The hallmarks of cancer consist of ten categories that serve as an organizing principle and framework for understanding neoplastic disease. The ten hallmarks include activating invasion and metastasis, avoiding immune destruction, deregulating cellular energetics, enabling replicative immortality, evading growth suppressors, genome instability and mutation, inducing angiogenesis, resisting cell death, sustaining proliferative signaling, and tumor-promoting inflammation. In order to facilitate the use of the hallmarks of cancer in genomic studies, we have undertaken a systematic methodology to map genes to each individual hallmark of cancer. Assignment of genes to the individual hallmarks was performed leveraging gene ontology (GO) annotations. Specifically, for each hallmark, a term list was generated by cancer research experts, with biological terms associated with each hallmark. This term list was used to search and identify matching GO terms. Matched terms were then mapped to the highest, most specific, representative, non-redundant GO term in the GO hierarchy. Genes for each term were restricted by selecting only human taxon, and high confidence experimental evidence codes. For specific GO terms, where present, the Regulation of ‘GO term’ category was included. The genes for each GO term for each hallmark were combined in a non-redundant list of genes for each hallmark. In the case of the evading growth suppressors category, a different strategy was employed. In this instance, genes were identified through ‘tumor suppressor’ gene/protein annotation in the DGIdb resource and UniProt. Additionally, text-mining methods on biomedical research literature were used to defined a high confidence list of genes for this hallmark. As an example use case of the hallmarks we applied them to the analysis of the TCGA Pan-Cancer RNAseq data set. Clustering across the individual hallmarks led to interesting differences within particular tumor tissue types, as well as shared hallmarks for other tumor types. For example, use of the genome instability and mutation hallmark for Pan-Cancer clustering, revealed a tight tissue-based cluster of low-grade gliomas and GBMs. This result suggests the importance of this hallmark expression in this tumor type. This high confidence mapping of genes to hallmarks of cancer provides a unique dataset to facilitate analyses of cancer genome data.

Citation Format: Jeff Kiefer, Sara Nasser, John Graf, Chinnappa Kodira, Fiona Ginty, Lee Newberg, Anup Sood, Michael E. Berens. A systematic approach toward gene annotation of the hallmarks of cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3589. doi:10.1158/1538-7445.AM2017-3589