ECR session
Mar 12, 2026
Scalable and accurate rare-variant association tests for whole genome sequencing time-to-event analysis in large biobanks
Feb 27, 2026
Significance: Rare variants (RV) identified through whole genome sequencing hold great promise for elucidating the genetic basis of disease onset, but existing methods for RV association testing are not well suited for time-to-event phenotypes. Here, we develop GATE-STAAR, a scalable and accurate framework integrating frailty modeling with functional annotations. We propose a rare-variant saddlepoint approximation to handle heavy censoring. Through comprehensive simulations and large-scale analysis of approximately 400 K UK Biobank participants, with replication in approximately 230 K All of Us participants, GATE-STAAR uncovers biologically meaningful RV associations while ensuring rigorous control of type I error. This framework is powerful to dissect the genetic architecture of disease onset and progression and advance precision medicine.