Polygenic risk score prediction accuracy convergence

Summary: Polygenic risk scores (PRSs) models trained from genome-wide association study (GWAS) results are set to play a pivotal role in biomedical research addressing multifactorial human diseases. The prospect of using these risk scores in clinical care and public health is generating both enthusi...

Full description

Saved in:
Bibliographic Details
Main Authors: Léo Henches, Jihye Kim, Zhiyu Yang, Simone Rubinacci, Gabriel Pires, Clara Albiñana, Christophe Boetto, Hanna Julienne, Arthur Frouin, Antoine Auvergne, Yuka Suzuki, Sarah Djebali, Olivier Delaneau, Andrea Ganna, Bjarni Vilhjálmsson, Florian Privé, Hugues Aschard
Format: Article
Language:English
Published: Elsevier 2025-07-01
Series:HGG Advances
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666247725000600
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850235085690765312
author Léo Henches
Jihye Kim
Zhiyu Yang
Simone Rubinacci
Gabriel Pires
Clara Albiñana
Christophe Boetto
Hanna Julienne
Arthur Frouin
Antoine Auvergne
Yuka Suzuki
Sarah Djebali
Olivier Delaneau
Andrea Ganna
Bjarni Vilhjálmsson
Florian Privé
Hugues Aschard
author_facet Léo Henches
Jihye Kim
Zhiyu Yang
Simone Rubinacci
Gabriel Pires
Clara Albiñana
Christophe Boetto
Hanna Julienne
Arthur Frouin
Antoine Auvergne
Yuka Suzuki
Sarah Djebali
Olivier Delaneau
Andrea Ganna
Bjarni Vilhjálmsson
Florian Privé
Hugues Aschard
author_sort Léo Henches
collection DOAJ
description Summary: Polygenic risk scores (PRSs) models trained from genome-wide association study (GWAS) results are set to play a pivotal role in biomedical research addressing multifactorial human diseases. The prospect of using these risk scores in clinical care and public health is generating both enthusiasm and controversy, with varying opinions among experts about their strengths and limitations. The performance of existing polygenic scores is still limited but is expected to improve with increasing GWAS sample sizes and the development of new, more powerful methods. Theoretically, the variance explained by PRS can be as high as the total additive genetic variance, but it is unclear how much of that variance has already been captured by PRS. Here, we conducted a retrospective analysis to assess progress in PRS prediction accuracy since the publication of the first large-scale GWASs, using data from six common human diseases with sufficient GWAS information. We show that although PRS accuracy has grown rapidly over the years, the pace of improvement from recent GWAS has decreased substantially, suggesting that merely increasing GWAS sample sizes may lead to only modest improvements in risk discrimination. We next investigated the factors influencing the maximum achievable prediction using whole-genome sequencing data from 125,000 UK Biobank participants and state-of-the-art modeling of polygenic outcomes. Our analyses suggest that increasing the variant coverage of PRS, using either more imputed variants or sequencing data, is a key component for future improvements in prediction accuracy.
format Article
id doaj-art-293258ce375a4fb380a212bfb0fdd4eb
institution OA Journals
issn 2666-2477
language English
publishDate 2025-07-01
publisher Elsevier
record_format Article
series HGG Advances
spelling doaj-art-293258ce375a4fb380a212bfb0fdd4eb2025-08-20T02:02:25ZengElsevierHGG Advances2666-24772025-07-016310045710.1016/j.xhgg.2025.100457Polygenic risk score prediction accuracy convergenceLéo Henches0Jihye Kim1Zhiyu Yang2Simone Rubinacci3Gabriel Pires4Clara Albiñana5Christophe Boetto6Hanna Julienne7Arthur Frouin8Antoine Auvergne9Yuka Suzuki10Sarah Djebali11Olivier Delaneau12Andrea Ganna13Bjarni Vilhjálmsson14Florian Privé15Hugues Aschard16Institut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, FranceDepartment of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USAInstitute for Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki, FinlandInstitute for Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki, FinlandInstitut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, FranceNational Centre for Register-Based Research, Aarhus University, 8210 Aarhus, DenmarkInstitut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, FranceInstitut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, FranceInstitut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, FranceInstitut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, FranceInstitut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, FranceIRSD, Université