Investigators at Cedars-Sinai Health Sciences University have developed a computational framework to improve how genetic data is used to estimate an individual's inherited risk of developing conditions like Alzheimer's disease, type 2 diabetes, high blood pressure and breast cancer.
The framework, called Adaptive Boosting of Pre-trained Polygenic Risk Scores (AB-PRS) and described in Nature Communications, builds on existing polygenic risk scores—tools that estimate disease risk based on many genetic variants across the genome—and identifies additional genetic signals that may not be fully captured by current scoring methods.
"As more people undergo genetic testing, it's becoming increasingly important to make the best use of that information," said Ruowang Li, Ph.D., co-corresponding author of the study and assistant professor of computational biomedicine at Cedars-Sinai. "Our findings aim to improve the accuracy of genetic risk prediction, allowing us to better predict disease risk and guide more personalized care."
Polygenic risk scores are increasingly used in research and clinical settings to estimate genetic susceptibility to common diseases, but their accuracy varies across diseases, data sets and populations. Investigators trained AB-PRS to identify genetic signals that existing risk scores may miss. They tested the approach using genetic data from the UK Biobank and validated the findings in three independent biobanks.
