Work

Temporal REGENIE mahdollistaa ajassa muuttuvien geneettisten vaikutusten laajamittaisen kartoittamisen

Year: 2026
Type: dissertation
Abstract: Cyclical biological processes, including diurnal and annual rhythms, shape human physiology, yet standard genome-wide association studies (GWAS) test only for effects that stay constant across time and so cannot represent them. We extend REGENIE, a whole-genome regression framework used at biobank scale, to test genetic effects that vary cyclically with the time of measurement. The method expands the time-of-day or time-of-year covariate into a cosinor basis entered as a genotype interaction. Fitting the sine and cosine terms together gives a phase-free test of rhythmicity, which recovers a time-varying effect at any phase without pre-specifying when it peaks; fitting a single term tests one phase at a time. Time is read directly from the sample timestamp, so the analysis reuses data biobanks already collect and adds no data-collection burden. Including the rhythmic basis in the null model keeps the interaction tests well-calibrated, and the tool ships as a Docker image for reproducible execution. Benchmarking across three serum biomarkers in the UK Biobank, the method recovered the known circadian glucose loci CRY2 and MTNR1B from a single-term diurnal scan, where the original analysis combined time-stratified GWAS with interaction tests whose sine and cosine covariates were constructed outside REGENIE. Temporal REGENIE makes time-varying association testing a routine, low-cost addition to existing biobank analyses, and brings time-varying GxE into large-scale genomics software.
Language: English
Subfield: Genetics
Open Access status: green
OpenAlex ID: W7213992792
Cites: 30
Cited by:
Related to: 10