Code and data for the article Core liver homeostatic co-expression networks are preserved but respond to perturbations in an organism and disease specific manner

This repository contains R analysis code and (nearly) all data necessary to reproduce the analyses in the article Core liver homeostatic co-expression networks are preserved but respond to perturbations in an organism and disease specific manner.

The code runs in reasonably recent versions of R (R 3.5.0 or newer should work). The code requires various R packages available from CRAN and Bioconductor. At a minimum, the code needs the WGCNA package and its dependencies and packages anRichment and anRichmentMethods. Some of the plotting code needs the vioplot package.

Each script requires that it be run in with the directory where it resides to be set as the working directory in R (use setwd to set the working directory. Each script creates, as needed, up to 3 subdirectories named Plots (for PDF plots), Results (for tables with result, usually in csv format and compressed using gzip) and RData (for intermediate data that can be re-used when re-running the analysis to save time).

Folders in the main repository are named with a numerical prefix that roughly represents the order in which the scripts or analysis steps should be performed.

1. 010-Preprocessing: Preprocessing of the mouse liver RNA-seq data.
2. 020-IndividualAnalysis: Individual gene DE analysis of the mouse liver RNA-seq data.
3. 030-NetworkAnalysis: The first part of the analysis should be run before the network stability analysis in 029-StabilityAnalysis is carried out. The second part needs the results of the stability analysis.
4. 029-StabilityAnalysis: Network stability analysis of the mouse liver RNA-seq data. This should be run after the first step in the network analysis in folder 030-NetworkAnalysis.
5. 039-DownloadGEOData: This is an optional step of downloading the human NAFLD (GSE126848) data from GEO. The data is already included in the Data folder.
6. 040-Preprocessing-Suppli2019-HumanNAFLD-GSE126848: Preprocessing of the human NAFLD data.
7. 050-IndividualAnalysis-Suppli2019-HumanNAFLD-GSE126848: Individual gene DE analysis of the human NAFLD data.
8. 060-NetworkAnalysis-Suppli2019-HumanNAFLD-GSE126848 and 059-StabilityAnalysis-Suppli2019-HumanNAFLD-GSE126848: as with the mouse network analysis, the first part of the network analysis should be carried out first to set up the network stability analysis whose results are used in the second part of the main network analysis script.
9. 070-ModulePreservation-MouseAndHuman: Network module preservation analysis between mouse and human NAFLD data.
10. 110-Preprocessing-TCGA: Preprocessing of TCGA data. Note that the first reformatting step requires the original TCGA files that are too large for github; either start from the second step (after reformatting) or contact us for the original TCGA files.
11. 120-IndividualAnalysis-TCGA: Individual gene DE analysis of TCGA data using limma-voom.
12. 120-IndividualAnalysis-TCGA-V2-DESeq2: Individual gene DE analysis of TCGA data using DESeq2.
13. 120-IndividualAnalysis-TCGA-CompareLimmaAndDESeq2: A short script comparing limma-voom and DESeq2 results from TCGA.
14. 130-NetworkAnalysis-TCGAPhase and 129-StabilityAnalysis-TCGA: The first part of the network analysis should be carried out first to set up the network stability analysis whose results are used in the second part of the main network analysis script.
15. 140-ModulePreservation-TCGA: Module preservation calculations between TCGA and the mouse and human NAFLD data.
16. 150-TCGA-eQTLAnalysis: eQTL analysis of TCGA data and modules.
17. 200-MODifieR: Network analysis using MODA, ARACNE and MRNet and comparison with WGCNA.
18. 210-EnrichmentInPPINetworks: Enrichment analysis of WGCNA modules in PPI networks.

Folder Data contains the raw and preprocessed data used in this study as well as annotation data needed for the analysis. Please note that not all annotation files used by the analysis scripts are included here. For licensing reasons, we are unable to distribute Enrichr libraries and the Molecular Signatures Database (MSigDB). Analysts who wish to re-run and/or adapt our code should (complying with all licensing requirements) download the necessary files from these web sites, or adapt the code to avoid using these files.

Folder Functions contains supporting functions scattered over multiple files.