LPWC: Lag Penalized Weighted Correlation for Time Series Clustering

Authors: Thevaa Chandereng and Anthony Gitter

Overview

Lag Penalized Weighted Correlation (LPWC) is a method for clustering short time series data. It is designed to identify groups of biological entities (for example, genes or phosphosites) that exhibit the same pattern of activity changes over time. LPWC allows lags to incorporate delayed responses in the biological data. For example, two genes may have similar expression changes over time, but one initiates those changes 5 minutes after the other. LPWC also supports irregular time intervals between time points collected in biological data. The LPWC website is available here.

Installation

Prior to analyzing your data, the R package needs to be installed.

The easiest way to install LPWC is through CRAN:

install.packages("LPWC")

There are other additional ways to download LPWC. The first option is most useful if want to download a specific version of LPWC (which can be found at https://github.com/gitter-lab/LPWC/releases).

devtools::install_github("gitter-lab/LPWC@vx.xx.x")
# OR 
devtools::install_version("LPWC", version = "x.x.x", repos = "http://cran.us.r-project.org")

The second option is to download through GitHub.

devtools::install_github("gitter-lab/LPWC")

After successful installation, the package must be loaded into the working space:

library(LPWC)

Usage

See the vignette for usage instructions.

Examples

The LPWC example repository is available here. All the example code can be executed in binder.

Reference

If you use LPWC, please cite

Chandereng, T., Gitter, A. Lag penalized weighted correlation for time series clustering. BMC Bioinformatics 21, 21 (2020). https://doi.org/10.1186/s12859-019-3324-1

License

LPWC is available under the open source MIT license.