Repositiry for the article: "Gene regulatory network inference using mixed-norms regularized multivariate model with covariance selection" by Alain Mbebi & Zoran Nikoloski

This is the repository for the manuscript "Gene regulatory network inference using mixed-norms regularized multivariate model with covariance selection" by Alain J. Mbebi &amp; Zoran Nikoloski. <strong>Organisation</strong> The folder Codes contains the following R scripts with the K-folds cross-validation option to learn the hyperparameters: Mixed_L1L21_GRN.R which computes L1L21-solution Mixed_L1L21G_GRN.R which computes L1L21G-solution Mixed_L2L21_GRN.R which computes L2L21-solution Mixed_L2L21G_GRN.R which computes L2L21G-solution L1L21_Dream5_Scerevisiae_example_run.R is an example run using the L1L21-solution with S. cerevisiae data (Network 4 in DREAM5 challenge) All files needed to successfully run "L1L21_Dream5_Scerevisiae_example_run" are locaded in the folder Codes. 2. The folder Figures contains all figures in the manuscript. 3. The folder Inferred-networks contains all network objects for each dataset and each inference methods in the comparative analysis. <strong>Dependencies and required packages</strong> The following packages are required for the contending approaches in the comparative analysis: "devtools", "foreach", "plyr", "glmnet" and "randomForest". <strong>GENIE3</strong> The GENIE3 package can be installed from: http://bioconductor.org/packages/release/bioc/html/GENIE3.html <strong>TIGRESS</strong> The TIGRESS repository can be obtained from: https://github.com/jpvert/tigress <strong>ENNET</strong> The ENNET repository can be obtained from: https://github.com/slawekj/ennet <strong>PLSNET</strong> The Matlab source code of PLSNET can be obtained from: https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-016-1398-6#Sec17 <strong>PORTIA</strong> The PORTIA repository can be obtained from: https://github.com/AntoinePassemiers/PORTIA <strong>D3GRN</strong> The Matlab source code of D3GRN can be obtained from: https://github.com/chenxofhit/D3GRN <strong>Fused-LASSO</strong> The fused-LASSO repository can be obtained from: https://github.com/omranian/inference-of-GRN-using-Fused-LASSO <strong>ANOVerence</strong> Because of some technical issues (e.g code's accessibility: http://www2.bio.ifi.lmu.de/˜kueffner/anova.tar.gz), we were not able to reproduce ANOVerence results and used the inferred network from DREAM5 challenge instead. 4. Although the codes here were tested on Fedora 29 (Workstation Edition) using R (version 4.2.2), they can run under any Linux or Windows OS distributions, as long as all the required packages are compatible with the desired R version.