Quintile Of Genetic Risk Score Research Articles

Abstract P554: Association of a Genetic Risk Score With Cholesterol and Cholesterol Control: Findings From the Hispanic Community Heath Study/Study of Latinos

Background: Hypercholesterolemia is a prevalent cardiovascular (CVD) risk factor among Hispanics/Latinos; however, rates of cholesterol awareness and treatment are low in this population. Using a genetic risk score (GRS) associated with low-density lipoprotein cholesterol (LDL) levels may help identify those who would benefit most from statin therapy. A 223-SNP GRS was previously shown to be associated with baseline LDL levels and CVD risk in European, African, and East Asian populations, but it was not evaluated for use in a Hispanic/Latino population. Methods: The Hispanic Community Health Study/Study of Latinos (HCHS/SOL) is a prospective study that enrolled 16,415 Hispanic/Latino adults aged 18-74 years from four US communities in 2008-2011. We assessed the association of a 223-SNP GRS with baseline LDL levels and cholesterol control among 8,868 HCHS/SOL participants with complete data and consent to conduct genetic research. A weighted GRS was calculated using effect sizes estimated from the Million Veteran Program Genome Wide Association Study that included about 25,000 Hispanic/Latino participants. Multivariable linear regression analysis was used to derive effect size estimates (Betas (β), 95% confidence intervals (CI)) for the association between the GRS, modeled continuously and by quintiles, and LDL levels. Multivariable logistic regression was used to derive odds ratios and 95% CI for the association between GRS and LDL control in statin users. LDL control was defined as LDL <130mg/dL or <100 mg/dL in those with diabetes. Models were adjusted for predefined confounders and accounted for survey weights. Results: For each standard deviation increase in GRS, LDL levels increased on average by 8.61mg/dL (95% CI 7.37, 9.83) when adjusted by age, sex, and the first 5 principal components. GRS associations with LDL differed by strata of sex, Hispanic/Latino background group, and European genetic ancestry proportion. In 1,114 statin users, odds of cholesterol control were 38% less likely for those in the highest quintile of GRS compared to the lowest quintile, although not significant (OR 0.62, 95% CI 0.36, 1.08). Conclusion: The 223-SNP GRS shows promise to identify those who are more likely to develop higher LDL and may benefit most from statin therapy in a Hispanic/Latino population. However, additional studies are needed in a larger population of Hispanic/Latino statin users for confirmation of these results.

A higher burden of multiple sclerosis genetic risk confers an earlier onset

Background: Age at onset of multiple sclerosis (MS) is an objective, influential predictor of the evolution of MS independent of disease duration. Objectives: Determine the influence of MS genetic predisposition on age of onset. Methods: We conducted a comprehensive investigation of MS risk variants and age at onset in 3495 non-Latinx white individuals, including for combinations of HLA-DRB1*15:01 alleles and quintiles of an unweighted genetic risk score (GRS) for 198 of 200 autosomal MS risk variants that reside outside the major histocompatibility complex. Results: The mean age at onset was 32 years, 29% were male, and 46% were HLA-DRB1*15:01 carriers. For those with the greatest genetic risk burden (the highest GRS quintile with two HLA-DRB1*15:01 alleles) were on average 5 years younger at onset (p = 0.002) than those with the lowest genetic risk burden (the lowest GRS quintile with no HLA-DRB1*15:01 alleles). There was a strong inverse relationship between the MS genetic risk burden and age at onset of MS (p < 5 × 10−8). Conclusion: We demonstrate a significant gradient between elevated MS genetic risk burden and an earlier onset of MS, suggesting that a higher MS genetic risk burden accelerates onset of the disease.

Targeted Coronary Artery Calcium Screening in High-Risk Younger Individuals Using Consumer Genetic Screening Results