de Toulouse, INSERM, INRAE, ENVT, University Toulouse III - Paul Sabatier (UPS), Toulouse, FranceDepartment of Computational Biology, University of Lausanne, Lausanne, SwitzerlandInstitute for Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki, FinlandNational Centre for Register-Based Research, Aarhus University, 8210 Aarhus, Denmark; Bioinformatics Research Centre, Aarhus University, 8000 Aarhus, DenmarkNational Centre for Register-Based Research, Aarhus University, 8210 Aarhus, DenmarkInstitut Pasteur, Université de Paris, Department of Computational Biology, 75015 Paris, France; Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA; Corresponding authorSummary: Polygenic risk scores (PRSs) models trained from genome-wide association study (GWAS) results are set to play a pivotal role in biomedical research addressing multifactorial human diseases. The prospect of using these risk scores in clinical care and public health is generating both enthusiasm and controversy, with varying opinions among experts about their strengths and limitations. The performance of existing polygenic scores is still limited but is expected to improve with increasing GWAS sample sizes and the development of new, more powerful methods. Theoretically, the variance explained by PRS can be as high as the total additive genetic variance, but it is unclear how much of that variance has already been captured by PRS. Here, we conducted a retrospective analysis to assess progress in PRS prediction accuracy since the publication of the first large-scale GWASs, using data from six common human diseases with sufficient GWAS information. We show that although PRS accuracy has grown rapidly over the years, the pace of improvement from recent GWAS has decreased substantially, suggesting that merely increasing GWAS sample sizes may lead to only modest improvements in risk discrimination. We next investigated the factors influencing the maximum achievable prediction using whole-genome sequencing data from 125,000 UK Biobank participants and state-of-the-art modeling of polygenic outcomes. Our analyses suggest that increasing the variant coverage of PRS, using either more imputed variants or sequencing data, is a key component for future improvements in prediction accuracy.http://www.sciencedirect.com/science/article/pii/S2666247725000600polygenic risk predictionGWASsample sizepolygenicity
spellingShingle Léo Henches
Jihye Kim
Zhiyu Yang
Simone Rubinacci
Gabriel Pires
Clara Albiñana
Christophe Boetto
Hanna Julienne
Arthur Frouin
Antoine Auvergne
Yuka Suzuki
Sarah Djebali
Olivier Delaneau
Andrea Ganna
Bjarni Vilhjálmsson
Florian Privé
Hugues Aschard
Polygenic risk score prediction accuracy convergence
HGG Advances
polygenic risk prediction
GWAS
sample size
polygenicity
title Polygenic risk score prediction accuracy convergence
title_full Polygenic risk score prediction accuracy convergence
title_fullStr Polygenic risk score prediction accuracy convergence
title_full_unstemmed Polygenic risk score prediction accuracy convergence
title_short Polygenic risk score prediction accuracy convergence
title_sort polygenic risk score prediction accuracy convergence
topic polygenic risk prediction
GWAS
sample size
polygenicity
url http://www.sciencedirect.com/science/article/pii/S2666247725000600
work_keys_str_mv AT leohenches polygenicriskscorepredictionaccuracyconvergence
AT jihyekim polygenicriskscorepredictionaccuracyconvergence
AT zhiyuyang polygenicriskscorepredictionaccuracyconvergence
AT simonerubinacci polygenicriskscorepredictionaccuracyconvergence
AT gabrielpires polygenicriskscorepredictionaccuracyconvergence
AT claraalbinana polygenicriskscorepredictionaccuracyconvergence
AT christopheboetto polygenicriskscorepredictionaccuracyconvergence
AT hannajulienne polygenicriskscorepredictionaccuracyconvergence
AT arthurfrouin polygenicriskscorepredictionaccuracyconvergence
AT antoineauvergne polygenicriskscorepredictionaccuracyconvergence
AT yukasuzuki polygenicriskscorepredictionaccuracyconvergence
AT sarahdjebali polygenicriskscorepredictionaccuracyconvergence
AT olivierdelaneau polygenicriskscorepredictionaccuracyconvergence
AT andreaganna polygenicriskscorepredictionaccuracyconvergence
AT bjarnivilhjalmsson polygenicriskscorepredictionaccuracyconvergence
AT florianprive polygenicriskscorepredictionaccuracyconvergence
AT huguesaschard polygenicriskscorepredictionaccuracyconvergence