(233c) Drl Take the Wheel: Employing Deep Reinforcement Learning to Drive Vaccine Models

Highly mutable infectious disease pathogens (hm-IDPs) such as HIV and influenza evolve faster than the human immune system, allowing them to circumvent traditional vaccination approaches and cause over one million deaths annually. Agent-based models can be used to describe the complex interactions that occur between immune cells and hm-IDP-like proteins (antigens) during affinity maturation—the process by which antibodies evolve. As such, compared to existing experimental approaches, these models offer a safe, low-cost, and rapid route to study and assess the immune response to vaccines spanning a wide range of design variables. However, the highly stochastic nature of affinity maturation and vast sequence space of hm-IDPs render standard agent-based modeling approaches intractable for exploring all pertinent vaccine design variables and the subset of immunization protocols encompassed therein. Herein, we present a multi-scale model of the affinity maturation process—taking into account both sequence and structural level information—enabling the rapid screening of potential HIV-1 vaccination protocols (antigen design, concentration, number of inoculations). Further, we use a deep symbolic optimization framework to determine sequence-function determinants in the antigen sequence. Molecular dynamics simulations were then employed to develop an atomistic understanding of the mechanism by which the antibodies evolved are binding to the HIV Env protein. Future works include in vitro and in vivo studies to determine antibody stability and antigen immunogenicity.