ObjectivesThe goal of this study was to assess the utility of a genetic risk score (GRS) in targeted coronary artery calcium (CAC) screening among young individuals. BackgroundEarly CAC screening and preventive therapy may reduce long-term risk of a coronary heart disease (CHD) event. However, identifying younger individuals at increased risk remains a challenge. GRS for CHD are age independent and can stratify individuals on various risk trajectories. MethodsUsing 142 variants associated with CHD events, we calculated a GRS in 1,927 individuals in the CARDIA (Coronary Artery Risk Development in Young Adults) cohort (aged 32 to 47 years) and 6,600 individuals in the MESA (Multi-Ethnic Study of Atherosclerosis) cohort (aged 44 to 87 years). We assessed GRS utility to predict CAC presence in the CARDIA cohort and stratify individuals of varying risk for CAC presence over the lifetime in both cohorts. ResultsThe GRS predicted CAC presence in CARDIA males. It was not predictive in CARDIA females, which had a CAC prevalence of 6.4%. In combined analysis of the CARDIA and MESA cohorts, the GRS was predictive of CAC in both males and females and was used to derive an equation for the age at which CAC probability crossed a predetermined threshold. When assessed in combination with traditional risk factors, the GRS further stratified individuals. For individuals with an equal number of traditional risk factors, probability of CAC reached 25% approximately 10 years earlier for those in the highest GRS quintile compared to the lowest. ConclusionsThe GRS may be used to target high-risk younger individuals for early CAC screening.

Polygenetic risk for coronary artery disease increases hospitalization burden and mortality

BackgroundCoronary artery disease (CAD) is a leading cause of death worldwide and increasing cost for society. Genome wide association studies (GWAS) have identified common variants associated with CAD. Combining single nucleotide polymorphisms (SNPs) into a genetic risk score (GRS) can estimate an individual's genetic burden. ObjectivesTo investigate whether GRS for CAD can predict hospitalization and mortality. Methods23,594 individuals without CAD at baseline and with full data for all covariates from the population based prospective study Malmö diet and cancer study were investigated. The association between hospitalizations was calculated by negative binomial regression and risk of mortality was calculated by Cox proportional hazards regression. The GRS was constructed from 50 SNPs. ResultsThe study population was divided into quintiles according to the value of GRS. During the mean follow-up time of 17.8 years, 17,254 individuals were hospitalized at least once. Individuals in the highest quintile of GRS were hospitalized 10% more often than individuals in the lowest quintile (IRR: 1.10 [95% CI 1.04–1.16], p = 0.001), mainly for cardiovascular reasons (IRR: 1.31 [95% CI 1.20–1.43], p = 5.17 × 10−10). These individuals had highly increased risk of CVD mortality (HR: 1.44 [1.25–1.66], p = 6.56 × 10−7) but not the risk of mortality due to other causes. ConclusionOur results suggest that genetic predisposition for CAD can predict hospitalization burden and mortality, especially due to cardiovascular causes, independently of traditional risk factors. As the risk conferred by the GRS is partially modifiable, our results may help to reduce societal costs, individual suffering and prolong life.

Does genetic risk of obesity modify associations between the neighbourhood environment and body-mass index? A cross-sectional analysis of the UK Biobank dataset

BackgroundIn studies of gene–environment interactions and obesity, the term environment usually refers to individual behavioural factors that influence energy balance. However, evidence suggests that the built environments in which individuals live, work, and play are associated with obesity and may therefore be environmental triggers of genetic risk. No studies have explored whether individual response to such environmental determinants of obesity varies with genetic risk. We aimed to examine whether genetic risk of obesity modifies associations between various neighbourhood characteristics and body-mass index (BMI). MethodsOur sample included 245 532 adults aged 40–69 years from UK Biobank with cross-sectional data on neighbourhood exposures, BMI, potential confounders, and BMI-linked single nucleotide polymorphisms (SNPs). We derived a composite measure of residential neighbourhood environment based on proximity of fast-food stores, availability of physical activity facilities, and neighbourhood green-ness. We also examined each component separately. We constructed a genetic risk score (GRS) as the weighted average of 70 SNPs, where weights were published estimates of genetic effect sizes, scaled to account for the number of available SNPs. In adjusted mixed-effects regression models of neighbourhood environment–BMI associations we examined possible effect modification by genetic risk by testing GRS-by-neighbourhood measure interaction terms and estimating GRS-stratified associations. FindingsSome evidence of an interaction with genetic risk was observed for the composite measure (pinteraction=0·040); the magnitude of the main association with BMI was somewhat larger among people in the highest genetic risk quintile (0·12 lower BMI for each unit increase in composite score [95% CI −0·15 to −0·08]) compared with those in the lowest GRS quintile (β=–0·07; 95% CI −0·11 to −0·03). Less evidence of interaction was observed for the component neighbourhood exposures (fast-food proximity, availability of PA facilities, and neighbourhood greenness), but some results were sensitive (particularly fast-food proximity findings) to alternative model specification and GRS definitions. InterpretationThis preliminary analysis suggests that any gene–environment interaction is fairly weak. People at some levels of genetic risk of obesity might be more sensitive to some neighbourhood characteristics in combination. It is possible that obesogenic environments are so pervasive that low genetic risk affords limited protection. Results were sensitive to analytical decisions, and further examination of this possible gene–environment interaction is necessary. Measurement error in the neighbourhood exposures and UK Biobank's low response fraction remain potential sources of bias. FundingCommonwealth Scholarship Commission.

