smoking across multiple populations, combined with its known biological significance, suggests rs16969968 is most likely a functional variant that alters risk for heavy smoking. We interpret additional association results that differ across populations as providing evidence for additional functional variants, but we are unable to further localize the source of this association. Using the cross-population study paradigm provides valuable insights to narrow regions of interest and inform future biological experiments.

Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk

Blood pressure (BP) is a heritable trait¹ influenced by multiple biological pathways and is responsive to environmental stimuli. Over one billion people worldwide have hypertension (BP ≥140 mm Hg systolic [SBP] or ≥90 mm Hg diastolic [DBP])². Even small increments in BP are associated with increased risk of cardiovascular events³. This genome-wide association study of SBP and DBP, which used a multi-stage design in 200,000 individuals of European descent, identified 16 novel loci: six of these loci contain genes previously known or suspected to regulate BP (GUCY1A3-GUCY1B3; NPR3-C5orf23; ADM; FURIN-FES; GOSR2; GNAS-EDN3); the other 10 provide new clues to BP physiology. A genetic risk score based on 29 genome-wide significant variants was associated with hypertension, left ventricular wall thickness, stroke, and coronary artery disease, but not kidney disease or kidney function. We also observed associations with BP in East Asian, South Asian, and African ancestry individuals. Our findings provide new insights into the genetics and biology of BP, and suggest novel potential therapeutic pathways for cardiovascular disease prevention.

Genetic variation in NCAM1 contributes to left ventricular wall thickness in hypertensive families

RATIONALE

Left ventricular (LV) mass and related phenotypes are heritable, important predictors of cardiovascular disease, particularly in hypertensive individuals.

OBJECTIVE

Identify genetic predictors of echocardiographic phenotypes in hypertensive families.

METHODS AND RESULTS

A multistage genome-wide association study (GWAS) was conducted in hypertensive-ascertained black families (HyperGEN, stage I; GENOA, stage II); findings were replicated in HyperGEN white families (stage III). Echocardiograms were collected using a common protocol, and participants were genotyped with the Affymetrix Genome-Wide Human SNP 6.0 Array. The following were analyzed using mixed models adjusted for ancestry: in stages I and II, 1258 and 989 blacks, respectively; and in stage III, 1316 whites. Phenotypes included LV mass, LV internal dimension (LVID), wall thicknesses (posterior [PWT] and intraventricular septum [IVST]), and relative wall thickness (RWT). In stage I, 5 single nucleotide polymorphisms (SNPs) had P≤10(-6). In stage II, 1 SNP (rs1436109; NCAM1 intron 1) replicated with the same phenotype (PWT, P=0.025) in addition to RWT (P=0.032). In stage III, rs1436109 was associated with RWT (P=5.47×10(-4)) and LVID (P=1.86×10(-4)). Fisher combined probability value for all stages was RWT=3.80×10(-9), PWT=3.12×10(-7), IVST=8.69×10(-7), LV mass=2.52×10(-3), and LVID=4.80×10(-4).

CONCLUSIONS

This GWAS conducted in hypertensive families identified a variant in NCAM1 associated with LV wall thickness and RWT. NCAM is upregulated during the remodeling period of hypertrophy to heart failure in Dahl salt-sensitive rats. Our initial screening in hypertensive blacks may have provided the context for this novel locus.

Association of novel risk factors with the ankle brachial index in African American and non-Hispanic white populations

OBJECTIVES

To investigate whether novel risk factors, including C-reactive protein (CRP), fibrinogen, lipoprotein(a) [Lp(a)], and homocysteine levels, are associated with the ankle brachial index (ABI) in African American and non-Hispanic white populations and whether novel risk factors account for ethnic differences in peripheral arterial disease (PAD).

