The immune system is almost fantastically complex, and many basic questions remain unanswered about how it manages to keep us safe from intruders without attacking our own tissues. The Santa Fe Institute helped pioneer the field of theoretical immunology, seeding a vibrant community of modelers whose work has led to new HIV treatments, better methods to choosing vaccine strains, and improved cancer immunotherapy treatments. The connections have also circled back to create improvements in computer science, with applications in computer security, anomaly detection in manufacturing, robotics, and more.
A June 8-10 SFI working group, Distributed Computing Perspectives on Theoretical Immunology, gathered a diverse community of researchers to both revisit classic problems in immunology with a fresh face and ask what new questions have arisen, taking advantage of recent developments in both biology and computer science.
Consider, for example, the now-famous spike protein on the new coronavirus. The immune system recognizes the virus by targeting the spike protein and a few other antigens, ignoring many other proteins on the surface that might serve as red flags for the intruder. This strategy gives the immune system fewer proteins to remember and reduces the chances that it will react to the body’s own proteins, creating autoimmunity — with the downside that a few mutations in those key proteins can allow the virus to effectively don an invisibility cloak. This raises questions that an algorithmic mindset might shed light on: How does the immune system decide which proteins to remember? How might we quantify the trade-offs of this strategy compared to alternatives, particularly with pathogens that evolve fast, like the new coronavirus or the flu?
Another example of an area ripe for exploration is the analogy between cybersecurity and immunity. By comparing the two systems, the workshop planned to tackle questions including: When does an effective defense invite increasingly damaging attacks? Can defense be structured to make less damaging attacks advantageous to the attacker? Can some attacks simply be tolerated, so that the attackers face less evolutionary pressure, with the goal of creating an equilibrium in the arms race?
This team of computer scientists, mathematicians, experimental immunologists, and experienced modelers aims to crack these puzzles and more.
This working group was co-hosted by SFI External Professor Stephanie Forrest (Arizona State University), Saket Navlakha (Cold Spring Harbor Laboratory), and Joshua Daymude (Arizona State University)