Polygenic scoring accuracy varies across the genetic ancestry continuum

Nature, Journal Year: 2023, Volume and Issue: 618(7966), P. 774 - 781

Published: May 17, 2023

Abstract Polygenic scores (PGSs) have limited portability across different groupings of individuals (for example, by genetic ancestries and/or social determinants health), preventing their equitable use 1–3 . PGS has typically been assessed using a single aggregate population-level statistic R 2 ) 4 , ignoring inter-individual variation within the population. Here, large and diverse Los Angeles biobank 5 (ATLAS, n = 36,778) along with UK Biobank 6 (UKBB, 487,409), we show that accuracy decreases individual-to-individual continuum 7 in all considered populations, even traditionally labelled ‘homogeneous’ ancestries. The decreasing trend is well captured continuous measure distance (GD) from training data: Pearson correlation −0.95 between GD averaged 84 traits. When applying models trained on as white British UKBB to European ATLAS, furthest decile 14% lower relative closest decile; notably, Hispanic Latino American similar performance significantly correlated estimates themselves for 82 traits, further emphasizing importance incorporating interpretation. Our results highlight need move away discrete ancestry clusters towards when considering PGSs.

Language: Английский

---

Quantifying portable genetic effects and improving cross-ancestry genetic prediction with GWAS summary statistics

Nature Communications, Journal Year: 2023, Volume and Issue: 14(1)

Published: Feb. 14, 2023

Abstract Polygenic risk scores (PRS) calculated from genome-wide association studies (GWAS) of Europeans are known to have substantially reduced predictive accuracy in non-European populations, limiting their clinical utility and raising concerns about health disparities across ancestral populations. Here, we introduce a statistical framework named X-Wing improve performance ancestrally diverse quantifies local genetic correlations for complex traits between employs an annotation-dependent estimation procedure amplify correlated effects combines multiple population-specific PRS into unified score with GWAS summary statistics alone as input. Through extensive benchmarking, demonstrate that pinpoints portable improves showing 14.1%–119.1% relative gain R 2 compared state-of-the-art methods based on statistics. Overall, addresses critical limitations existing approaches may broad applications cross-population polygenic prediction.

Language: Английский

---

Extremely sparse models of linkage disequilibrium in ancestrally diverse association studies

Nature Genetics, Journal Year: 2023, Volume and Issue: 55(9), P. 1494 - 1502

Published: Aug. 28, 2023

Language: Английский

---

Simple scaling laws control the genetic architectures of human complex traits

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2022, Volume and Issue: unknown

Published: Oct. 7, 2022

Abstract Genome-wide association studies have revealed that the genetic architectures of complex traits vary widely, including in terms numbers, effect sizes, and allele frequencies significant hits. However, at present we lack a principled way understanding similarities differences among traits. Here, describe probabilistic model combines mutation, drift, stabilizing selection individual sites with genome-scale phenotypic variation. In this model, architecture trait arises from distribution coefficients mutations two scaling parameters. We fit for 95 diverse, highly polygenic quantitative UK Biobank. Notably, infer similar distributions across all these This shared implies arise mainly parameters: mutational target size heritability per site, which by orders magnitude When scale factors are accounted for, nearly identical.

Language: Английский

---

Fast and accurate Bayesian polygenic risk modeling with variational inference

The American Journal of Human Genetics, Journal Year: 2023, Volume and Issue: 110(5), P. 741 - 761

Published: April 7, 2023

Language: Английский

---

Tree-based QTL mapping with expected local genetic relatedness matrices

The American Journal of Human Genetics, Journal Year: 2023, Volume and Issue: 110(12), P. 2077 - 2091

Published: Dec. 1, 2023

Language: Английский

---

SDPRX: A statistical method for cross-population prediction of complex traits

The American Journal of Human Genetics, Journal Year: 2022, Volume and Issue: 110(1), P. 13 - 22

Published: Dec. 1, 2022

Language: Английский

---

Optimizing and benchmarking polygenic risk scores with GWAS summary statistics

Genome biology, Journal Year: 2024, Volume and Issue: 25(1)

Published: Oct. 8, 2024

Polygenic risk score (PRS) is a major research topic in human genetics. However, significant gap exists between PRS methodology and applications practice due to often unavailable individual-level data for various tasks including model fine-tuning, benchmarking, ensemble learning.

Language: Английский

---

Unifying approaches from statistical genetics and phylogenetics for mapping phenotypes in structured populations

PLoS Biology, Journal Year: 2024, Volume and Issue: 22(10), P. e3002847 - e3002847

Published: Oct. 9, 2024

In both statistical genetics and phylogenetics, a major goal is to identify correlations between genetic loci or other aspects of the phenotype environment focal trait. these 2 fields, there are sophisticated but disparate traditions aimed at tasks. The disconnect their respective approaches becoming untenable as questions in medicine, conservation biology, evolutionary biology increasingly rely on integrating data from within among species, once-clear conceptual divisions blurred. To help bridge this divide, we lay out general model describing covariance contributions quantitative phenotypes different individuals. Taking approach shows that standard models (e.g., genome-wide association studies; GWAS) phylogenetic comparative regression) can be interpreted special cases more quantitative-genetic model. fact share same core architecture means build unified understanding strengths limitations methods for controlling structure when testing associations. We develop intuition why spurious may occur analytically conduct population-genetic simulations traits. structural similarity problems phylogenetics enables us take methodological advances one field apply them other. demonstrate by showing how GWAS technique-including relatedness matrix (GRM) well its leading eigenvectors, corresponding principal components genotype matrix, regression model-can mitigate analyses. As case study, re-examine an analysis coevolution expression levels genes across fungal phylogeny show including eigenvectors covariates decreases false positive rate while simultaneously increasing true rate. More generally, work provides foundation integrative processes shape it.

Language: Английский

---

Improving polygenic prediction from summary data by learning patterns of effect sharing across multiple phenotypes

PLoS Genetics, Journal Year: 2025, Volume and Issue: 21(1), P. e1011519 - e1011519

Published: Jan. 7, 2025

Polygenic prediction of complex trait phenotypes has become important in human genetics, especially the context precision medicine. Recently, mr.mash , a flexible and computationally efficient method that models multiple jointly leverages sharing effects across such to improve accuracy, was introduced. However, drawback is it requires individual-level data, which are often not publicly available. In this work, we introduce mr.mash-rss an extension model only summary statistics from Genome-Wide Association Studies (GWAS) linkage disequilibrium (LD) estimates reference panel. By using achieve twin goal increasing applicability data sets available making scalable biobank-size data. Through simulations, show competitive with, outperforms, current state-of-the-art methods for single- multi-phenotype polygenic variety scenarios differ pattern effect phenotypes, number causal variants, genomic heritability. We also present real analysis 16 blood cell UK Biobank, showing achieves higher accuracy than competing majority traits, when set smaller sample size.

Language: Английский

---