Abstract
Neoepitopes have attracted much attention as targets for immunotherapy against cancer. Therefore, efficient neoepitope screening technology is an essential step in the development of personalized vaccines. Circular RNAs (circRNAs) are generated by back-splicing and have a single-stranded continuous circular structure. So far, various circRNAs have been poorly characterized, though new evidence suggests that a few translated circRNAs may play a role in cancer. In the present study, circRNA was used as a source of neoepitope, a novel strategy as circRNA-derived neoepitopes have never been previously explored. The present study reports CIRC_neo (circRNA-derived neoepitope prediction pipeline), which is a comprehensive and automated bioinformatic pipeline for the prediction of circRNA-derived neoepitopes from RNA sequencing data. The computational prediction from sequencing data requires complex computational workflows to identify circRNAs, derive the resulting peptides, infer the types of human leukocyte antigens (HLA I and HLA II) in patients, and predict the neoepitopes binding to these antigens. The present study proposes a novel source of neoepitopes. The study focused on cancer-specific circRNAs, which have greatly expanded the source pool for neoepitope discovery. The statistical analysis of different features of circRNA-derived neoepitopes revealed that circRNAs could produce long proteins or truncated proteins. Because the peptides were completely foreign to the human body, they could be highly immunogenic. Importantly, circRNA-derived neoepitopes capable of binding to HLA were discovered. In the current study, circRNAs were systematically analyzed, revealing potential targets and novel research clues for cancer diagnosis, treatment, and prospective personalized vaccine research.