Roberto Fritsche-Neto poses with the drone and the sensor his lab uses to evaluate crops in the field. Photo by Kyle Peveto
Modern approaches to crop breeding will soon help the LSU AgCenter deliver new rice breeds in a much shorter time frame.
Last year, analytics expert Roberto Fritsche-Neto brought his knowledge of qualitative genetics to the AgCenter H. Rouse Caffey Rice Research Station and is now assisting rice breeders and plant pathologists in creating better rice varieties tailored to the needs of Louisiana producers.
His system of statistical analyses and computer modeling can knock years off the rice breeding timeline while also honing the characteristics of each variety to fit specific regions.
“You can reduce it by two or three years, but we can also increase the potential of each new variety,” Fritsche-Neto said. “We can select which one is the best one.”
Fritsche-Neto works in a field he likes to call prediction-based breeding that combines statistics, genetics and breeding and new technologies such as artificial intelligence. He assists multiple AgCenter scientists with their research.
“Sometimes you cannot directly see our work,” Fritsche-Neto said. “But we are partnering with other labs who are much nearer to the farmers.”
Fritsche-Neto joined the AgCenter in 2022 after working as a senior scientist in breeding analytics for a year and a half at the International Rice Research Institute in the Philippines. Previously, he was a professor of genetics and plant breeding at University of São Paulo in Brazil for seven years.
“In my work, we need to optimize breeding programs to reduce the time that we need to release a new variety,” he said. “So, how you speed up some selections is how you can speed up the cycle.”
Fritsche-Neto attempts to optimize the components of plant breeding, seeking the best “number of parents, the number of crosses per parent, the project size and how many individuals per cross we should evaluate.”
To do this, he follows two approaches. In one, Fritsche-Neto builds computer simulations of breeding programs that allow him to compare key performance indicators such as the response to selection and genetic gains. This allows the rice breeding team to “fine tune” the components of the breeding program, he said.
The second approach relies on analyzing historical data collected from rice breeding trials. These datasets include weather and soil information, molecular markers, and traditional phenotype, which are the observable characteristics of the varieties. This data is layered onto a statistical model that Fritsche-Neto can run to determine which potential new varieties would be the best for each trial location.
“We can then use the second approach to combine all layers of data and try to see how you can identify the best potential new rice varieties,” Fritsche-Neto said.
Fritsche-Neto is also using his analytical approach to evaluate the rice variety trials at locations across the state. He wants to ensure that the trials represent the varied soil types, weather, historical production numbers and other data points.
“The idea is not to reduce the number of trials, but to move one location or another location to cover some spots,” he said.
Also, Fritsche-Neto has assisted plant pathologist Felipe Dalla Lana on new ways to find sheath blight and blast in rice. They are using Fritsche-Neto's statistical analysis expertise to examine imagery data from drones to evaluate large numbers of plants. They plan to develop tools for the breeding program to select new resistant varieties and to help control the diseases in crops.
“You can set up a spray area where the farm can spray fungicide in only specific locations,” he said. “So, we can save money, we can save time and we can have a more sustainable crop production.”
Fritsche-Neto is also assisting rice breeder Adam Famoso with his breeding program. New lines that are developing with Fritsche-Neto's statistical approach are performing better than lines produced in the traditional manner, he said.
“There are many things running behind the scenes in the precision breeding program,” Fritsche-Neto said. “The results are evident; they are very clear.”