Statistical Inference of Large-Scale Gaussian Graphical Model in Gene Networks

Provides a general framework to perform statistical inference of each gene pair and global inference of whole-scale gene pairs in gene networks using the well known Gaussian graphical model (GGM) in a time-efficient manner. We focus on the high-dimensional settings where p (the number of genes) is allowed to be far larger than n (the number of subjects). Four main approaches are supported in this package: (1) the bivariate nodewise scaled Lasso (Ren et al (2015) ) (2) the de-sparsified nodewise scaled Lasso (Jankova and van de Geer (2017) ) (3) the de-sparsified graphical Lasso (Jankova and van de Geer (2015) ) (4) the GGM estimation with false discovery rate control (FDR) using scaled Lasso or Lasso (Liu (2013) ). Windows users should install 'Rtools' before the installation of this package.