Performs the Joint Graphical Lasso for Sparse Inverse Covariance Estimation on Multiple Classes

The Joint Graphical Lasso is a generalized method for estimating Gaussian graphical models/ sparse inverse covariance matrices/ biological networks on multiple classes of data. We solve JGL under two penalty functions: The Fused Graphical Lasso (FGL), which employs a fused penalty to encourage inverse covariance matrices to be similar across classes, and the Group Graphical Lasso (GGL), which encourages similar network structure between classes. FGL is recommended over GGL for most applications. Reference: Danaher P, Wang P, Witten DM. (2013) .

This package runs the Joint Graphical Lasso (JGL) method for estimating sparse inverse covariance matrices across multiple similar datasets.


Danaher P, Wang P, Witten DM. The joint graphical lasso for inverse covariance estimation across multiple classes. Journal of the Royal Statistical Society: Series B (Statistical Methodology). 2014 Mar 1;76(2):373-97.

Install the package from CRAN with install.packages("JGL")


JGL update

This quick 2018 update makes this old package (est. 2012) compatible with current CRAN requirements. No changes to the package's core functions have been made.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


2.3.1 by Patrick Danaher, 2 years ago

Browse source code at

Authors: Patrick Danaher

Documentation:   PDF Manual  

GPL-2 license

Depends on igraph

Imported by fgm.

Depended on by JointNets, pGMGM.

Suggested by EstimateGroupNetwork.

See at CRAN