The Molecular Signatures Database (MSigDB) (Liberzon et al., 2011), originally developed for use with GSEA and now employed by many similar approaches, remains one of the largest and most popular repositories of gene sets. The latest version of MSigDB consists of seven collections C1-C7 which include: genes grouped by their location in the human genome (C1), canonical pathways and experimental signatures curated from publications (C2), genes sharing cis-regulatory motifs up- or downstream of their coding sequences (C3), clusters of genes co-expressed in microarray compendia (C4), genes grouped according to gene ontology (GO) categories (C5), signatures of oncogenic pathway activation (C6), and a large collection of immunological conditions (C7). All of the gene sets in MSigDB are reviewed, curated, and annotated manually by the MSigDB curator. They are all represented as lists of human gene symbols from the HUGO Gene Nomenclature Committee at the European Bioinformatics Institute (Gray et al., 2015).