PARTICIPANTS AND METHODS

Between December 2000 and October 2004, original participants in the Genetic Epidemiology Network of Arteriopathy study returned for a second study visit to undergo measurement of risk factors and ABI. The CRP, Lp(a), and homocysteine levels were log transformed to reduce skewness. Multivariable regression analyses were used to assess whether a novel risk factor was associated with ABI after adjustment for conventional risk factors and whether ethnicity was associated with PAD (ABI, <or=0.95) after adjustment for conventional and novel risk factors.

RESULTS

Of 2229 study participants, the ABI was determined in 1395 African American participants (mean +/- SD age, 63 +/- 9 years; 71% women) and 834 white participants (mean +/- SD age, 58 +/- 9 years; 62% women) who belonged to hypertensive sibships. The mean ABI was lower in African American than in white individuals (0.99 +/- 0.1 vs 1.13 +/- 0.1; P < .001). In both ethnic groups, higher levels of CRP, fibrinogen, and homocysteine were each associated with a lower ABI after adjustment for conventional risk factors. In African American participants, the Lp(a) level was also significantly associated with the ABI. African American ethnicity was associated with the presence of PAD after adjustment for conventional risk factors (men: odds ratio [OR], 3.04; 95% confidence interval [CI], 1.80-5.15; women: OR, 2.82; 95% CI, 1.85-4.29), but the risk was significantly attenuated after additional adjustment for novel risk factors (men: OR, 2.11; 95% CI, 1.21-3.70; women: OR, 1.98; 95% CI, 1.26-3.11).

CONCLUSION

Novel risk factors are associated with interindividual variation in ABI in African American and non-Hispanic white populations and partly account for the increased risk of PAD associated with African American ethnicity.

Lack of association between lipoprotein(a) and coronary artery calcification in the Genetic Epidemiology Network of Arteriopathy (GENOA) study

OBJECTIVE

To investigate the relationship between lipoprotein(a) [Lp(a)] levels and the extent of coronary atherosclerosis in a cohort that consisted predominantly of hypertensive patients.

PATIENTS AND METHODS

Patients were ascertained through sibships that contained at least 2 individuals with essential hypertension diagnosed before the age of 60 years. The 10-year coronary heart disease (CHD) risk was estimated on the basis of the Framingham risk equation. Serum Lp(a) was measured by an immunoturbidimetric assay. Coronary artery calcification (CAC) was measured noninvasively by electron beam computed tomography and CAC score calculated using the Agatston score.

RESULTS

Patients included 765 non-Hispanic, white individuals (59% women) participating in the Genetic Epidemiology Network of Arteriopathy study. The mean +/- SD age of the patients was 62 +/- 8 years, and 77% had hypertension. The prevalence of detectable CAC was 87% in men and 60% in women. The CAC scores did not differ significantly across quintiles of Lp(a) levels in either men or women. In a multiple regression model that included conventional risk factors, Lp(a) levels were not related to CAC quantity in either sex. No significant interactions were noted between Lp(a) levels and the conventional risk factors in the prediction of CAC quantity. When stratified on the basis of the 10-year CHD risk, 26.5% of the patients were low risk (< 6%), 60.5% were intermediate risk (6%-20%), and 12.9% were high risk (> 20%). Lipoprotein(a) was not associated with CAC quantity within subgroups based on 10-year CHD risk.

CONCLUSION

In this cohort enriched with hypertensive patients, the estimated 10-year CHD risk did not appear to modify the lack of an association between Lp(a) levels and CAC.

Lack of genetic linkage evidence for a trans-acting factor having a large effect on plasma lipoprotein[a] levels in African Americans

The distribution of plasma lipoprotein[a] (Lp[a]) concentrations, a risk factor for cardiovascular disease, varies greatly among racial groups, with African Americans having values that are shifted toward higher levels than those of whites. The underlying cause of this heterogeneity is unknown, but a role for "trans-acting" factors has been hypothesized. This study used genetic linkage analysis to localize genetic factors influencing Lp[a] levels in African Americans that were absent in other populations; linkage results were analyzed separately in non-Hispanic whites, Hispanic whites, and African Americans. As expected, all three samples showed highly significant linkage at the approximate location of the lysophosphatidic acid locus. The white populations also independently had regions of significant linkage on chromosome 19 (LOD 3.80) and suggestive linkage on chromosomes 12 (LOD 1.60), 14 (LOD 2.56), and 19 (LOD 2.52). No linkage evidence was found to support the hypothesis of another single gene with large effects specifically segregating in African Americans that may account for their elevated Lp[a] levels.

