Algorithmic fairness in artificial intelligence for medicine and healthcare

Nature Biomedical Engineering, Journal Year: 2023, Volume and Issue: 7(6), P. 719 - 742

Published: June 28, 2023

Language: Английский

---

Leveraging fine-mapping and multipopulation training data to improve cross-population polygenic risk scores

Nature Genetics, Journal Year: 2022, Volume and Issue: 54(4), P. 450 - 458

Published: April 1, 2022

Language: Английский

---

Principles and methods for transferring polygenic risk scores across global populations

Nature Reviews Genetics, Journal Year: 2023, Volume and Issue: 25(1), P. 8 - 25

Published: Aug. 24, 2023

Language: Английский

---

Public human microbiome data are dominated by highly developed countries

PLoS Biology, Journal Year: 2022, Volume and Issue: 20(2), P. e3001536 - e3001536

Published: Feb. 15, 2022

The importance of sampling from globally representative populations has been well established in human genomics. In microbiome research, however, we lack a full understanding the global distribution research studies. This information is crucial to better understand patterns microbiome-associated diseases and extend health benefits this all populations. Here, analyze country origin 444,829 samples that are available world's 3 largest genomic data repositories, including Sequence Read Archive (SRA). 2,592 studies 19 body sites, 220,017 gut microbiome. We show more than 71% with known come Europe, United States, Canada, 46.8% US alone, despite representing only 4.3% population. also find central southern Asia most underrepresented region: Countries such as India, Pakistan, Bangladesh account for quarter world population but make up 1.8% samples. These results demonstrate critical need ensure representation participants

Language: Английский

---

Challenges and Opportunities for Developing More Generalizable Polygenic Risk Scores

Annual Review of Biomedical Data Science, Journal Year: 2022, Volume and Issue: 5(1), P. 293 - 320

Published: May 16, 2022

Polygenic risk scores (PRS) estimate an individual's genetic likelihood of complex traits and diseases by aggregating information across multiple variants identified from genome-wide association studies. PRS can predict a broad spectrum have therefore been widely used in research settings. Some work has investigated their potential applications as biomarkers preventative medicine, but significant is still needed to definitively establish communicate absolute patients for modifiable factors demographic groups. However, the biggest limitation currently that they show poor generalizability diverse ancestries cohorts. Major efforts are underway through methodological development data generation initiatives improve generalizability. This review aims comprehensively discuss current progress on PRS, affect generalizability, promising areas improving accuracy, portability, implementation.

Language: Английский

---

Recommendations on the use and reporting of race, ethnicity, and ancestry in genetic research: Experiences from the NHLBI TOPMed program

Cell Genomics, Journal Year: 2022, Volume and Issue: 2(8), P. 100155 - 100155

Published: July 26, 2022

How race, ethnicity, and ancestry are used in genomic research has wide-ranging implications for how is translated into clinical care incorporated public understanding. Correlation between race genetic contributes to unresolved complexity the scientific community, as illustrated by heterogeneous definitions applications of these variables. Here, we offer commentary recommendations on use across arc research, including data harmonization, analysis, reporting. While informed our experiences researchers affiliated with NHLBI Trans-Omics Precision Medicine (TOPMed) program, applicable basic translational diverse populations genome-wide data. Moving forward, considerable collaborative effort will be required ensure that described appropriately generate knowledge yields broad equitable benefit.

Language: Английский

---

A new method for multiancestry polygenic prediction improves performance across diverse populations

Nature Genetics, Journal Year: 2023, Volume and Issue: 55(10), P. 1757 - 1768

Published: Sept. 25, 2023

Language: Английский

---

Improving polygenic risk prediction in admixed populations by explicitly modeling ancestral-differential effects via GAUDI

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Feb. 3, 2024

Abstract Polygenic risk scores (PRS) have shown successes in clinics, but most PRS methods focus only on participants with distinct primary continental ancestry without accommodating recently-admixed individuals mosaic backgrounds for different segments of their genomes. Here, we develop GAUDI, a novel penalized-regression-based method specifically designed admixed individuals. GAUDI explicitly models ancestry-differential effects while borrowing information across shared We demonstrate marked advantages over other through comprehensive simulation and real data analyses traits associated variants exhibiting ancestral-differential effects. Leveraging from the Women’s Health Initiative study, show that improves prediction white blood cell count C-reactive protein African Americans by > 64% compared to alternative methods, even outperforms PRS-CSx large European GWAS some scenarios. believe will be valuable tool mitigate disparities performance

Language: Английский

---

An ensemble penalized regression method for multi-ancestry polygenic risk prediction

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: April 15, 2024

Abstract Great efforts are being made to develop advanced polygenic risk scores (PRS) improve the prediction of complex traits and diseases. However, most existing PRS primarily trained on European ancestry populations, limiting their transferability non-European populations. In this article, we propose a novel method for generating multi-ancestry Polygenic Risk scOres based enSemble PEnalized Regression models (PROSPER). PROSPER integrates genome-wide association studies (GWAS) summary statistics from diverse populations ancestry-specific with improved predictive power minority The uses combination $${{{{{{\mathscr{L}}}}}}}_{1}$$ L 1 (lasso) $${{{{{{\mathscr{L}}}}}}}_{2}$$ 2 (ridge) penalty functions, parsimonious specification parameters across an ensemble step combine generated different parameters. We evaluate performance other methods large-scale simulated real datasets, including those 23andMe Inc., Global Lipids Genetics Consortium, All Us. Results show that can substantially compared alternative wide variety genetic architectures. data analyses, example, increased out-of-sample R 2 continuous by average 70% state-of-the-art Bayesian (PRS-CSx) in African population. Further, is computationally highly scalable analysis large SNP contents many

Language: Английский

---

MUSSEL: Enhanced Bayesian polygenic risk prediction leveraging information across multiple ancestry groups

Cell Genomics, Journal Year: 2024, Volume and Issue: 4(4), P. 100539 - 100539

Published: April 1, 2024

Polygenic risk scores (PRSs) are now showing promising predictive performance on a wide variety of complex traits and diseases, but there exists substantial gap across populations. We propose MUSSEL, method for ancestry-specific polygenic prediction that borrows information in summary statistics from genome-wide association studies (GWASs) multiple ancestry groups via Bayesian hierarchical modeling ensemble learning. In our simulation data analyses four distinct studies, totaling 5.7 million participants with ancestral diversity, MUSSEL shows compared to alternatives. For example, has an average gain R

Language: Английский

---