Abstract P094: Genetic Risk, Incident Coronary Heart Disease Events, and the Benefits of a Healthy Lifestyle: Joint and Interacting Effects Across Four US Cohorts

Background: Recent GWAS revealed single nucleotide polymorphisms (SNPs) that cumulatively increase risk of CHD in those of European ancestry. We quantified the magnitude of preventive effects of healthy lifestyle against the countervailing effects of genetic risk on incident CHD, and assessed potential interactions. Methods: We extracted SNPs of genome-wide significance (&lt;5 x 10 -8 ) for CHD from CARDIoGRAMplusC4 GWAS, and created a weighted 57-SNP genetic risk score (GRS). Baseline covariates, healthy lifestyle score (including diet pattern, physical activity, smoking and BMI), and incident CHD were harmonized across 4 US cohorts (WHI, ARIC, CHS, and MESA) using dbGaP data. Pooled analyses were run using one-step individual participant data multivariate regressions in Caucasians without prevalent CVD. Results: Among 22,412 adults, the GRS was associated with incident CHD (n = 3001 cases) across each cohort (pooled GRS quintile (Q)5 vs Q1, RR (95% CI): 1.59 (1.41-1.79)), independent of all demographic and lifestyle variables. The lifestyle score was more strongly associated with CHD (pooled Q1 vs Q5, RR: 1.94 (1.71-2.20)) than the GRS, and better discriminated those who had a CHD event ( P =0.025 for difference in AUROC curves). When compared to population-level analysis, joint effects analysis revealed significant heterogeneity within quartiles of lifestyle score on risk of incident CHD ( Table ). Interactions between the GRS and diet pattern ( P =0.022) and trans fats ( P =0.046) were observed. Effects of trans fats on CHD risk appear to be concentrated in those with a higher GRS: for trans fats &gt;2.4% of energy, RRs for increasing GRS quartiles are 0.92 (0.62-1.36), 1.13 (0.79-1.62), 1.43 (1.01-2.01), and 1.84 (1.30-2.61), compared to those in the lowest quartile of trans fat (&lt;1.3% of energy) and lowest GRS quartile. Conclusion: A joint effects approach reveals significant inter-individual variability in lifestyle on CHD risk by underlying genetic risk that is obscured by a conventional population-level effects approach.

Explaining Disparities in Ovarian Cancer Incidence Rates between Women of African and European Ancestry: The Role of Genetic Factors

Abstract Non-Hispanic White (NHW) women are at higher risk of ovarian cancer than African-American (AA) women. Approximately 30% of the difference in age-adjusted invasive epithelial ovarian cancer incidence rates (AAIR) between the two groups can be explained by differing oophorectomy rates and the prevalence of non-genetic risk and protective factors. Our purpose was to determine how much of the remaining difference in AAIRs could be explained by varying allele frequencies between NHWs and AAs for 18 genome-wide significant common susceptibility variants for ovarian cancer. Using data on 13,385 cases and 24,875 controls from the Ovarian Cancer Association Consortium, a genetic risk score (GRS) was created from 18 single nucleotide polymorphisms (SNPs) associated with ovarian cancer risk following the Collaborative Oncological Gene-environment Study (COGS) effort. Relative risks for each GRS quintile were estimated using conditional logistic regression, adjusting for genetic ancestry and conditioning on study site, age, and race. The population attributable risk percent (PAR) for GRS above the lowest quintile was calculated using the Bruzzi method. Previously reported oophorectomy and non-genetic risk factor (talc, oral contraceptive use, family history of ovarian cancer, endometriosis, parity and tubal ligation) adjusted incidence rates for ovarian cancer in NHWs and AA's were 7.2 and 5.8 per 100,000 respectively. These incidence rates were further adjusted for the contribution of the GRS from this analysis. The subsequent genetic PAR adjusted rate was 5.1 per 100,000 for the European ancestry group and 4.9 for the African ancestry group, after taking into account the different oophorectomy rates and prevalence of non-genetic risk factors. These incidence rates show the unexplained difference in incidence rates between NHWs and AAs is only 3.9%. Future efforts should focus on incorporating novel non-genetic and genetic factors into this analysis to determine whether essentially all of the difference in incidence between these groups can be explained.

