Scientists Develop BigMHC Model for Accurate Neoantigen Prediction

Scientists from Johns Hopkins University have developed a machine learning model called BigMHC to predict and classify cancer-specific antigens called neoantigens, according to a study published in Nature Machine Intelligence. The researchers employed transfer learning, training the model on mass spectrometry datasets of peptide-MHC fragments and fine-tuning the model’s predictions using T cell receptor (TCR) response assays. While the BigMHC model showed promise in predicting immunogenicity, scientists acknowledged the limitations of machine learning-based predictions and the challenges posed by biological redundancy. Incorporating other ligand-receptor interactions and addressing data redundancy issues could further improve neoantigen prediction models. Despite these challenges, the development of such tools holds great potential for large-scale discovery in personalized immunotherapy.