AI Antibody Design to Combat Viruses
Research led by Vanderbilt University Medical Center highlights how artificial intelligence (AI) and protein language models can expedite the design of monoclonal antibodies aimed at combating serious viral infections. The pivotal study, published on November 4 in the journal Cell, emphasizes the potential of these advancements in addressing both current and emergent viral challenges, such as RSV (respiratory syncytial virus) and avian influenza.
Dr. Ivelin Georgiev, the study’s corresponding author and a prominent figure in computational microbiology, characterized this research as a significant milestone towards the ultimate goal of utilizing AI for the efficient design of novel biologics. Georgiev notes, “Such approaches will have significant positive impacts on public health and can be applied to a broad range of diseases, including cancer, autoimmunity, neurological diseases, and many others.” This assertion underscores the far-reaching implications of their findings.
The project received substantial backing through a grant from the Advanced Research Projects Agency for Health (ARPA-H), amounting to up to $30 million, aimed at enhancing the application of AI in antibody development. Perry Wasdin, a data scientist within Georgiev’s lab, served as the first author and contributed significantly throughout the research process.
Co-researchers from a diverse set of institutions, as well as collaborators from Australia and Sweden, successfully demonstrated that their protein language model could generate effective human antibodies that specifically recognized unique antigen sequences of various viruses. Significantly, this process did not necessitate existing antibody sequence templates, indicating a novel approach in antibody development.
Relying on their protein language model, known as MAGE (Monoclonal Antibody Generator), the team trained it on established antibodies against the H5N1 influenza virus. Remarkably, MAGE produced antibodies effective against a different, yet related, strain of influenza not previously encountered. This innovative capability suggests that MAGE could significantly shorten the time needed to develop antibodies for emerging health threats compared to conventional methods, which typically require blood samples or antigen proteins from infected individuals.
Additional co-authors included several Vanderbilt faculty members with expertise across various scientific disciplines. The funding, shared between ARPA-H and NIH grants, illustrated the collaborative effort underlying this promising research initiative. Overall, the outcomes indicate a transformative potential for AI technologies in the realm of public health, particularly in the swift response to viral threats.