AI Accelerates Antibody Design to Combat Viral Infections
Groundbreaking research has revealed that artificial intelligence (AI) can significantly enhance the design of monoclonal antibodies aimed at preventing or mitigating severe viral infections. A multi-institutional study led by researchers at Vanderbilt University Medical Center emphasizes the potential of AI and “protein language” models in developing effective therapeutic interventions against both existing and emerging viral threats, such as respiratory syncytial virus (RSV) and avian influenza.
The findings, published in the journal Cell, mark a significant step toward leveraging computational tools for the efficient design of novel biologics. According to the study’s corresponding author, Ivelin Georgiev, Ph.D., this research is a vital early milestone in a larger effort to utilize AI in the clinical development of new treatments. Georgiev, who serves as a professor of Pathology, Microbiology and Immunology and directs the Vanderbilt Program in Computational Microbiology and Immunology, stated, “Such approaches will have a significant positive impact on public health and can be applied to a broad range of diseases, including cancer, autoimmunity, neurological diseases, and many others.”
The collaborative team, which included scientists from various institutions across the United States, Australia, and Sweden, demonstrated that AI-driven protein language models could create functional human antibodies. These antibodies are designed to recognize specific viral antigens without needing any part of the antibody sequence as a template.
Innovative Use of AI in Antibody Generation
The research team employed a protein language model dubbed MAGE (Monoclonal Antibody Generator), which was trained on previously characterized antibodies against a known strain of the H5N1 influenza virus. This model successfully generated antibodies targeting a related, yet unencountered strain of the virus. The researchers concluded that MAGE could potentially expedite the development of antibodies against new health threats, outperforming traditional antibody discovery methods that typically rely on blood samples from infected individuals or antigen proteins derived from novel viruses.
Dr. Perry Wasdin, a data scientist in Georgiev’s lab and the paper’s first author, contributed extensively throughout the research process. The collaborative effort yielded promising results, indicating that AI could revolutionize the field of antibody design.
Other co-authors from Vanderbilt include Alexis Janke, Ph.D., Toma Marinov, Ph.D., Gwen Jordaan, Olivia Powers, Matthew Vukovich, Ph.D., Clinton Holt, Ph.D., and Alexandra Abu-Shmais.
The implications of this study extend beyond viral infections. The ability to harness AI for antibody design holds promise for addressing a wide array of diseases, enhancing the speed and efficiency of therapeutic development. As Dr. Georgiev outlined, the ultimate objective is to create new biologics from scratch, translating these innovations into clinical applications that could transform treatment paradigms across multiple medical fields.
For further details, refer to the study: Perry T. Wasdin et al, “Generation of antigen-specific paired-chain antibodies using large language models,” published in Cell in October 2025. DOI: 10.1016/j.cell.2025.10.006.
