Large uncertainty in individual polygenic risk score estimation impacts PRS-based risk stratification.
Ontology highlight
ABSTRACT:
SUBMITTER: Ding Y
PROVIDER: S-EPMC8758557 | biostudies-literature | 2022 Jan
REPOSITORIES: biostudies-literature
Publications
Large uncertainty in individual polygenic risk score estimation impacts PRS-based risk stratification.
Ding Yi Y Hou Kangcheng K Burch Kathryn S KS Lapinska Sandra S Privé Florian F Vilhjálmsson Bjarni B Sankararaman Sriram S Pasaniuc Bogdan B
Nature genetics 20211220 1
Although the cohort-level accuracy of polygenic risk scores (PRSs)-estimates of genetic value at the individual level-has been widely assessed, uncertainty in PRSs remains underexplored. In the present study, we show that Bayesian PRS methods can estimate the variance of an individual's PRS and can yield well-calibrated credible intervals via posterior sampling. For 13 real traits in the UK Biobank (n = 291,273 unrelated 'white British'), we observe large variances in individual PRS estimates wh ...[more]
Similar Datasets
Project description:BackgroundAlthough polygenic risk score (PRS) has emerged as a promising tool for predicting cancer risk from genome-wide association studies (GWAS), the individual-level accuracy of lung cancer PRS and the extent to which its impact on subsequent clinical applications remains largely unexplored.MethodsLung cancer PRSs and confidence/credible interval (CI) were constructed using two statistical approaches for each individual: (1) the weighted sum of 16 GWAS-derived significant SNP loci and the CI through the bootstrapping method (PRS-16-CV) and (2) LDpred2 and the CI through posteriors sampling (PRS-Bayes), among 17,166 lung cancer cases and 12,894 controls with European ancestry from the International Lung Cancer Consortium. Individuals were classified into different genetic risk subgroups based on the relationship between their own PRS mean/PRS CI and the population level threshold.ResultsConsiderable variances in PRS point estimates at the individual level were observed for both methods, with an average standard deviation (s.d.) of 0.12 for PRS-16-CV and a much larger s.d. of 0.88 for PRS-Bayes. Using PRS-16-CV, only 25.0% of individuals with PRS point estimates in the lowest decile of PRS and 16.8% in the highest decile have their entire 95% CI fully contained in the lowest and highest decile, respectively, while PRS-Bayes was unable to find any eligible individuals. Only 19% of the individuals were concordantly identified as having high genetic risk (> 90th percentile) using the two PRS estimators. An increased relative risk of lung cancer comparing the highest PRS percentile to the lowest was observed when taking the CI into account (OR = 2.73, 95% CI: 2.12-3.50, P-value = 4.13 × 10-15) compared to using PRS-16-CV mean (OR = 2.23, 95% CI: 1.99-2.49, P-value = 5.70 × 10-46). Improved risk prediction performance with higher AUC was consistently observed in individuals identified by PRS-16-CV CI, and the best performance was achieved by incorporating age, gender, and detailed smoking pack-years (AUC: 0.73, 95% CI = 0.72-0.74).ConclusionsLung cancer PRS estimates using different methods have modest correlations at the individual level, highlighting the importance of considering individual-level uncertainty when evaluating the practical utility of PRS.
| S-EPMC10840262 | biostudies-literature
Project description:Polygenic risk scores (PRS) have been successfully developed for the prediction of human diseases and complex traits in the past years. For drug response prediction in randomized clinical trials, a common practice is to apply PRS built from a disease genome-wide association study (GWAS) directly to a corresponding pharmacogenomics (PGx) setting. Here, we show that such an approach relies on stringent assumptions about the prognostic and predictive effects of the selected genetic variants. We propose a shift from disease PRS to PGx PRS approaches by simultaneously modeling both the prognostic and predictive effects and further make this shift possible by developing a series of PRS-PGx methods, including a novel Bayesian regression approach (PRS-PGx-Bayes). Simulation studies show that PRS-PGx methods generally outperform the disease PRS methods and PRS-PGx-Bayes is superior to all other PRS-PGx methods. We further apply the PRS-PGx methods to PGx GWAS data from a large cardiovascular randomized clinical trial (IMPROVE-IT) to predict treatment related LDL cholesterol reduction. The results demonstrate substantial improvement of PRS-PGx-Bayes in both prediction accuracy and the capability of capturing the treatment-specific predictive effects while compared with the disease PRS approaches.