Personalized risk prediction for type 2 diabetes: the potential of genetic risk scores

Purpose:Using effect estimates from genome-wide association studies (GWAS), we identified a genetic risk score (GRS) that has the strongest association with type 2 diabetes (T2D) status in a population-based cohort and investigated its potential for prospective T2D risk assessment.Methods:By varying the number of single-nucleotide polymorphisms (SNPs) and their respective weights, alternative versions of GRS can be computed. They were tested in 1,181 T2D cases and 9,092 controls of the Estonian Biobank cohort. The best-fitting GRS was chosen for the subsequent analysis of incident T2D (386 cases).Results:The best fit was provided by a novel doubly weighted GRS that captures the effect of 1,000 SNPs. The hazard for incident T2D was 3.45 times (95% CI: 2.31–5.17) higher in the highest GRS quintile compared with the lowest quintile, after adjusting for body mass index and other known predictors. Adding GRS to the prediction model for 5-year T2D risk resulted in continuous net reclassification improvement of 0.324 (95% CI: 0.211–0.444). In addition, a significant effect of the GRS on all-cause and cardiovascular mortality was observed.Conclusion:The proposed GRS would improve the accuracy of T2D risk prediction when added to the currently used set of predictors.Genet Med19 3, 322–329.

Genome-wide assessment of Parkinson's disease in a Southern Spanish population

Here, we set out to study the genetic architecture of Parkinson's disease (PD) through a Genome-Wide Association Study in a Southern Spanish population. About 240 PD cases and 192 controls were genotyped on the NeuroX array. We estimated genetic variation associated with PD risk and age at onset (AAO). Risk profile analyses for PD and AAO were performed using a weighted genetic risk score. Total heritability was estimated by genome-wide complex trait analysis. Rare variants were screened with single-variant and burden tests. We also screened for variation in known PD genes. Finally, we explored runs of homozygosity and structural genomic variations. We replicate PD association (uncorrected p-value < 0.05) at the following loci: ACMSD/TMEM163, MAPT, STK39, MIR4697, and SREBF/RAI1. Subjects in the highest genetic risk score quintile showed significantly increased risk of PD versus the lowest quintile (odds ratio = 3.6, p-value < 4e−7), but no significant difference in AAO. We found evidence of runs of homozygosity in 2 PD-associated regions: one intersecting the HLA-DQB1 gene in 6 patients and 1 control; and another intersecting the GBA-SYT11 gene in PD case. The GBA N370S and the LRRK2 G2019S variants were found in 8 and 7 cases, respectively, replicating previous work. A structural variant was found in 1 case in the PARK2 gene locus. This current work represents a comprehensive assessment at a genome-wide level characterizing a novel population in PD genetics.

Common sequence variants associated with coronary artery disease correlate with the extent of coronary atherosclerosis.

Single-nucleotide polymorphisms predisposing to coronary artery disease (CAD) have been shown to predict cardiovascular risk in healthy individuals when combined into a genetic risk score (GRS). We examined whether the cumulative burden of known genetic risk variants associated with risk of CAD influences the development and progression of coronary atherosclerosis. We investigated the combined effects of all known CAD variants in a cross-sectional study of 8622 Icelandic patients with angiographically significant CAD (≥ 50% diameter stenosis). We constructed a GRS based on 50 CAD variants and tested for association with the number of diseased coronary arteries on angiography. In models adjusted for traditional cardiovascular risk factors, the GRS associated significantly with CAD extent (difference per SD increase in GRS, 0.076; P=7.3 × 10(-17)). When compared with the bottom GRS quintile, patients in the top GRS quintile were roughly 1.67× more likely to have multivessel disease (odds ratio, 1.67; 95% confidence interval, 1.45-1.94). The GRS significantly improved prediction of multivessel disease over traditional cardiovascular risk factors (χ(2) likelihood ratio 48.1; P<0.0001) and modestly improved discrimination, as estimated by the C-statistic (without GRS versus with GRS, 64.0% versus 64.8%) and the integrated discrimination improvement (0.52%). Furthermore, the GRS associated with an earlier age at diagnosis of angiographic CAD. These findings were replicated in an independent sample from the Emory Biobank study (n=1853). When combined into a single GRS, known genetic risk variants for CAD contribute significantly to the extent of coronary atherosclerosis in patients with significant angiographic disease.

Abstract 60: Combined effects of low-penetrance variants on breast cancer risk: Results from the Seoul Breast Cancer Study.

