Researchers at the University of Guelph have sprouted what they’re calling a ‘super bean’, with some help from artificial intelligence. The team of dry bean breeders with the Ontario Agricultural College has used the technology to speed up the selection process when cross-breeding seeds. The college’s growing fields in Elora feature a wide range of beans including kidney, pinto, navy and black beans. But the bean breeding and computational biology team are looking to set a new standard of legume. The process begins with selecting female and male seeds to cross breed and create a population. “Then it gives us the ability to select out of this population, the superior beings that are producing high yield, high quality, earlier maturity and actually has better resistance to different diseases,” explained Mohsen Yoosefzadeh Najafabadi, an associate professor with the Ontario Agricultural College. “We needed to find something to narrow it down from 100,000 potential combinations to 250. So we started to use AI to better analyze the data that we are collecting in the field.” Once digitized and combined with over 40 years of archived college data, Yoosefzadeh Najafabadi and his team merged the information with hundreds of thousands of published archived papers on dry beans to create Bean GPT. “We are the first breeding program that is using hybrid AI,” Yoosefzadeh Najafabadi said, adding that other programs more commonly use predictive AI in genomic selection. “I think there are a lot of different people with a lot of different opinions. But when it comes to science, I think it’s important that we’re using the technology that we have for things that will help people in the future,” said Trinity Solinger, an undergraduate student at the University of Guelph. The technology also helped with this year’s growing season that started in late May. It proved to be a difficult one due to fluctuating weather in southwestern Ontario. “We experienced a couple of thunderstorms that washed off some of the seeds. Then after that we experienced uneven heat waves, which was devastating, and then our beans were affected by the wildfire smoke and we experienced that for almost a week,” Yoosefzadeh Najafabadi said. The population that was AI-influenced was less affected by those weather events because they can better tolerate different climate variables and variations. “If [the smoke is] something that continues over the years, maybe that’s an avenue we actually have to start breeding for. Developing these lines that are more resistant to ozone damage,” said Lucas Rupert, another University of Guelph student. The super beans are set to hit the market in the next few years and those involved with the project believe it could revolutionize the growing market. “Every dollar invested in research translates into about $40 of commercial value for a farmer,” explained John Cranfield, dean of the Ontario Agricultural College. “For the sector, that’s really an impactful number, because that 40 to 1 return ratio is something that is almost unparalleled in terms of the scope and the size of that impact.”