| S-EPMC9458667 | biostudies-literature
Project description:AimsTo construct a polygenic risk score (PRS) for coronary artery disease (CAD) and comprehensively evaluate its potential in clinical utility for primary prevention in Chinese populations.Methods and resultsUsing meta-analytic approach and large genome-wide association results for CAD and CAD-related traits in East Asians, a PRS comprising 540 genetic variants was developed in a training set of 2800 patients with CAD and 2055 controls, and was further assessed for risk stratification for CAD integrating with the guideline-recommended clinical risk score in large prospective cohorts comprising 41 271 individuals. During a mean follow-up of 13.0 years, 1303 incident CAD cases were identified. Individuals with high PRS (the highest 20%) had about three-fold higher risk of CAD than the lowest 20% (hazard ratio 2.91, 95% confidence interval 2.43-3.49), with the lifetime risk of 15.9 and 5.8%, respectively. The addition of PRS to the clinical risk score yielded a modest yet significant improvement in C-statistic (1%) and net reclassification improvement (3.5%). We observed significant gradients in both 10-year and lifetime risk of CAD according to the PRS within each clinical risk strata. Particularly, when integrating high PRS, intermediate clinical risk individuals with uncertain clinical decision for intervention would reach the risk levels (10-year of 4.6 vs. 4.8%, lifetime of 17.9 vs. 16.6%) of high clinical risk individuals with intermediate (20-80%) PRS.ConclusionThe PRS could stratify individuals into different trajectories of CAD risk, and further refine risk stratification for CAD within each clinical risk strata, demonstrating a great potential to identify high-risk individuals for targeted intervention in clinical utility.
| S-EPMC9076396 | biostudies-literature
Project description:BackgroundPolygenic risk scores (PRSs) are a summarization of an individual's genetic risk for a disease or trait. These scores are being generated in research and commercial settings to study how they may be used to guide healthcare decisions. PRSs should be updated as genetic knowledgebases improve; however, no guidelines exist for their generation or updating.MethodsHere, we characterize the variability introduced in PRS calculation by a common computational process used in their generation-genotype imputation. We evaluated PRS variability when performing genotype imputation using 3 different pre-phasing tools (Beagle, Eagle, SHAPEIT) and 2 different imputation tools (Beagle, Minimac4), relative to a WGS-based gold standard. Fourteen different PRSs spanning different disease architectures and PRS generation approaches were evaluated.ResultsWe find that genotype imputation can introduce variability in calculated PRSs at the individual level without any change to the underlying genetic model. The degree of variability introduced by genotype imputation differs across algorithms, where pre-phasing algorithms with stochastic elements introduce the greatest degree of score variability. In most cases, PRS variability due to imputation is minor (<?5 percentile rank change) and does not influence the interpretation of the score. PRS percentile fluctuations are also reduced in the more informative tails of the PRS distribution. However, in rare instances, PRS instability at the individual level can result in singular PRS calculations that differ substantially from a whole genome sequence-based gold standard score.ConclusionsOur study highlights some challenges in applying population genetics tools to individual-level genetic analysis including return of results. Rare individual-level variability events are masked by a high degree of overall score reproducibility at the population level. In order to avoid PRS result fluctuations during updates, we suggest that deterministic imputation processes or the average of multiple iterations of stochastic imputation processes be used to generate and deliver PRS results.
| S-EPMC7682022 | biostudies-literature
Project description:BackgroundParkinson's disease (PD) is a highly age-related disorder, where common genetic risk variants affect both disease risk and age at onset. A statistical approach that integrates these effects across all common variants may be clinically useful for individual risk stratification. A polygenic hazard score methodology, leveraging a time-to-event framework, has recently been successfully applied in other age-related disorders.ObjectivesWe aimed to develop and validate a polygenic hazard score model in sporadic PD.MethodsUsing a Cox regression framework, we modeled the polygenic hazard score in a training data set of 11,693 PD patients and 9841 controls. The score was then validated in an independent test data set of 5112 PD patients and 5372 controls and a small single-study sample of 360 patients and 160 controls.ResultsA polygenic hazard score predicts the onset of PD with a hazard ratio of 3.78 (95% confidence interval 3.49-4.10) when comparing the highest to the lowest risk decile. Combined with epidemiological data on incidence rate, we apply the score to estimate genetically stratified instantaneous PD risk across age groups.ConclusionsWe demonstrate the feasibility of a polygenic hazard approach in PD, integrating the genetic effects on disease risk and age at onset in a single model. In combination with other predictive biomarkers, the approach may hold promise for risk stratification in future clinical trials of disease-modifying therapies, which aim at postponing the onset of PD. © 2021 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.
