Project leader: François Belzile
Sector: Agrifood
Budget: 2 984 240,00 $

Start date: 31 March 2026 End date: 31 March 2028

User: Prograin

With Canada’s seed market set to grow 60% in 10 years, soybean seed value should rise from 500M$US to 800M$US. Rising global demand of plant-based protein (>50% by 2030), places soybeans as a cornerstone of sustainable food systems. Canada, a top producer of high-quality soybeans, plays a key role in supplying international markets. A notable portion of this market is held by multinational seed companies (MSCs) that have successfully integrated the latest tools (genomics, phenomics and AI) in their breeding programs. These allow MSCs to expedite the development of superior varieties and gain market share. To stay competitive, it is critical for Prograin to level the playing field through successful adoption of such powerful technologies. The aim of this project is to improve the accuracy of decisions made at key early stages of a breeding program. A data-driven breeding strategy allows the early identification of superior lines based on their genetic makeup.

A first objective will be to implement and validate high-throughput phenotyping, using sensors mounted on drones, to rapidly capture data on thousands of small plots instead of relying on the work of ground crews. A second activity will center on the incorporation of both climate and soil data to our prediction models to more accurately capture the interaction that occur between the genetic makeup of a line and the environment in which it is grown. A third objective will focus on making use of the latest genomic and phenotyping tools to identify lines with superior disease resistance at both early and late stages. Finally, a fourth activity will enable us to develop the databases and analytical tools to harness of this information to allow breeders to make the best possible decisions as to which breeding lines are the most promising. Prograin and collaborators will provide genetic materials and perform field trials to feed the model-building process. Family-specific field data (drone flights) as well as weather/soil data will enrich our existing genomic prediction model with the help of AI.

All major methods needed for the success of this project exist, but a coordinated and collaborative approach is required to enable their implementation to enable data-driven decisions. The tools and technologies developed through this collaboration will be a valuable trade secret allowing Prograin to achieve increased rates of genetic gain in their breeding program by shortening the cycle, increasing selection intensity & accuracy. This will directly result in the more rapid and efficient development of commercially attractive varieties for Canadian growers. For example, we estimate we will be able to shorten a breeding cycle by ~25% and increase the rate of genetic gain by 50% using these tools. These advances will empower Prograin to continue producing high-yielding, disease resistant varieties whose seed will be sold on the Canadian market and earn us a growing share of new markets.