Evaluating Performance and Agreement of Coronary Heart Disease Polygenic Risk Scores

JAMA, Год журнала: 2024, Номер unknown

Опубликована: Ноя. 16, 2024

Importance Polygenic risk scores (PRSs) for coronary heart disease (CHD) are a growing clinical and commercial reality. Whether existing provide similar individual-level assessments of susceptibility remains incompletely characterized. Objective To characterize the agreement CHD PRSs that perform similarly at population level. Design, Setting, Participants Cross-sectional study participants from diverse backgrounds enrolled in All Us Research Program (AOU), Penn Medicine BioBank (PMBB), University California, Los Angeles (UCLA) ATLAS Precision Health Biobank with electronic health record genotyping data. Exposures published new developed separately testing samples. Main Outcomes Measures performed population-level prediction were identified by comparing calibration discrimination models prevalent CHD. Individual-level was tested intraclass correlation coefficient (ICC) Light κ. Results A total 48 calculated 171 095 AOU participants. The mean (SD) age 56.4 (16.8) years. 104 947 (61.3%) female. 35 590 (20.8%) most genetically to an African reference population, 29 801 (17.4%) admixed American 100 493 (58.7%) European remaining Central/South Asian, East Middle Eastern populations. There 17 589 (10.3%) 153 506 without (89.7%) When included model CHD, 46 had practically equivalent Brier area under receiver operator curves (region practical equivalence ±0.02). Twenty percent least 1 score both top bottom 5% risk. Continuous individual predictions poor (ICC, 0.373 [95% CI, 0.372-0.375]). κ, used evaluate consistency assignment, did not exceed 0.56. Analysis among 41 193 PMBB 53 092 yielded different sets scores, which also lacked agreement. Conclusions Relevance level demonstrated highly variable estimates Recognizing may generate incongruent estimates, effective implementation will require refined statistical methods quantify uncertainty strategies communicate this patients clinicians.

Язык: Английский

---

Methylation Clocks Do Not Predict Age or Alzheimer’s Disease Risk Across Genetically Admixed Individuals

Опубликована: Янв. 31, 2025

Epigenetic clocks that quantify rates of aging from DNA methylation patterns across the genome have emerged as a potential biomarker for risk age-related diseases, like Alzheimer’s disease (AD), and environmental social stressors. However, not been validated in genetically diverse cohorts. Here we evaluate set 621 AD patients matched controls African American, Hispanic, white co-horts. The are less accurate at predicting age admixed individuals, especially those with substantial ancestry, than cohort. also do consistently identify acceleration cases compared to controls. Methylation QTL (meQTL) commonly influence CpGs clocks, these meQTL significantly higher frequencies genetic ancestries. Our results demonstrate often fail predict beyond their training populations suggest avenues improving portability.

Язык: Английский

---

Integrated clinical risk prediction of type 2 diabetes with a multifactorial polygenic risk score

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Авг. 22, 2024

Combining information from multiple GWASs for a disease and its risk factors has proven powerful approach development of polygenic scores (PRSs). This may be particularly useful type 2 diabetes (T2D), highly heterogeneous where the additional predictive value PRS is unclear. Here, we use meta-scoring to develop metaPRS T2D that incorporated genome-wide associations both European non-European genetic ancestries factors. We evaluated performance this benchmarked it against existing in 620,059 participants 50,572 cases amongst six diverse UK Biobank, INTERVAL, All Us Research Program, Singapore Multi-Ethnic Cohort. show our was most predicting population-based cohorts had comparable top ancestry-specific PRS, highlighting transferability. In stronger power 10-year than all individual apart BMI biomarkers dysglycemia. The modestly improved stratification QDiabetes prediction, when prioritising individuals blood tests Overall, present transferrable demonstrate potential incrementally improve prediction into guideline-recommended screening with clinical score.

Язык: Английский

---

Methylation Clocks Do Not Predict Age or Alzheimer’s Disease Risk Across Genetically Admixed Individuals

Опубликована: Янв. 31, 2025

Epigenetic clocks that quantify rates of aging from DNA methylation patterns across the genome have emerged as a potential biomarker for risk age-related diseases, like Alzheimer’s disease (AD), and environmental social stressors. However, not been validated in genetically diverse cohorts. Here we evaluate set 621 AD patients matched controls African American, Hispanic, white co-horts. The are less accurate at predicting age admixed individuals, especially those with substantial ancestry, than cohort. also do consistently identify acceleration cases compared to controls. Methylation QTL (meQTL) commonly influence CpGs clocks, these meQTL significantly higher frequencies genetic ancestries. Our results demonstrate often fail predict beyond their training populations suggest avenues improving portability.

Язык: Английский

---

The Estonian Biobank’s journey from biobanking to personalized medicine

Nature Communications, Год журнала: 2025, Номер 16(1)

Опубликована: Апрель 5, 2025

Abstract Large biobanks have set a new standard for research and innovation in human genomics implementation of personalized medicine. The Estonian Biobank was founded quarter century ago, its biological specimens, clinical, health, omics, lifestyle data been included over 800 publications to date. What makes the biobank unique internationally is translational focus, with active efforts conduct clinical studies based on genetic findings, explore effects return results participants. In this review, we provide an overview Biobank, highlight strengths studying variation quantitative phenotypes health-related traits, development methods frameworks bringing into clinic, role as driving force implementing medicine national level beyond.

