Several CIMMYT breeders, pathologists, students from the Colegio de Postgraduado (Mexico), and Monsanto researchers from Argentina and the USA attended a statistical data analysis course on three-way interactions at El Batán 15-18 November. Analysis of plant performance in different locations is typically arranged in two-way tables of plant genotype and location. Taking this a step further, the workshop introduced participants to a statistical model and new computer software that can add another factor, such as the year, into the equation.
“The program will help us to determine a genotype’s stability across different environments and years,” says José Crossa, head of Biometrics and Statistics of the Crop Research Informatics Lab of CIMMYT and coordinator of the course, which was jointly funded by CIMMYT and Monsanto.
The technology, adapted for agricultural analysis by Dr. Mario Varela of the Department of Mathematics at the National Institute of Agriculture Science in La Havana, Cuba and the main instructor of the course, builds upon previous multiplicative models, such as the Additive Main effect and Multiplicative Interaction Model (AMMI) and the Sites Regression model (SREG). Course participants spent two days learning the statistical theory behind the program and the following two days testing real data in the 3-way software.
The software, though fully-functional, is still a work in progress. “The practical component isn’t commercial software, but rather software that simply exemplifies the theoretical part that was given in the course,” says Varela, who until recently was only using the program in his personal research. Expressed interest by Monsanto and CIMMYT researchers prompted Varela and Crossa to update the program and organize the 3-way course. Both have expressed plans to further improve the program.
“During the course, questions came up that will help us improve the software. The idea is not to commercialize it; but once it’s improved, it will be sent to those interested,” says Varela. “I think the proper dissemination of the course will depend on the participants who are going to transmit it to other colleagues.” Alberto Pepper, manager of development strategies for Monsanto- Argentina and course participant, said he gained valuable information that will benefit his company. “We’re going to be using this program with our databases that cover different years and trials, allowing us to run more efficient analysis. This is going to help us to better position our materials and identify in each zone the advantages our materials have in comparison to other materials, enabling us to better advise farmers.”