Abstract Background: Although several low-penetrance loci associated with breast cancer risk were identified and confirmed, knowledge of the effect of multiple risk alleles is limited especially in Asian women. Therefore we evaluated the association between the polygenic risk scores and breast cancer risk in Korean women using the most recent list of breast cancer susceptibility loci. Methods: We analyzed 51 single-nucleotide polymorphisms (SNPs) located in 34 loci in 1,774 cases and age-frequency matched 1,774 control subjects participating in Seoul Breast Case-Control Study. The fourteen independent SNPs associated with breast cancer risk were selected based on the results of single SNP analysis. The genetic risk score (GRS) were calculated using simple count method and weighted method. Tests of association were conducted using the logistic regression for quintiles of each GRS with or without adjustment. The c-statistic was estimated to evaluate the contribution of GRS to risk prediction model including non-genetic factors (age, family history of breast cancer, education, BMI, menopausal status, age at menarche, age at first full-term pregnancy, and number of children). Results: Fourteen SNPs (rs13393577, rs4973768, rs7716600, rs1092913, rs889312, rs9485372, rs2046210, rs1562430, rs704010, rs10736303, rs7107217, rs10771399, rs3803662, and rs4784227), each of which reflected a genetically independent locus, were found to be associated with breast cancer risk. A highly significant trend was observed between the GRS and the risk of breast cancer. The adjusted odds ratios for women in the highest quintile of count GRS or weighted GRS vs. those in the lowest were 2.64 (95% confidence interval (95% CI), 2.09- 3.35; Ptrend=9.3E-19) and 2.76 (95% CI, 2.18- 3.50; Ptrend=5.9E-21), respectively. The c-statistic for model including the GRS in additional to the conventional risk factors was 0.6389 (95% CI, 0.620-0.658) vs. 0.6041 (95% CI, 0.585-0.623) with the conventional risk factors only. Conclusions: Supporting the polygenic inheritance model of breast cancer, our study showed that GRS based on low-penetrance SNPs adds very modest improvement to risk prediction models. Citation Format: Hyuna Sung, Ji-Yeob Choi, Sue K. Park, Wonshik Han, Keun-Young Yoo, Sei-Hyun Ahn, Dong-Young Noh, Daehee Kang. Combined effects of low-penetrance variants on breast cancer risk: Results from the Seoul Breast Cancer Study. [abstract]. In: Proceedings of the AACR Special Conference on Post-GWAS Horizons in Molecular Epidemiology: Digging Deeper into the Environment; 2012 Nov 11-14; Hollywood, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2012;21(11 Suppl):Abstract nr 60.

Abstract 175: Performance of a Multilocus Genetic Risk Score Derived from Top Signals in the Cardiogram+C4d Consortium in Predicting Incident Coronary Disease in the Atherosclerosis Risk in Communities Study

Genome wide association studies (GWAS) to date have identified 30 CAD susceptibility loci but the ability to use this information to improve risk prediction remains limited. A meta-analysis of the GWAS and Cardio Metabochip data produced by the CARDIoGRAM+C4D consortium representing 63,253 cases and 126,820 controls has identified 1885 SNPs passing a False Discovery Rate (FDR) threshold of 0.5%. We hypothesized that an expanded multi locus genetic risk score (GRS) incorporating genotype information at all loci below an FDR of 0.5% would perform better than a GRS restricted to 42 loci reaching genome wide significance and tested this hypothesis in subjects of European ancestry participating in the Atherosclerosis Risk in the Community (ARIC) study. Models testing the GRS were either minimally (age and sex) or fully adjusted for traditional risk factors (TRFs). The Figure shows the hazard ratio (HZ) and 95% CI for incident events comparing each quintile of GRS to the middle quintile. The GRS including genotype information at all loci with an FDR of 0.5% noticeably improves risk prediction over the GRS restricted to genome wide significant loci in both the minimally and fully adjusted models based on several metrics including i) HR per GRS quintile, ii) the HR per SD of the GRS, and iii) the logistic regression pseudo R2, and iv) the c statistic. The HR per GRS quintile and per SD of GRS were all lower in the fully adjusted models compared to the respective minimally adjusted models but the reduction of the HR was more striking for the models that tested the more expansive GRS. These findings suggest that a larger proportion of novel GWAS CAD loci are mediating their effects through TRFs. While these findings demonstrate some progress in risk prediction using GWAS loci, both the limited and the expanded GRS continues to explain a relatively small proportion of the overall variance compared to TRF. Thus, the clinical utility of a CAD GRS remains to be determined.

