Set-Based Tests for Genetic Association Studies with Interval-Censored Competing Risks Outcomes.

Area of Science:

• Genetics
• Biostatistics
• Epidemiology

Background:

• Large genetic databases like the UK Biobank offer insights into complex diseases.
• Genetic association studies often face challenges with time-to-event data, specifically interval-censored outcomes and competing risks.
• Existing set-based genetic analysis tools lack methods for interval-censored data with competing risks.

Purpose of the Study:

• To develop novel statistical methods for set-based genetic association analysis.
• To address the specific challenges of interval-censored outcomes and competing risks in genetic studies.
• To enable robust genetic inference from large-scale biobanks using comprehensive genetic and phenotypic data.

Main Methods:

• Proposed two new set-based inference procedures: interval-censored competing risks sequence kernel association test (crSKAT) and interval-censored competing risks Burden (crBurden) test.
• crSKAT is a variance components approach suitable for heterogeneous genetic variant signals within a set.
• crBurden test is designed for homogeneous genetic variant signals within a set.

Main Results:

• Simulation studies demonstrated that the proposed methods effectively control type I error rates and enhance statistical power compared to alternative approaches.
• The new methods showed superiority in analyzing interval-censored competing risks data.
• The developed tests were successfully applied to UK Biobank data.

Conclusions:

• The newly developed crSKAT and crBurden tests provide powerful and reliable tools for genetic association studies with interval-censored outcomes and competing risks.
• These methods advance the analysis of complex diseases by enabling the use of comprehensive genetic and phenotypic data from large biobanks.
• Application to UK Biobank data identified genes associated with fracture risk, considering death as a competing outcome.