Autosomal genome-wide scan for coronary artery calcification loci in sibships at high risk for hypertension

Coronary artery disease (CAD) is the leading cause of mortality in the developed world. Although several CAD risk factors, including measures of lipid metabolism, obesity, and blood pressure, have a genetic basis, many genes for CAD susceptibility have yet to be identified. Coronary atherosclerosis is the major cause of CAD, but many with coronary atherosclerosis lack symptoms. Thus, a major limitation of using symptomatic CAD endpoints (eg, sudden coronary death, myocardial infarction) as a study outcome is substantial disease misclassification. Coronary artery calcification (CAC) is part of the atherosclerotic process and is an independent predictor of CAD endpoints. In the present study, CAC was noninvasively quantified by electron beam computed tomography. We performed genome-wide multipoint mode-of-inheritance-free linkage analysis on affected sib pairs, defined as being > or = the 70th sex- and age-specific percentile for CAC quantity, in a sample of 29 families enriched for hypertension. Almost 95% of participants were asymptomatic for CAD. Our LOD score (log10 odds in favor of linkage) results provide evidence that chromosomal regions 6p21.3 (maximum LOD score=2.22, P=0.00070) and 10q21.3 (maximum LOD score=3.24, P=0.000057) may harbor genes associated with subclinical coronary atherosclerosis.

A comparative molecular analysis of developing mouse forelimbs and hindlimbs using serial analysis of gene expression (SAGE)

The analysis of differentially expressed genes is a powerful approach to elucidate the genetic mechanisms underlying the morphological and evolutionary diversity among serially homologous structures, both within the same organism (e.g., hand vs. foot) and between different species (e.g., hand vs. wing). In the developing embryo, limb-specific expression of Pitx1, Tbx4, and Tbx5 regulates the determination of limb identity. However, numerous lines of evidence, including the fact that these three genes encode transcription factors, indicate that additional genes are involved in the Pitx1-Tbx hierarchy. To examine the molecular distinctions coded for by these factors, and to identify novel genes involved in the determination of limb identity, we have used Serial Analysis of Gene Expression (SAGE) to generate comprehensive gene expression profiles from intact, developing mouse forelimbs and hindlimbs. To minimize the extraction of erroneous SAGE tags from low-quality sequence data, we used a new algorithm to extract tags from -analyzed sequence data and obtained 68,406 and 68,450 SAGE tags from forelimb and hindlimb SAGE libraries, respectively. We also developed an improved method for determining the identity of SAGE tags that increases the specificity of and provides additional information about the confidence of the tag-UniGene cluster match. The most differentially expressed gene between our SAGE libraries was Pitx1. The differential expression of Tbx4, Tbx5, and several limb-specific Hox genes was also detected; however, their abundances in the SAGE libraries were low. Because numerous other tags were differentially expressed at this low level, we performed a 'virtual' subtraction with 362,344 tags from six additional nonlimb SAGE libraries to further refine this set of candidate genes. This subtraction reduced the number of candidate genes by 74%, yet preserved the previously identified regulators of limb identity. This study presents the gene expression complexity of the developing limb and identifies candidate genes involved in the regulation of limb identity. We propose that our computational tools and the overall strategy used here are broadly applicable to other SAGE-based studies in a variety of organisms. [SAGE data are all available at GEO (http://www.ncbi.nlm.nih.gov/geo/) under accession nos. GSM55 and GSM56, which correspond to the forelimb and hindlimb raw SAGE data.]