| S-EPMC9843635 | biostudies-literature
Project description:BackgroundPolygenic risk scores (PRS) could potentially improve breast cancer screening recommendations. Before a PRS can be considered for implementation, it needs rigorous evaluation, using performance measures that can inform about its future clinical value.ObjectivesTo evaluate the prognostic performance of a regression model with a previously developed, prevalence-based PRS and age as predictors for breast cancer incidence in women from the Estonian biobank (EstBB) cohort; to compare it to the performance of a model including age only.MethodsWe analyzed data on 30,312 women from the EstBB cohort. They entered the cohort between 2002 and 2011, were between 20 and 89 years, without a history of breast cancer, and with full 5-year follow-up by 2015. We examined PRS and other potential risk factors as possible predictors in Cox regression models for breast cancer incidence. With 10-fold cross-validation we estimated 3- and 5-year breast cancer incidence predicted by age alone and by PRS plus age, fitting models on 90% of the data. Calibration, discrimination, and reclassification were calculated on the left-out folds to express prognostic performance.ResultsA total of 101 (3.33‰) and 185 (6.1‰) incident breast cancers were observed within 3 and 5 years, respectively. For women in a defined screening age of 50-62 years, the ratio of observed vs PRS-age modelled 3-year incidence was 0.86 for women in the 75-85% PRS-group, 1.34 for the 85-95% PRS-group, and 1.41 for the top 5% PRS-group. For 5-year incidence, this was respectively 0.94, 1.15, and 1.08. Yet the number of breast cancer events was relatively low in each PRS-subgroup. For all women, the model's AUC was 0.720 (95% CI: 0.675-0.765) for 3-year and 0.704 (95% CI: 0.670-0.737) for 5-year follow-up, respectively, just 0.022 and 0.023 higher than for the model with age alone. Using a 1% risk prediction threshold, the 3-year NRI for the PRS-age model was 0.09, and 0.05 for 5 years.ConclusionThe model including PRS had modest incremental performance over one based on age only. A larger, independent study is needed to assess whether and how the PRS can meaningfully contribute to age, for developing more efficient screening strategies.
| S-EPMC8691010 | biostudies-literature
Project description:Polygenic risk scores (PRS) have the potential to identify individuals at risk of diseases, optimizing treatment, and predicting survival outcomes. Here, we construct and validate a genome-wide association study (GWAS) derived PRS for nasopharyngeal carcinoma (NPC), using a multi-center study of six populations (6 059 NPC cases and 7 582 controls), and evaluate its utility in a nested case-control study. We show that the PRS enables effective identification of NPC high-risk individuals (AUC = 0.65) and improves the risk prediction with the PRS incremental deciles in each population (Ptrend ranging from 2.79 × 10-7 to 4.79 × 10-44). By incorporating the PRS into EBV-serology-based NPC screening, the test's positive predictive value (PPV) is increased from an average of 4.84% to 8.38% and 11.91% in the top 10% and 5% PRS, respectively. In summary, the GWAS-derived PRS, together with the EBV test, significantly improves NPC risk stratification and informs personalized screening.
| S-EPMC9005522 | biostudies-literature
Project description:BackgroundStatistical associations of numerous single nucleotide polymorphisms with breast cancer (BC) have been identified in genome-wide association studies (GWAS). Recent evidence suggests that a Polygenic Risk Score (PRS) can be a useful risk stratification instrument for a BC screening strategy, and a PRS test has been developed for clinical use. The performance of the PRS is yet unknown in the Norwegian population.AimTo evaluate the performance of PRS models for BC in a Norwegian dataset.MethodsWe investigated a sample of 1053 BC cases and 7094 controls from different regions of Norway. PRS values were calculated using four PRS models, and their performance was evaluated by the area under the curve (AUC) and the odds ratio (OR). The effect of the PRS on the age of onset of BC was determined by a Cox regression model, and the lifetime absolute risk of developing BC was calculated using the iCare tool.ResultsThe best performing PRS model included 3820 SNPs, which yielded an AUC = 0.625 and an OR = 1.567 per one standard deviation increase. The PRS values of the samples correlate with an increased risk of BC, with a hazard ratio of 1.494 per one standard deviation increase (95% confidence interval of 1.406-1.588). The individuals in the highest decile of the PRS have at least twice the risk of developing BC compared to the individuals with a median PRS. The results in this study with Norwegian samples are coherent with the findings in the study conducted using Estonian and UK Biobank samples.ConclusionThe previously validated PRS models have a similar observed accuracy in the Norwegian data as in the UK and Estonian populations. A PRS provides a meaningful association with the age of onset of BC and lifetime risk. Therefore, as suggested in Estonia, a PRS may also be integrated into the screening strategy for BC in Norway.