Язык: Английский

---

All of Us diversity and scale improve polygenic prediction contextually with greatest improvements for under-represented populations

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Авг. 6, 2024

Abstract Recent studies have demonstrated that polygenic risk scores (PRS) trained on multi-ancestry data can improve prediction accuracy in groups historically underrepresented genomic studies, but the availability of linked health and genetic from large-scale diverse cohorts representative a wide spectrum human diversity remains limited. To address this need, All Us research program (AoU) generated whole-genome sequences 245,388 individuals who collectively reflect USA. Leveraging resource another widely-used population-scale biobank, UK Biobank (UKB) with half million participants, we developed PRS multi-biobank up to ∼750,000 participants for 32 common, complex traits diseases across range architectures. We then compared effects ancestry, methodology, architecture held out subset ancestrally AoU participants. Due more heterogeneous study design AoU, found lower heritability average UKB (0.075 vs 0.165), which limited maximal achievable AoU. Overall, increased significantly improved performance some especially individuals, multiple phenotypes. Notably, maximizing sample size by combining discovery is not optimal approach predicting phenotypes African ancestry populations; rather, using only these resulted greatest accuracy. This was true less large ancestry-enriched effects, such as neutrophil count ( R 2 : 0.055 vs. 0.035 cross-biobank meta-analysis, respectively, because e.g. DARC ). Lastly, calculated individual-level accuracies rather than grouping continental critical step towards interpretability precision medicine. Individualized decays linearly function divergence, slope smaller GWAS European GWAS. Our results highlight potential biobanks balanced representations facilitate accurate least represented studies.

Язык: Английский

---

One score to rule them all: regularized ensemble polygenic risk prediction with GWAS summary statistics

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Дек. 3, 2024

Ensemble learning has been increasingly popular for boosting the predictive power of polygenic risk scores (PRS), with almost every recent multi-ancestry PRS approach employing ensemble as a final step. Existing approaches rely on individual-level data model training, which severely limits their real-world applications, especially in non-European populations without sufficient genomic samples. Here, we introduce statistical framework to construct regularized PRS, allows us combine large number candidate models using only summary statistics from genome-wide association studies. We demonstrate its robust and substantial improvement over many existing both within- cross-ancestry applications. believe this is truly "one score rule them all" due capability continuously newly developed improve prediction performance, makes it universal that should always be employed future

Язык: Английский

---

Population Performance and Individual Agreement of Coronary Artery Disease Polygenic Risk Scores

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Июль 26, 2024

Abstract Importance Polygenic risk scores (PRSs) for coronary artery disease (CAD) are a growing clinical and commercial reality. Whether existing provide similar individual-level assessments of liability is critical consideration implementation that remains uncharacterized. Objective Characterize the reliability CAD PRSs perform equivalently at population level predicting risk. Design Cross-sectional Study. Setting All Us Research Program (AOU), Penn Medicine Biobank (PMBB), UCLA ATLAS Precision Health Biobank. Participants Volunteers diverse genetic backgrounds enrolled in AOU, PMBB, with available electronic health record genotyping data. Exposures from previously published new developed separately testing cohorts. Main Outcomes Measures Sets prediction were identified by comparing calibration discrimination (Brier score AUROC) generalized linear models prevalent using Bayesian analysis variance. Among performing scores, agreement between estimates was tested intraclass correlation (ICC) Light’s Kappa, measures inter-rater reliability. Results 50 calculated 171,095 AOU participants. When included model CAD, 48 had practically equivalent Brier AUROCs (region practical equivalence = 0.02). Across these 84% participants least one both top bottom quintile. Continuous individual predictions poor, an ICC 0.351 (95% CI; 0.349, 0.352). Agreement two statistically moderate, 0.649 0.646, 0.652). used to evaluate consistency assignment high-risk thresholds, did not exceed 0.56 (interpreted as ‘fair’) across scores. Repeating among 41,193 PMBB 50,748 yielded different sets which also lacked strong agreement. Conclusions Relevance three biobanks, performed produced unreliable estimates. Approaches must consider potential discordant otherwise indistinguishable

Язык: Английский

---

Improving on polygenic scores across complex traits using select and shrink with summary statistics (S4) and LDpred2

BMC Genomics, Год журнала: 2024, Номер 25(1)

Опубликована: Сен. 18, 2024

Язык: Английский

---

Lipid trajectories improve risk models for Alzheimer's disease and mild cognitive impairment

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Сен. 28, 2024

Abstract To assess the relationship between lipids and cognitive dysfunction, we retrospectively analyzed blood-lipid levels in clinically well-characterized individuals with stable mild impairment (MCI) or Alzheimer’s disease (AD) over decade prior to first symptoms. In this case/control cohort study, AD MCI cases were diagnosed using DSM-IV criteria; had not progressed dementia for ≥5 years; controls propensity matched at age of symptom onset (MCI: 116 cases, 435 controls; AD: 215 483 controls). Participants grouped based on longitudinal trajectories quintile variability independent mean (VIM) total cholesterol, HDL-C, LDL-C, non-HDL-C ln(triglycerides). Models risk dysfunction evaluated trajectory VIM groups, APOE genotype, polygenic scores (PRS) lipid levels, age, comorbidities, correlates concentrations. Lower HDL-C (OR = 3.8, 95% CI 1.3–11.3) lowest 2.2, 1.3–3.0) associated higher risk. 3.0, 1.6–5.7) cholesterol 2.4, 1.5–3.9) The inclusion lipid-trajectory groups improved risk-model predictive performance genotype PRS levels. These results provide an important real-world perspective influence metabolism development AD.

Язык: Английский

---