Identification and prevention of a GC content bias in SAGE libraries

Serial Analysis of Gene Expression (SAGE) is becoming a widely used gene expression profiling method for the study of development, cancer and other human diseases. Investigators using SAGE rely heavily on the quantitative aspect of this method for cataloging gene expression and comparing multiple SAGE libraries. We have developed additional computational and statistical tools to assess the quality and reproducibility of a SAGE library. Using these methods, a critical variable in the SAGE protocol was identified that has the potential to bias the Tag distribution relative to the GC content of the 10 bp SAGE Tag DNA sequence. We also detected this bias in a number of publicly available SAGE libraries. It is important to note that the GC content bias went undetected by quality control procedures in the current SAGE protocol and was only identified with the use of these statistical analyses on as few as 750 SAGE Tags. In addition to keeping any solution of free DiTags on ice, an analysis of the GC content should be performed before sequencing large numbers of SAGE Tags to be confident that SAGE libraries are free from experimental bias.

Genome-wide linkage analysis reveals evidence of multiple regions that influence variation in plasma lipid and apolipoprotein levels associated with risk of coronary heart disease

Results of genome-wide linkage analyses to identify chromosomal regions that influence interindividual variation in plasma lipid and apolipoprotein levels in the Rochester, Minn, population are reported. Analyses were conducted for total cholesterol (total-C), triglycerides (TGs), high density lipoprotein cholesterol (HDL-C), apolipoprotein A-I, apolipoprotein A-II, apolipoprotein B, apolipoprotein C-II, apolipoprotein C-III, apolipoprotein E, the total-C/HDL-C ratio, and the TG/HDL-C ratio. Genotypes were measured for 373 genome-wide marker loci on 1484 individuals distributed among 232 multigeneration pedigrees sampled without regard to health status. LOD scores and estimates of additive genetic variance associated with map locations were obtained by using the variance-component method of linkage analysis. No evidence of linkage with genes influencing variation in age served as a negative control. Plasma apolipoprotein E levels and the apolipoprotein E gene served as a positive control (LOD score 4.20). Evidence (LOD score >2.00) was provided that was suggestive of a gene or genes on chromosomes 4 and 5 influencing variation in the apolipoprotein A-II level, on chromosome 12 influencing variation in the apolipoprotein A-I level, and on chromosome 17 influencing variation of total-C/HDL-C. These analyses provide new information about genomic regions in humans that influence interindividual variation in plasma lipid and apolipoprotein levels and serve as a basis for further fine-mapping studies to identify new genes involved in lipid metabolism.

A combinatorial partitioning method to identify multilocus genotypic partitions that predict quantitative trait variation

Recent advances in genome research have accelerated the process of locating candidate genes and the variable sites within them and have simplified the task of genotype measurement. The development of statistical and computational strategies to utilize information on hundreds -- soon thousands -- of variable loci to investigate the relationships between genome variation and phenotypic variation has not kept pace, particularly for quantitative traits that do not follow simple Mendelian patterns of inheritance. We present here the combinatorial partitioning method (CPM) that examines multiple genes, each containing multiple variable loci, to identify partitions of multilocus genotypes that predict interindividual variation in quantitative trait levels. We illustrate this method with an application to plasma triglyceride levels collected on 188 males, ages 20--60 yr, ascertained without regard to health status, from Rochester, Minnesota. Genotype information included measurements at 18 diallelic loci in six coronary heart disease--candidate susceptibility gene regions: APOA1--C3--A4, APOB, APOE, LDLR, LPL, and PON1. To illustrate the CPM, we evaluated all possible partitions of two-locus genotypes into two to nine partitions (approximately 10(6) evaluations). We found that many combinations of loci are involved in sets of genotypic partitions that predict triglyceride variability and that the most predictive sets show nonadditivity. These results suggest that traditional methods of building multilocus models that rely on statistically significant marginal, single-locus effects, may fail to identify combinations of loci that best predict trait variability. The CPM offers a strategy for exploring the high-dimensional genotype state space so as to predict the quantitative trait variation in the population at large that does not require the conditioning of the analysis on a prespecified genetic model.

