Micronutrient deficiencies are estimated to affect more than two billion people worldwide. One method of addressing this public health issue is Biofortification: the improvement of crop nutritional quality through plant breeding and/or agronomic approaches (Bouis and Welch 2010). Biofortification has been a cost-effective solution to help mitigate micronutrient deficiencies in certain regional contexts, as the added nutritional value is bred into the seed/plant without requiring additional inputs, regular transportation networks, or other changes to production or consumption. Breeders typically carry out biofortification in tandem with improvement of other traits such as yield and disease resistance.

A central challenge in breeding programs is how to identify varieties that have the best combinations of multiple traits in a cost-effective and timely manner. This research project seeks to develop an AI-enabled toolkit that will: create predictors of nutritional quality and develop machine learning tools to evaluate the quality of the breeding material. In addition to improving biofortification, the open-source code, documentation, and datasets will be useful for the broader community to train and benchmark similar efforts.