zeitzeiger is a package for regularized supervised learning on high-dimensional data from an oscillatory system. zeitzeiger can quantify rhythmic behavior, make accurate predictions, identify major patterns and important features, and detect when the oscillator is perturbed.

Update (Nov 2018): ZeitZeiger now uses limma internally, which makes training (previously the slowest step by far) about 250x faster. Additional optimizations have made calculation of the sparse principal components about 6x faster, and prediction about 20% faster.

For details about the method and to see how we used it to analyze circadian gene expression in mice, check out Hughey et al. (2016) and the accompanying results.

To see how we used zeitzeiger to analyze the phasing of circadian clocks in humans and other mammals, check out Hughey and Butte (2016) and the accompanying results.

To see how we used zeitzeiger to predict circadian time from gene expression in human blood, check out Hughey (2017) and the accompanying results.

Installation

1. Install BiocManager.

if (!requireNamespace('BiocManager', quietly = TRUE))
install.packages('BiocManager')
2. If you use RStudio, go to Tools → Global Options… → Packages → Add… (under Secondary repositories), then enter:

You only have to do this once. Then you can install or update the package by entering:

BiocManager::install('zeitzeiger')

Alternatively, you can install or update the package by entering:

BiocManager::install('zeitzeiger', site_repository = 'https://hugheylab.github.io/drat/')

Usage

For an introduction to the package, read the vignette. For detailed help on specific functions, check out the reference documentation.