Researchers have developed RareEffect, a new computational method to estimate the specific impact of rare genetic variants on complex diseases. While common genetic variants are widely used for risk prediction, rare variants have historically been difficult to measure due to their low frequency and the complexity of their biological effects. This new tool changes that by providing stable effect size estimations at both the gene and individual variant levels.
Using data from the UK Biobank, the team applied this method to 100 different traits. Their findings show that RareEffect provides a significant boost to polygenic risk scores. By integrating these rare variant scores with traditional common variant analysis, the team can now identify individuals at high risk for diseases like type 2 diabetes and various cancers who might otherwise be missed. This approach outperforms existing methods that rely on predicted functional impact scores, offering higher accuracy for clinical applications.
The method is designed for efficiency. By implementing a factored spectrally transformed linear mixed model, the researchers reduced computational requirements significantly compared to traditional ridge regression. For binary traits, they also introduced a fast implementation of the Firth bias reduction method, which corrects for case-control imbalances without the standard time penalty.
These results provide a practical framework for future genetic studies. By pinpointing the specific contribution of rare variants, clinicians and researchers have a better opportunity to interpret genetic data and improve the precision of individual disease risk profiles. The tool is now available as part of the SAIGE software package for broader use in the research community.