NYU research could help farmers cut fertilizer costs and reduce environmental harm.
New research from New York University offers a promising path to growing corn with less fertilizer — using artificial intelligence to pinpoint genes that make plants more efficient at using nitrogen.
“By identifying genes-of-importance to nitrogen utilization, we can select for or even modify certain genes to enhance nitrogen use efficiency in major U.S. crops like corn,” said Gloria Coruzzi, senior author and Carroll & Milton Petrie Professor in NYU’s Department of Biology and Center for Genomics and Systems Biology.
The study, published in The Plant Cell, outlines how NYU scientists combined plant genetics and machine learning to discover groups of genes — called “regulons”—that help control how well corn uses nitrogen.
Fertilizer has been essential to growing higher-yielding crops over the past 50 years. But up to 45% of the nitrogen fertilizer applied to fields goes unused, seeping into groundwater or turning into nitrous oxide — a greenhouse gas 265 times more potent than carbon dioxide.
Corn, the top crop in the U.S., consumes large amounts of nitrogen fertilizer. Yet its low nitrogen use efficiency creates both economic and environmental costs.
To improve this, NYU researchers trained AI models to find patterns in how genes respond to nitrogen in both corn and Arabidopsis, a model plant commonly used in genetic research. They discovered sets of nitrogen-responsive genes and the transcription factors that regulate them — essentially building a map of how nitrogen efficiency is genetically controlled.
“We showed that traits like nitrogen use efficiency or photosynthesis are never controlled by a single gene,” said Coruzzi. “The beauty of machine learning is it learns which sets of genes collectively drive a trait — and identifies the factors that regulate those sets.”
The researchers validated their machine learning predictions through lab studies, confirming two transcription factors in corn — ZmMYB34 and R3 — and one in Arabidopsis — AtDIV1 — regulate genes responsible for nitrogen use. Feeding this data back into the AI model improved its ability to predict nitrogen efficiency across different field-grown corn varieties.
This knowledge could speed up crop improvement. Instead of waiting to see how corn hybrids perform in the field, researchers can now screen seedlings for key gene expression levels and select those with higher nitrogen use efficiency.
“This will not only result in a cost savings for farmers,” said Coruzzi, “but also reduce the harmful effects of nitrogen pollution of groundwaters and nitrous oxide greenhouse gas emissions.”
