Genome-wide association studies have been crucial for identifying genetic variants linked to diseases, yet current methods primarily focus on single nucleotide polymorphisms (SNPs) individually. This common practice is prone to errors, missing genetic variations that could contribute to disease (false negatives) on the one hand, and yielding others that cannot be repeated (likely false positives).
A new study reported by Dr. Heping Zhang, PhD, the Susan Dwight Bliss Professor of Biostatistics, and Dr. Yiran Jiang, PhD, a postdoctoral associate at Yale School of Public Health, introduces a more robust and accurate technique that synthesizes the information of SNPs in local regions and reduces both false positive and false negative errors.
The Regional Association Score (RAS) method proposed by Jiang and Zhang quantifies associations between SNPs and disease traits by converting regional SNP data into a time-series format. The RAS method applies a change point detection algorithm that Dr. Zhang originally developed for time series data monitoring of patients with HIV.
Specifically, their simulation studies show that while controlling the false discovery rate, RAS increases detection power by over 20% in complex genetic landscapes, particularly in cases where causal variants are sparse. Additionally, when applied to real genetic data from the Adolescent Brain Cognitive Development (ABCD) study, the method successfully identified regions associated with mental health traits, reinforcing its potential impact on disease research.
This study highlights an innovative step forward in genetic research as a result of the elegant use of statistical methods and theory, offering a more accurate and efficient approach for detecting disease-related genetic variants. The authors anticipate that RAS could significantly enhance genetic studies and precision medicine, paving the way for improved disease prediction and treatment strategies.
Empowering Genome-Wide Association Studies via a Visualizable Test Based on Regional Association Score (PNAS, published online February 25, 2025)
— Janice Hur