The glutathione S-transferase-mu and -theta genotypes in the etiology of prostate cancer: genotype-environment interactions with smoking

It has been reported that individuals who express GSTT1, the gene coding for the theta class of the glutathione S-transferases (GSTs), have an elevated risk of prostate cancer (CaP). This result is supported by studies that show glutathione conjugation of some xenobiotics by the GSTs can produce mutagenic intermediates. However, the potential role of environmental factors in modifying the risk of CaP conferred by GSTT1 is not known. We investigated whether there was an interaction between smoking and the non-deleted genotypes of the mu (GSTM1) and theta (GSTT1) GST genes using a clinic-based study of 276 CaP cases and 499 controls. We observed no main effect of smoking (odds ratio, 0.95; confidence interval, 0.69-1.29) or GSTM1 (odds ratio, 1.00; confidence interval, 0.73-1.36) with CaP, but did observe a statistically significant main effect of GSTT1 with CaP (odds ratio, 1.61; confidence interval, 1.14-2.28) as reported previously. No interaction between smoking and GSTM1 was observed. A significant increase in the probability of having CaP was observed in men who were both smokers and carried a non-deleted GSTT1 genotype compared with men who had neither or only one of these risk factors (P = 0.049). Approximately 30.9% of CaP cases in this study could be attributed to the smoking x GSTT1 interaction. Whereas the mechanism of this interaction is not known, it is plausible that the metabolism of carcinogenic intermediates or the response to chronic inflammation associated with smoking may be modulated by the GSTT1 genotype and may modify CaP risk.

Syndrome X: is it for real?

The term syndrome X has been applied to the association of hypertension, non-insulin-dependent diabetes mellitus (NIDDM), android obesity, insulin resistance, and dyslipidemia. In this paper, based on population samples from Tecumseh, Michigan, and Hiroshima, Japan, characterized by persons > or = 40 years of age, we examine the validity of regarding this constellation of traits as a true syndrome, i.e., an array of traits with a single, unifying pathophysiology underlying its components. Data were not available on insulin resistance and dyslipidemia, and obesity was expressed as body mass index (BMI) without the division into android and non-android types. The four ethnic-gender data sets were analyzed on the basis of two age classes, age > or = 40 years and age > or = 50 years, and two obesity classes, BMI > or = 27 and > or = 30. A simple chi 2 test of goodness-of-fit under a model of independence revealed non-random associations between hypertension, NIDDM, and BMI which were in part attributable to an excess of persons with all three traits. However, when the four data sets were subjected to separate log-linear analyses of the three-way association tables, none of the three-factor interaction terms (i.e., syndrome X) was significant. High significance was, however, observed in the two-factor interaction term for BMI*hypertension. It is concluded that the significant association between these three traits is driven by the BMI*hypertension interaction, and there is no evidence in these data sets of a significant role for a syndrome X. Genet.

Association and linkage analysis of the alpha-adducin gene and blood pressure

In Milan hypertensive rats, a variant in the alpha-adducin gene has been shown to account for approximately 50% of the interindividual variation in blood pressure levels between these animals and their normotensive counterparts. Additional studies have suggested that a polymorphism within exon 10 of the human alpha-adducin gene (Gly-460-Trp) may be associated with hypertension and salt sensitivity. On the basis of these observations, we investigated variation within or near the human alpha-adducin gene for linkage and association with a locus influencing blood pressure levels in 281 nuclear families (774 siblings aged 5 to 37 years; 380 parents aged 26 to 57 years), selected from the white population of Rochester, Minnesota, without regard to health. Sib pair linkage analyses (n = 852 sibling pairs) using a dinucleotide repeat marker (D4S43) that maps approximately 660 kb from the alpha-adducin gene provided no evidence of linkage between this marker locus and a locus influencing systolic, diastolic, or mean blood pressure levels. Allele frequencies for the Gly-460-Trp polymorphism were similar to those reported in other white populations (Gly = 0.812, Trp = 0.188); however, this polymorphism was not associated with any measure of blood pressure level in either parents or siblings. Therefore, variation within the alpha-adducin gene does not appear to have a major influence on measures of blood pressure in white families from Rochester, Minnesota.

