Can Scientists Predict a Peanut’s Flavor?
For peanut breeders, developing a new variety is a long process. A line may show promise in the field, but determining whether consumers will enjoy its flavor often comes much later.
Now, researchers at the University of Georgia are working to change that. According to the University of Georgia College of Agricultural and Environmental Sciences, scientists are developing a machine learning model that uses chemical data to evaluate the flavor potential of roasted peanuts. The goal is to supplement traditional sensory testing, which can be time-consuming and expensive.
Researchers say the new approach requires only about one-tenth of the sample material needed for conventional flavor evaluations. That could allow breeders to screen breeding lines much earlier in the development process and identify poor-flavor candidates before investing additional years of research and seed production.
The project could also help researchers identify genetic markers associated with desirable flavor traits. Ultimately, that would give breeders another tool to develop peanut varieties that combine strong agronomic performance with the flavor quality consumers have come to expect from U.S. peanuts.
