SAIGEgds

SAIGEgds is a high-performance statistical R package designed for large-scale Phenome-wide Association Studies (PheWAS). PheWAS is a powerful tool for discovering and replicating genetic associations across a wide range of phenotypes. To address computational challenges in analyzing large cohorts, such as the UK Biobank, SAIGE was introduced to handle case-control imbalance and sample relatedness efficiently. However, SAIGE remains computationally intensive, especially when dealing with thousands of ICD10-coded phenotypes and whole-genome imputed genotype data.

SAIGEgds optimizes the SAIGE method by implementing it in C++ codes and incorporating sparse genotype dosages and an efficient genomic data structure file format. This results in a 5-6 times faster analysis compared to the original SAIGE R package.