phenotype classes. The refinement methodology allows the hallmark to contain the most transcriptionally coherent set of genes, which serve as more effective and accurate transcriptional signatures for detecting specific biological processes. By summarizing relevant information from thousands of founder gene sets across diverse collections in MSigDB, hallmarks greatly reduce redundancy and produce more robust and concise GSEA results that facilitate interpretation and follow-up analysis, as well as substantially reduce the chances of missing potentially important findings. The fifty hallmarks described here represent 48% (4,022 out of 8,380) of MSigDB gene sets. The hallmarks also capture 52.7% (452 out of 858) gene sets from the C4 collection of co-expression modules, which cover the global landscape of the transcriptome.