Context-dependent genetic effects in hypertension

There is a growing frustration with the limitations and inconsistencies of single locus candidate gene association and linkage studies. This frustration is exacerbated by the knowledge that a large influx of genotypic and gene expression data is expected to emerge over the next 5 years, and we are not prepared for the type of multigenic conceptual framework that will be necessary to analyze that data. A review of the hypertension genetic literature reveals substantial evidence for the importance of both genetic and environmental contexts on the mapping between single locus polymorphisms and risk of disease. These trends indicate that the current reliance on simple single gene studies to elucidate the complex etiology of hypertension needs to undergo some kind of transformation. It is suggested that even a minor shift to a more systematic investigation of context-dependent effects will increase our understanding of the multidimensional genetic and environmental realities underneath current studies. This shift is also necessary if we are to gain a deeper understanding of the specific ways in which an individual's risk of hypertension is a consequence of the inter- actions among genes and environments.

Utility of the predictors of coronary heart disease mortality in a longitudinal study of elderly Finnish men aged 65 to 84 years is dependent on context defined by Apo E genotype and area of residence

A common assumption underlying most genetic studies is that individuals with different genotypes respond similarly to exposure to internal (epigenetic and background genotype effects) and external (ecological) environments. Here we evaluate whether this assumption is true in individuals with different genotypes of the gene coding for the apolipoprotein E (Apo E) molecule, an important determinant of the metabolic fate of plasma lipids and lipoproteins. We addressed whether the utility of known risk factors of coronary heart disease (CHD) in the prediction of CHD death in a 5-year follow-up is the same for the two most common Apo E genotypes, epsilon3/3 and epsilon4/3, in two cohorts of elderly Finnish men (age at baseline: 65-84 years), one in Eastern and the other in Southwestern Finland. The CHD mortality rate was higher in the epsilon4/3 than in the epsilon3/3 genotype in both cohorts (11.1 versus 7.8%, Pr = 0.281 in the Eastern cohort and 19.6 versus 8.2%, Pr = 0.002 in the Southwestern cohort). In the Eastern cohort, serum high density lipoprotein (HDL) cholesterol level was identified as a strong predictor of CHD death in the epsilon3/3 genotype (beta = -2.155, Pr = 0.019). In the Southwestern cohort, age (beta = 0.139, Pr = 0.006), body mass index (BMI) (beta = 0.149, Pr = 0.016), and serum total cholesterol level (beta = 0.453, Pr = 0.051) were identified as strong predictors in the epsilon3/3 genotype, as were smoking (beta = 0.236, Pr = 0.008) and BMI (beta = -0.124, Pr = 0.057) in the epsilon4/3 genotype. The latter observation indicates that in Southwestern Finland the probability of CHD death decreases with increasing BMI in elderly men with the epsilon4/3 genotype, while in their counterparts with the epsilon3/3 genotype the risk increases with increasing BMI. This difference was statistically significant (Pr = 0.002). These observations clearly argue against the assumption that individuals with different genotypes respond similarly to exposures to internal and/or external environments. These observations are consistent with a complex pathobiology of CHD involving biochemical and physiological agents that are under the influence of interactions between genetic and environmental factors. Information about these interactions is necessary for developing a more precise risk assessment and ultimately to improve public health and clinical strategies to prevent this devastating disease both at the individual and population levels.