| S-EPMC10452897 | biostudies-literature
Project description:BackgroundSubsequent thyroid cancer (STC) is one of the most common malignancies in childhood cancer survivors. We aimed to evaluate the polygenic contributions to STC risk and potential utility in improving risk prediction.MethodsA polygenic risk score (PRS) was calculated from 12 independent SNPs associated with thyroid cancer risk in the general population. Associations between PRS and STC risk were evaluated among survivors from St. Jude Lifetime Cohort (SJLIFE) and were replicated in survivors from Childhood Cancer Survivor Study (CCSS). A risk prediction model integrating the PRS and clinical factors, initially developed in SJLIFE, and its performance were validated in CCSS.ResultsAmong 2,370 SJLIFE survivors with a median follow-up of 28.8 [interquartile range (IQR) = 21.9-36.1] years, 65 (2.7%) developed STC. Among them, the standardized PRS was associated with an increased rate of STC [relative rate (RR) = 1.57; 95% confidence interval (CI) = 1.24-1.98; P < 0.001]. Similar associations were replicated in 6,416 CCSS survivors, among whom 121 (1.9%) developed STC during median follow-up of 28.9 (IQR = 22.6-34.6) years (RR = 1.52; 95% CI = 1.25-1.83; P < 0.001). A risk prediction model integrating the PRS with clinical factors showed better performance than the model considering only clinical factors in SJLIFE (P = 0.004, AUC = 83.2% vs. 82.1%, at age 40), which was further validated in CCSS (P = 0.010, AUC = 72.9% vs. 70.6%).ConclusionsIntegration of the PRS with clinical factors provided a statistically significant improvement in risk prediction of STC, although the magnitude of improvement was modest.ImpactPRS improves risk stratification and prediction of STC, suggesting its potential utility for optimizing screening strategies in survivorship care.
| S-EPMC8568663 | biostudies-literature
Project description:BackgroundPolygenic risk scores (PRS) hold the promise to refine prognostication in hepatocellular cancer (HCC). The few available HCC PRS include germline risk variants identified among individuals of mostly European ancestry, but data are lacking on the transportability of these PRS in multiethnic U.S patients with cirrhosis from multiple etiologies.MethodsWe used data from 1644 patients with cirrhosis enrolled in two prospective cohort studies in the U.S. Patients were followed until HCC diagnosis, death, liver transplantation, or last study visit through June 30, 2021. The high-risk variants in PNPLA3-MBOAT7-TM6SF2-GCKR were combined in a PRS and we evaluated its association with HCC. Discriminatory accuracy was assessed using the C-statistic.ResultsDuring 4,759 person-years of follow-up, 93 patients developed HCC. Mean age was 59.8 years, 68.6% were male, 27.2% Hispanic, 25.1% non-Hispanic Black, 25.7% had NAFLD, 42.1% had heavy alcohol use, and 19.5% had active HCV. HCC risk increased by 134% per unit increase in PRS (HR = 2.30; 95% CI, 1.35-3.92). Compared to cirrhosis patients in the lowest tertile of the PRS, those in the highest tertile had 2-fold higher risk of HCC (HR = 2.05; 95% CI, 1.22-3.44). The PRS alone had modest discriminatory ability (C-statistic = 0.58; 95% CI, 0.52-0.63); however, adding PRS to a predictive model with traditional HCC risk factors had a C-statistic of 0.70 (95% CI, 0.64-0.76), increasing from 0.68 without the PRS (p = 0.0012).ConclusionsOur findings suggest that PRS may enhance risk prediction for HCC in contemporary U.S. cirrhosis patients.
| S-EPMC9974109 | biostudies-literature