Neon Therapeutics, Inc. announced publication in the scientific journal Immunityof a breakthrough process for predicting which neoantigens will be presented by MHC class II molecules in the tumor microenvironment. Predicting the relevant cancer-specific antigens is a crucial precursor to developing immunotherapies that effectively train T cells to traffic to the tumor and destroy malignant cells. In the paper, titled “Defining HLA-II ligand processing and binding rules with mass spectrometry enhances cancer epitope prediction,” Neon’s proprietary mono-allelic profiling technology called MAPTAC™ facilitated the development of convolutional neural network-based predictors. These algorithms achieved up to a 61-fold improvement in predicting MHC class II peptides compared to publicly available tools. This MHC class II technology will be integrated into Neon’s RECON bioinformatics platform and is expected to improve the efficacy of immunotherapies developed by Neon by predicting recruitment of CD4+ T cells, which are believed to be important in controlling tumor growth. Neon will be integrating its MHC class II prediction tool into its RECON bioinformatics platform, which is powered by machine-learning neural networks and trained on proprietary MAPTAC data sets. RECON identifies, predicts and selects the most therapeutically relevant neoantigen targets associated with each patient’s tumor. This information is used to design personal immunotherapies customized to each patient’s unique mutational fingerprint and precision immunotherapies targeting shared cancer neoantigens. The findings in the Immunity publication demonstrate that Neon’s proprietary class II prediction algorithms substantially outperform NetMHCIIpan, the current benchmark for class II prediction. Key findings in the research include the development of novel proteomic strategies that resolve over 40 MHC class II motifs and the observation that intra-tumoral MHC class II presentation is dominated by professional antigen presenting cells (APCs) rather than tumor cells. Tracking which tumor epitopes are most readily phagocytosed and presented by APCs further enhances the ability to pinpoint therapeutically relevant epitopes. A proprietary approach powering Neon’s RECON bioinformatics platform for MHC class I molecules was first described in an earlier article published in Immunity in 2017.