A literature-derived benchmarking database for single-ancestry polygenic risk score construction methods with spectral ranking inferences.
- 1Department of Mathematics, University of Pennsylvania, 209 South 33rd Street, Philadelphia, PA 19104-6395, United States.
- 2Department of Biostatistics, Epidemiology and Informatics, Perelman School of Medicine, University of Pennsylvania, 3600 Civic Center Boulevard, Philadelphia, PA 19104-6021, United States.
- 3Department of Statistics and Data Science, Washington University in St. Louis, 1 Brookings Drive, St. Louis, MO 63130, United States.
Abstract
Polygenic risk scores (PRSs) have emerged as a valuable tool for genetic risk prediction and stratification in human diseases. Over the past decade, extensive methodological efforts have focused on improving the predictive power of PRS, leading to the development of numerous methods for PRS construction. Benchmarking these various methods thus becomes an essential task that is crucial for guiding future PRS applications. While studies have benchmarked subsets of these methods on specific phenotypes and cohorts, the resulting evidence remains fragmented, with a lack of work that comprehensively assess the relative performance of the various PRS methods. We addressed this gap by constructing a PRS method benchmarking database, synthesizing published results in 2009-2025 and conducting spectral ranking inference with uncertainty quantification to rank 14 single ancestry summary statistics-based PRS methods that had sufficient benchmarking evidence to support reliable ranking. We constructed rankings using two complementary sources: original method-development studies and applications/benchmarking studies. While the highest-ranked methods (LDpred2 and AnnoPred) and the lowest-ranked method (C + T) were consistently identified from both sources, the relative ordering of most methods showed moderate variability. We further constructed phenotype-specific rankings, providing more detailed insights into the robustness and phenotype-specific strengths of individual methods. Collectively, the overall and phenotype-specific rankings of the PRS methods, along with the detailed model performance data curated from the literature, form a dynamic benchmarking database that can be continuously updated as new PRS methods and published benchmarking evidence emerge, providing a practical resource for future PRS method selection and application.
Related
- Time versus causal inferences: Action-outcome learning and prediction differentially shape the experience of causality and simultaneity
- How metapopulation dynamics shape evolutionary inferences from resurrection ecology
- Reference mapping and whole-genome alignment yield comparable phylogenomic inferences
- A deep dive into the fossil record of diving beetles (Coleoptera: Dytiscidae): Bayesian inferences elucidate their origin and pulsed diversification at the genus level
- Paleoenvironmental inferences from ungulate dental microwear texture analysis during the Middle to Upper Palaeolithic transition in Covalejos (northern Iberia)
BibTeX
@article{sebastian2026literaturederived,
title = {{A literature-derived benchmarking database for single-ancestry polygenic risk score construction methods with spectral ranking inferences.}},
author = {Chris Sebastian and Mengxin Yu and Jin Jin},
journal = {Briefings in bioinformatics},
year = {2026},
month = sep,
doi = {10.1093/bib/bbag598},
url = {https://europepmc.org/article/MED/42960992},
note = {CC BY},
}
RIS
TY - JOUR
TI - A literature-derived benchmarking database for single-ancestry polygenic risk score construction methods with spectral ranking inferences.
AU - Chris Sebastian
AU - Mengxin Yu
AU - Jin Jin
PY - 2026
DA - 2026/09/01
JO - Briefings in bioinformatics
DO - 10.1093/bib/bbag598
UR - https://europepmc.org/article/MED/42960992
AB - Polygenic risk scores (PRSs) have emerged as a valuable tool for genetic risk prediction and stratification in human diseases. Over the past decade, extensive methodological efforts have focused on improving the predictive power of PRS, leading to the development of numerous methods for PRS construction. Benchmarking these various methods thus becomes an essential task that is crucial for guiding future PRS applications. While studies have benchmarked subsets of these methods on specific phenotypes and cohorts, the resulting evidence remains fragmented, with a lack of work that comprehensively assess the relative performance of the various PRS methods. We addressed this gap by constructing a PRS method benchmarking database, synthesizing published results in 2009-2025 and conducting spectral ranking inference with uncertainty quantification to rank 14 single ancestry summary statistics-based PRS methods that had sufficient benchmarking evidence to support reliable ranking. We constructed rankings using two complementary sources: original method-development studies and applications/benchmarking studies. While the highest-ranked methods (LDpred2 and AnnoPred) and the lowest-ranked method (C + T) were consistently identified from both sources, the relative ordering of most methods showed moderate variability. We further constructed phenotype-specific rankings, providing more detailed insights into the robustness and phenotype-specific strengths of individual methods. Collectively, the overall and phenotype-specific rankings of the PRS methods, along with the detailed model performance data curated from the literature, form a dynamic benchmarking database that can be continuously updated as new PRS methods and published benchmarking evidence emerge, providing a practical resource for future PRS method selection and application.
SN - 1467-5463
SN - 1477-4054
ER -