Influence of apolipoprotein E genotype variation on the means, variances, and correlations of plasma lipids and apolipoproteins in children

The impact of the three most common apolipoprotein E (APOE) genotypes (epsilon 32, epsilson 33, and epsilon 43) on means, variances, and correlations of nine plasma lipid and apolipoprotein traits (total cholesterol, InTriglycerides, HDL cholesterol, and apolipoproteins AI, AII, B, CII, CIII, and InE) was studied in 212 unrelated female and 219 unrelated male children aged 5-21.5 years from 278 pedigrees ascertained without regard to health status from Rochester, Minnesota. There was significant heterogeneity (p < or = 0.05) among genotypes for the mean plasma levels of InApo E, Apo CII, Apo CIII, and InTriglycerides (InTrig) in females, and for the means of InApo E, Apo B, and total cholesterol (Total-C) in males. Significant heterogeneity of intragenotypic variance was observed in males for Apo CII, InTrig, and HDL-C; no significant heterogeneity was observed in females. Pairwise correlations between traits differed significantly among APOE genotypes in both females (6 of 36 pairs) and males (5 of 36 pairs). These results differ from those obtained from studies of the parental generation from the same sample of pedigrees. Our study further demonstrates that, with the exception of mean InApo E levels, the univariate and bivariate distributions of traits that are measures of lipoprotein metabolism are influenced by variation in the APOE gene in a gender- and generation-dependent manner.

The relationship between risk factor levels and presence of coronary artery calcification is dependent on apolipoprotein E genotype

An important research question in the study of the genetics of coronary artery disease (CAD) is whether information about genetic variation will improve our ability to predict CAD beyond established risk factors. This question is especially relevant to the goal of identifying young, asymptomatic adults with coronary atherosclerosis who would benefit most from interventions to reduce risk. Coronary artery calcification (CAC) detected by electron-beam computed tomography is a relatively new method for detecting coronary atherosclerosis in asymptomatic individuals that has been shown to be a more accurate indicator of coronary atherosclerosis in asymptomatic individuals than other noninvasive techniques. In a study of asymptomatic women (n=169) and men (n=160) between the ages of 20 and 59 representative of the Rochester, Minnesota population, we used logistic regression to ask whether the most common Apolipoprotein (Apo) E genotypes (epsilon3/2, epsilon3/3, and epsilon4/3) predict the presence of CAC. The addition of information about ApoE genotypes to logistic models containing each separate risk factor did not improve prediction of CAC (P>0.10 in both women and men). However, there was significant evidence (P<0.10) that associations between variation in the probability of having CAC and variation in body mass index, plasma total cholesterol, and plasma ApoB in men and body mass index, plasma triglycerides, plasma ApoA1, and plasma ApoE in women were dependent on ApoE genotype. Thus, variation in the gene coding for ApoE may play a role in determining the contribution of established risk factors to risk of CAC.

