Skip to main content
ARS Home » Southeast Area » New Orleans, Louisiana » Southern Regional Research Center » Food and Feed Safety Research » Research » Publications at this Location » Publication #414164
Research Project: Development of Aflatoxin Resistant Corn Lines Using Omic Technologies

Location: Food and Feed Safety Research

Title: Predicting fumonisins in Iowa corn: gradient boosting machine learning

Author
item BRANSTAD-SPATES, EMILY - Iowa State University
item Castano-Duque, Lina
item MOSHER, GRETCHEN - Iowa State University
item HURBURGH, JR, CHARLES - Iowa State University
item Rajasekaran, Kanniah - Rajah
item Owens, Phillip
item Winzeler, Hans - Edwin
item BOWERS, ERIN - Iowa State University

Submitted to: Cereal Chemistry
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/29/2024
Publication Date: 8/12/2024
Citation: Branstad-Spates, E., Castano-Duque, L.M., Mosher, G., Hurburgh, Jr, C., Rajasekaran, K., Owens, P.R., Winzeler, H.E., Bowers, E. 2024. Predicting fumonisins in Iowa corn: gradient boosting machine learning. Cereal Chemistry. https://doi.org/10.1002/cche.10824.
DOI: https://doi.org/10.1002/cche.10824

Interpretive Summary: This study evaluated pre-published Illinois-centric and an Iowa-centric predictive models with historical Iowa FUM contamination data using machine learning. Applying a 2 ppm (mg/kg) threshold for fumonisin high and low contamination events, the overall accuracy was 71.08% and 85.39% for the Illinois and Iowa-centric models in 2011. Balanced accuracies were 60.23% and 50.00% for the Illinois andIowa-centric models. For Iowa’s remaining years (testing data), the overall accuracy was 98.10%, and balanced accuracy was 50.00%. Fumonisin analyses determined the top influential predictor for the Illinois-centric model was satellite-acquired normalized difference vegetation index in March, whereas the top predictor for the Iowa-centric model was precipitation in October.

Technical Abstract: 1) Background and Objectives: Fumonisin (FUM), a secondary metabolite from Fusarium spp. poses major concerns for the United States (US) corn industry. This study evaluated a pre-published Illinois-centric predictive model with historical Iowa FUM contamination data using gradient boosting machine (GBM) learning and compared influential predictors with an Iowa-centric model. Corn samples (n=529) were collected from 2010, 2020, and 2021 in Iowa’s 99 counties; 2011 data was used for independent validation (n=89). 2) Findings: Applying a 2 ppm (mg/kg) threshold for FUM high and low contamination events, the overall accuracy was 71.08% and 85.39% for the Illinois and Iowa-centric models in 2011. Balanced accuracies were 60.23% and 50.00% for the Illinois andIowa-centric models. For Iowa’s remaining years (testing data), the overall accuracy was 98.10%, and balanced accuracy was 50.00%. 3) Conclusions: FUM-GBM analyses determined the top influential predictor for the Illinois-centric model was satellite-acquired normalized difference vegetation index (NDVI) (Veg_index) in March, whereas the top predictor for the Iowa-centric model was precipitation (PRCP) in October.