A multilocus genetic risk score for coronary heart disease: case-control and prospective cohort analyses

SummaryBackgroundComparison of patients with coronary heart disease and controls in genome-wide association studies has revealed several single nucleotide polymorphisms (SNPs) associated with coronary heart disease. We aimed to establish the external validity of these findings and to obtain more precise risk estimates using a prospective cohort design.MethodsWe tested 13 recently discovered SNPs for association with coronary heart disease in a case-control design including participants differing from those in the discovery samples (3829 participants with prevalent coronary heart disease and 48 897 controls free of the disease) and a prospective cohort design including 30 725 participants free of cardiovascular disease from Finland and Sweden. We modelled the 13 SNPs as a multilocus genetic risk score and used Cox proportional hazards models to estimate the association of genetic risk score with incident coronary heart disease. For case-control analyses we analysed associations between individual SNPs and quintiles of genetic risk score using logistic regression.FindingsIn prospective cohort analyses, 1264 participants had a first coronary heart disease event during a median 10·7 years' follow-up (IQR 6·7–13·6). Genetic risk score was associated with a first coronary heart disease event. When compared with the bottom quintile of genetic risk score, participants in the top quintile were at 1·66-times increased risk of coronary heart disease in a model adjusting for traditional risk factors (95% CI 1·35–2·04, p value for linear trend=7·3×10−10). Adjustment for family history did not change these estimates. Genetic risk score did not improve C index over traditional risk factors and family history (p=0·19), nor did it have a significant effect on net reclassification improvement (2·2%, p=0·18); however, it did have a small effect on integrated discrimination index (0·004, p=0·0006). Results of the case-control analyses were similar to those of the prospective cohort analyses.InterpretationUsing a genetic risk score based on 13 SNPs associated with coronary heart disease, we can identify the 20% of individuals of European ancestry who are at roughly 70% increased risk of a first coronary heart disease event. The potential clinical use of this panel of SNPs remains to be defined.FundingThe Wellcome Trust; Academy of Finland Center of Excellence for Complex Disease Genetics; US National Institutes of Health; the Donovan Family Foundation.

Genetic and Clinical Predictors for Breast Cancer Risk Assessment and Stratification Among Chinese Women

Most of the genetic variants identified from genome-wide association studies of breast cancer have not been validated in Asian women. No risk assessment model that incorporates both genetic and clinical predictors is currently available to predict breast cancer risk in this population. We analyzed 12 single-nucleotide polymorphisms (SNPs) identified in recent genome-wide association studies mostly of women of European ancestry as being associated with the risk of breast cancer in 3039 case patients and 3082 control subjects who participated in the Shanghai Breast Cancer Study. All participants were interviewed in person to obtain information regarding known and suspected risk factors for breast cancer. The c statistic, a measure of discrimination ability with a value ranging from 0.5 (random classification) to 1.0 (perfect classification), was estimated to evaluate the contribution of genetic and established clinical predictors of breast cancer to a newly established risk assessment model for Chinese women. Clinical predictors included in the model were age at menarche, age at first live birth, waist-to-hip ratio, family history of breast cancer, and a previous diagnosis of benign breast disease. The utility of the models in risk stratification was evaluated by estimating the proportion of breast cancer patients in the general population that could be accounted for above a given risk threshold as predicted by the models. All statistical tests were two-sided. Eight SNPs (rs2046210, rs1219648, rs3817198, rs8051542, rs3803662, rs889312, rs10941679, and rs13281615), each of which reflected a genetically independent locus, were found to be associated with the risk of breast cancer. A dose-response association was observed between the risk of breast cancer and the genetic risk score, which is an aggregate measure of the effect of these eight SNPs (odds ratio for women in the highest quintile of genetic risk score vs those in the lowest = 1.85, 95% confidence interval = 1.58 to 2.18, P(trend) = 2.5 x 10(-15)). The genetic risk score, the waist-to-hip ratio, and a previous diagnosis of benign breast disease were the top three predictors of the risk of breast cancer, each contributing statistically significantly (P < .001) to the full risk assessment model. The model, with a c statistic of 0.6295 after adjustment for overfitting, showed promise for stratifying women into different risk groups; women in the top 30% risk group accounted for nearly 50% of the breast cancers diagnosed in the general population. A risk assessment model that includes both genetic markers and clinical predictors may be useful to classify Asian women into relevant risk groups for cost-efficient screening and other prevention programs.