Leveraging AI to Aid the Development of T Cell Vaccine Candidates

An exciting collaboration between the Ragon Institute and the Jameel Clinic at MIT has achieved a significant milestone in leveraging artificial intelligence (AI) to aid the development of T cell vaccine candidates.

Ragon faculty member Gaurav Gaiha, MD, DPhil, and MIT Professor Regina Barzilay, PhD, AI lead of the Jameel Clinic for AI and Health, have published research in Nature Machine Intelligence introducing MUNIS-a deep learning tool designed to predict CD8+ T cell epitopes with unprecedented accuracy. This advancement has the potential to accelerate vaccine development against various infectious diseases.

The project marks a major first outcome from the Mark and Lisa Schwartz AI/ML Initiative at the Ragon Institute, which aims to integrate artificial intelligence, machine learning, and translational immunology to prevent and cure infectious diseases of global importance. This initiative was made possible through the generous support of Ragon Institute Board Chair Mark Schwartz and his wife, Lisa Schwartz.

By combining the Gaiha Lab's expertise in T cell immunology with the Barzilay Lab's pioneering work in AI, the team-led by co-first authors Jeremy Wohlwend, PhD, and Anusha Nathan, PhD-sought to address a longstanding challenge in vaccine development: the rapid and accurate identification of T cell epitopes in foreign pathogens. Epitopes are specific regions of an antigen that are recognized by the body's immune cells and are critical for activating targeted immune responses.

Traditional methods for predicting epitopes often fall short in speed and accuracy. By integrating machine learning, researchers can now achieve faster and more efficient identification of T cell epitopes.

Using a curated dataset of over 650,000 unique human leukocyte antigen (HLA) ligands and cutting-edge AI architectures, MUNIS significantly outperformed existing epitope prediction models. The tool was validated using experimental data from influenza, HIV, and Epstein-Barr virus (EBV) and was able to identify novel immunogenic epitopes in EBV, a virus that has been extensively studied. Remarkably, MUNIS achieved accuracy comparable to experimental stability assays, another form of epitope prediction, demonstrating its potential to reduce laboratory burdens and streamline vaccine design.

This is our first paper at the intersection of AI and immunology. Through this collaboration with Dr. Gaiha and his team, we learned a lot about this fascinating field and are excited about the immense possibilities in using AI algorithms to model the intricacies of the immune system."

Regina Barzilay, MIT Professor

A key factor in the development of MUNIS was the collaboration between immunologists and computer scientists. The partnership leveraged the unique skills and expertise of each team, ensuring the tool's effectiveness in addressing biological complexities.

"This is a wonderful application of artificial intelligence that benefited greatly from insights shared by both computer scientists and immunologists," Gaiha said. "The credit lies with the initiative for bringing us together, which has led to the creation of an exciting new tool for immunology and vaccine design."

The implications of this breakthrough extend beyond vaccine research. By providing a reliable method to predict immunodominant epitopes which are those most-easily recognized by the immune system, MUNIS lays the foundation for applications in cancer T cell immunotherapy and autoimmunity research. As the global community continues to confront emerging infectious diseases, tools like MUNIS offer promise for enhanced preparedness.

This innovation underscores the Ragon Institute's commitment to advancing science at the intersection of immunology and technology to save lives and promote global health.