Gordon, Rachel. “MIT Researchers Develop AI Tool to Improve Flu Vaccine Strain Selection.” MIT News, 28 Aug. 2025, news.mit.edu/2025/vaxseer-ai-tool-to-improve-flu-vaccine-strain-selection-0828.
VaxSeer is a new AI-driven tool developed at MIT’s CSAIL and Jameel Clinic. This innovative system aims to assist in selecting the optimal influenza strains for seasonal vaccines. VaxSeer utilizes deep learning models that have been trained on decades of viral sequences and laboratory test results to simulate viral evolution and antigenicity, which refers to how effectively immune responses might neutralize future strains. The tool combines two main prediction engines, one forecasts which viral strains are likely to dominate, while the other estimates how effectively a vaccine will neutralize those strains. Together, these engines generate a predicted coverage score. A retrospective analysis spanning 10 years revealed that VaxSeer’s strain selections either outperformed or matched those made by the World Health Organization for many flu seasons. The researchers also suggest that future versions of VaxSeer could incorporate additional viral proteins, consider immune history, account for manufacturing constraints, or even be extended to include viruses beyond influenza.
This fascinating development represents a significant advancement at the intersection of machine learning and public health decision-making. Rather than relying primarily on human judgment and best guesses, VaxSeer offers a data-driven, predictive tool that could minimize mismatches in flu vaccines, ultimately improving health outcomes for the population.
Leave a Reply