limorhyde2 is an approach to analyze rhythmic, genome-scale data in a way that focuses on effect sizes. limorhyde2 fits linear models based on periodic splines, then moderates the model coefficients using multivariate adaptive shrinkage. The periodic splines capture non-sinusoidal rhythmicity, while adaptive shrinkage learns patterns in the data and accounts for different genomic features having fits with different levels of noise.