Correlates of family history of coronary artery disease in children

The atherosclerotic process begins in childhood but, in general, does not reach the clinical horizon until after the fifth decade of life, at which point the best opportunities for prevention and intervention have been lost. In order to identify children with a high risk of developing coronary artery disease (CAD), risk factors measured in children that are the most informative indicators of future risk must be identified. Using a novel analytical strategy that incorporates a continuum of information about context dependency, we investigated whether there were significant differences in intermediate biochemical and physiological traits between children (189 females and 188 males, ages 5-20.5 years) with and without a strong family history of clinically-defined CAD at three levels of context dependency (coarse grain, medium grain, and fine grain). In the coarse-grained analysis we tested for differences in mean levels of nine intermediate traits (lipids, apolipoproteins, blood pressure traits) and indices of external and internal environmental context (age, body mass index, smoking status). Female children with a strong family history had higher average levels for total cholesterol, triglyceride, Apo B, and systolic blood pressure and were on average older and weighed more than female children with a weak family history of CAD. Male children with a strong family history of CAD had higher average levels of triglycerides and were on average older than male children with a weak family history. In the medium-grained analysis we investigated whether the regression relationships between each intermediate trait and each measure of environmental context was significantly different between children with and without a strong family history of CAD. Our results indicate that children with a strong family history of CAD have a significantly different relationship between their intermediate traits and environmental contexts than children with a weak family history. In the fine-grained analysis, we stratified the sample into age, BMI, and smoking subgroups and tested for mean differences in the intermediate traits between children with and without a strong family history. For seven of the nine intermediate traits we found evidence of significant mean differences between children with and without a strong family history of CAD in particular age and BMI subgroups in nonsmokers that were not expected given the results from separate age-dependent or BMI-dependent marginal analyses. From these analyses, we conclude that the inferences about intermediate biochemical and physiological trait associations with family history of CAD depend on where on the coarse-grain to fine-grain continuum of context dependency the analysis is performed. In many cases, inferences at one level of investigation are different than the inferences made at a coarser or finer level. This study documents the complexity of the associations between intermediate traits and risk of CAD and raises the question of how many models are needed to maximize disease prediction and where these models should fall on the coarse- to fine-grain continuum.

Relationship between renal plasma flow response to angiotensin II and blood pressure in a population-based sample

OBJECTIVE

To assess whether interindividual variation in renal plasma flow or in its response to angiotensin II infusion is associated with interindividual differences in blood pressure in a population-based sample of 287 non-Hispanic whites (143 women and 144 men), aged 20-49.9 years.

METHODS

After seven days of eating a high-sodium diet (260 mmol/day), the renal plasma flow was determined by measuring the clearance of p-aminohippurate before and after infusion of 3 ng/kg per min angiotensin II. Multiple linear regression methods were used to assess whether measures of the renal plasma flow and of its response to angiotensin II infusion were predictive of systolic or diastolic blood pressures measured prior to administration of the high-sodium diet, on day 6 of the high-sodium diet, or during the renal clearance procedure on day 7 prior to angiotensin II infusion.

RESULTS

There was some evidence that measures of the renal plasma flow and of its response to angiotensin II infusion during the high-sodium diet were statistically significant predictors of measures of blood pressure in women; there was less evidence for this for blood pressures in men. Interindividual variation in measures of the renal plasma flow and of its response to angiotensin II infusion explained less than 10% of the interindividual variation in any measure of the blood pressure in both sexes.

CONCLUSION

These results suggest that interindividual variation in renal plasma flow ad in its response to angiotensin II infusion during a high-sodium diet will be of limited utility in elucidating the basis for interindividual differences in blood pressure.

The response of renal plasma flow to angiotensin II infusion in a population-based sample and its association with the parental history of essential hypertension

BACKGROUND

Results from previous studies suggested that a blunted response of renal plasma flow (RPF) to angiotensin II infusion during a high-sodium diet (a phenotype associated with nonmodulation) is an intermediate phenotype for essential hypertension.

OBJECTIVE

To determine whether RPF traits used to investigate nonmodulation have the characteristics of intermediate traits when examined in a population-based sample of adults aged 20-49.9 years.

DESIGN AND METHODS

We examined the frequency distribution of baseline RPF and of its response to All infusion using maximum-likelihood commingling analysis in order to investigate the null hypothesis that the distributions of these traits are unimodal. We also examined the null hypothesis that there is no association between these candidate intermediate traits and the parental history of essential hypertension.

RESULTS

There was some evidence for the commingling of multiple distributions underlying these traits both for women and for men but the commingled distributions overlapped substantially and the inferences about the commingling of distributions were sensitive to the method of RPF measurement, exclusion of outliers, and the method of adjustment for concomitants. There was no statistically significant association between any of the RPF traits and a parental history of essential hypertension.

CONCLUSIONS

There is not sufficiently strong evidence to advocate the use of this set of intermediate traits to identify high-risk individuals or to relate genetic variation to the variation in risk of essential hypertension within this age range in the population at large.