Machine-Learning Assisted Rapid Selection of Influenza Vaccine Seeds

Machine-Learning Assisted Rapid Selection of Influenza Vaccine Seeds

Innovation

Innovators at the University of Missouri (MU) have developed a machine-learning assisted framework to reduce the time needed for influenza vaccine seed identification from months to days. Background Vaccination remains crucial in preventing influenza infections. Yet, the virus’s ability to alter its hemagglutinin (HA) and neuraminidase (NA) often enables it to evade immunity conferred by vaccination or prior infection. Annual vaccine updates enable maintained protection from circulating viruses.

Background

Vaccine seed identification requires antigenic alignment, genetic stability, and high yield production. Conventional methods rely on laborious and time-consuming iterative rounds of physical passages and screening susceptible to catastrophic delays that can cost lives due to delayed vaccinations.

MU researchers have developed a machine-learning assisted framework that can identify naturally circulating, antigenically matched, high-yield influenza vaccine strains from clinical samples in days rather than the months required by conventional methods, enabling accelerated vaccine development, and avoidance of delays that can cost lives.

State of Development

Methods were validated through successful prediction of candidate H1N1 (A(H1N1)pdm09) viruses.

Inventors

Xiufeng (Henry) Wan
Cheng Gao

Applications
Fast identification of antigenically matched, genetically stable, high-yield influenza vaccine seeds from naturally occurring strains.

Advantages
Reduces the time needed to identify vaccine seed from months to days.

Additional Details