How are immune cells able to recognize and kill cells infected with SARS-CoV-2? Why does one person’s lymph nodes mount a strong immune response and another person’s doesn’t?
“We need to answer questions like these to better prepare for future infectious diseases,” says Bruce Walker, MD, director of the Ragon Institute of Mass General Hospital, MIT and Harvard. “For it’s the immune system—the body’s line of defense against infection--that will deliver us from future threats.”
Immunology and vaccinology are still in their infancy, he argues, and while scientists alone can make some headway, a powerful tool could supercharge their work: artificial intelligence.
That’s the idea behind a new initiative by the Ragon Institute, MIT’s Abdul Latif Jameel Clinic for Machine Learning in Health and the MIT Schwarzman College of Computing. Launched with a $2 million gift by longtime Ragon Institute supporters Mark and Lisa Schwartz, the program will fund graduate students and postdoctoral fellows to be mentored by both an MIT faculty member with AI expertise and a Ragon Institute immunologist.
“The human immune system involves trillions of cells communicating with one another. In an immune response, there are so many elements working at the same time that it’s nearly impossible for the human brain to figure out how it’s all interconnected,” says Walker. “The power of this collaboration is that it brings immunologists and computer scientists together, applying a technology that is ideally suited to finding patterns in a complex system.”
In doing so, they can jointly redefine the field, says collaborator Regina Barzilay, PhD, AI faculty lead at the Jameel Clinic and a 2017 MacArthur fellow. “AI has already been successfully applied to many areas of life science,” she says. “Surprisingly, its application to immunology is still in very early stages. While one can readily apply existing tools from computer vision and natural language processing, real breakthroughs in understanding the immune system with AI require development of novel methods that can accurately capture complex, dynamic interactions inherent in the immune response.”
As the COVID-19 pandemic has demonstrated all too well, humankind is extremely vulnerable to new infectious diseases. Says Walker: “The current states of AI and immunology research and the urgency lent by COVID make this the right time for this initiative.”