Neoantigens play important roles in cancer immunotherapy. Current methods used for neoantigen prediction focus on the binding between human leukocyte antigens (HLAs) and peptides, which is insufficient for high-confidence neoantigen prediction. In this study, we apply deep learning techniques to predict neoantigens considering both the possibility of HLA-peptide binding (binding model) and the potential immunogenicity (immunogenicity model) of the peptide-HLA complex (pHLA). The binding model achieves comparable performance with other well-acknowledged tools on the latest Immune Epitope Database (IEDB) benchmark datasets and an independent mass spectrometry (MS) dataset. The immunogenicity model could significantly improve the prediction precision of neoantigens. The further application of our method to the mutations with pre-existing T-cell responses indicating its feasibility in clinical application. DeepHLApan is freely available at https://github.com/jiujiezz/deephlapan and http://biopharm.zju.edu.cn/deephlapan.

Tumor-specific neoantigens have attracted much attention since they can be used as biomarkers to predict therapeutic effects of immune checkpoint blockade therapy and as potential targets for cancer immunotherapy. In this study, we developed a comprehensive tumor-specific neoantigen database (TSNAdb v1.0), based on pan-cancer immunogenomic analyses of somatic mutation data and human leukocyte antigen (HLA) allele information for 16 tumor types with 7748 tumor samples from The Cancer Genome Atlas (TCGA) and The Cancer Immunome Atlas (TCIA). We predicted binding affinities between mutant/wild-type peptides and HLA class I molecules by NetMHCpan v2.8/v4.0, and presented detailed information of 3,707,562/1,146,961 potential neoantigens generated by somatic mutations of all tumor samples. Moreover, we employed recurrent mutations in combination with highly frequent HLA alleles to predict potential shared neoantigens across tumor patients, which would facilitate the discovery of putative targets for neoantigen-based cancer immunotherapy. TSNAdb is freely available at http://biopharm.zju.edu.cn/tsnadb.

Tumour antigens have attracted much attention because of their importance to cancer diagnosis, prognosis and targeted therapy. With the development of cancer genomics, the identification of tumour-specific neoantigens became possible, which is a crucial step for cancer immunotherapy. In this study, we developed software called the tumour-specific neoantigen detector for detecting cancer somatic mutations following the best practices of the genome analysis toolkit and predicting potential tumour-specific neoantigens, which could be either extracellular mutations of membrane proteins or mutated peptides presented by class I major histocompatibility complex molecules. This pipeline was beneficial to the biologist with little programmatic background. We also applied the software to the somatic mutations from the International Cancer Genome Consortium database to predict numerous potential tumour-specific neoantigens. This software is freely available from https://github.com/jiujiezz/tsnad.

Neoantigens are optimal tumor-specific targets for T-cell based immunotherapy, especially for patients with “undruggable” mutated driver genes. T-cell immunotherapy can be a “universal” treatment for HLA genotype patients sharing same oncogenic mutations. To identify potential neoantigens for therapy in gastric cancer, 32 gastric cancer patients were enrolled in our study. Whole exome sequencing data from these patients was processed by TSNAD software to detect cancer somatic mutations and predict neoantigens. The somatic mutations between different patients suggested a high interpatient heterogeneity. C>A and C>T substitutions are common, suggesting an active nucleotide excision repair. The number of predicted neoantigens was significantly higher in patients at stage T1a compared to in patients at T2 or T4b. Six genes (PIK3CA, FAT4, BRCA2, GNAQ, LRP1B, and PREX2) were found as recurrently mutated driver genes in our study. Combining with highly frequent HLA alleles, several neoantigens derived from six recurrently mutated genes were considered as potential targets for further immunotherapy.

Neoantigen, as an important member of tumor-specific antigens, has attracted a great deal of attention as a target for immunotherapy. Neoantigens are potential targets for personalized vaccines and adoptive cell transfer therapies. However, most of the neoantigen-targeted immunotherapies in the process are customized and costly. So, we are inclined to find shared neoantigens suitable for more patients. With the help of existing neoantigen prediction algorithms, we found that the most frequent shared neoantigens occurred in more than 1% of patients for 17 tumor types and the ten most frequent shared neoantigens covered approximately 50% of pancreatic cancer patients, providing a potential list of targets for off-the-shelf immunotherapy.