Principal Investigator: François Belzile
Research Institution: Laval University
Timeline: April 2013 – March 2018
- Develop a genomic selection model in soybean germplasm that can accurately infer field performance strictly through genetic information.
- Optimize high throughput methods to genotype 1,000 individual plants and to retain the individuals in each cross that have the highest predicted yield.
- Compare how the lines selected strictly on genotype compare with those selected by the breeder based on individual plant selections in field trials for two years to determine yield.
- Can help breeders more accurately predict the field performance of new soybean lines for yield and protein and oil content, based on their genetic makeup.
- The ability to improve the genetic characterization processes for large numbers of soybean plants may lead to the development and commercialization of superior soybean varieties.
One of the key challenges with plant breeding is identifying the best progeny obtained in a cross. In advanced generations, information from extensive field testing is available to provide a good basis for selection. But in early generations, breeders must make decisions based on the appearance of a single plant, and work with very limited information to make decisions about whether to keep or discard a line.
Recent technology advances in DNA sequencing and genotyping now make it possible to quickly and cost effectively examine genetic markers in individual plants. This technology is a radical innovation for how marker information, or genomic selection, can support breeding efforts. The premise with genomic selection is that the performance of a line can be predicted from genetic makeup only, when there is sufficient marker information and a good model linking genetic information with agronomic performance.
This project set out to test this premise using genomic selection by collaborating with three public soybean breeding programs in Guelph and Ottawa, ON and Beloeil, QC. This work has the potential to help breeders more accurately predict the field performance of a new soybean line for yield and protein and oil content, based on its genetic makeup. The ability to improve the genetic characterization process for large numbers of individual soybean plants may lead to commercialization of superior soybean varieties.
Genomic selection model:
Using field trial and genetic marker data from a total of 275 soybean lines, researchers built a genomic selection model to predict six traits solely on genotype – yield, maturity, height, seed weight, seed protein content and seed oil content.
More than 10,000 individual plants have been genotyped, from three different breeding programs. Individuals were selected in each cross with the highest predicted performance.
Progeny from the individual plants selected were field tested to determine their agronomic performance based on six traits – yield, maturity, height, seed weight, seed protein content and seed oil content. Comparisons were then made on the performance of lines selected only on genotype, compared with those selected by the breeder based on individual plant selection. Because only preliminary yield trials were completed by March 2018, this only allowed for a preliminary assessment of the two breeding approaches. However, in the context of the Genome Canada project described below, this work will be continued and allow for a more definitive assessment by the end of 2019.
One PhD candidate worked on this project, and a research assistant gained valuable experience in genomic selection.
This project demonstrated a great collaborative effort between plant breeders and genomicists, working together to develop a workflow to very rapidly produce predictions for soybean breeders.
Through this project, researchers have performed the most extensive characterization of a relevant collection of soybean lines in eastern Canada and have improved the efficiency of the genomic selection process at all stages. These improvements hold future promise for soybean growers to have access to new superior soybean varieties.
Researchers are waiting for final field results to demonstrate whether the genomic selection approach has potential as an alternative means to identify the most promising lines in a soybean breeding program.
The cost of gene testing and predicting varietal performance is about $15-$20, and a future opportunity would be to find ways to bring this price down significantly to make this technology more affordable.
Researchers were able to leverage this work into a much larger Genome Canada Soyagen project to expand the scope of this work into new areas. Specifically, the work was expanded to provide breeders with enhanced tools to develop adapted varieties for western Canada (early maturity) and to address Phytophthora root rot, soybean cyst nematode and white mold related issues as well. This additional funding has resulted in much broader outcomes of this genomic selection project.
While this project was anchored in eastern Canada, if it becomes a promising approach to soybean breeding, it could be initiated in western Canada with western germplasm.
External Funding Partners:
This project was research activity #9 within the Canadian Field Crop Genetics Improvement Cluster andwas a collaboration between the Canadian Field Crop Research Alliance (CFCRA) and Agriculture and Agri-Food Canada (AAFC).
Funding for this project was provided in part by Agriculture and Agri-Food Canada through the Growing Forward 2 (GF2) AgriInnovation Program and in part by the CFCRA. Grain Farmers of Ontario is a founding member of the CFCRA.
Project Related Publications:
Belzile, F. 2018. Canadian research consortium for next generation selection in soybean. Canadian Field Crop Genetics Improvement Cluster: Summary Report – Growing Forward 2 – AgriInnovation Program (2013-2018). http://www.fieldcropresearch.ca/uploads/2/4/1/2/24120027/2019_09_25_- _cfcra_gf2_cluster_-_summary_report_-_english.pdf”
Sonah H, O’Donoughue L, Cober E, Rajcan I, Belzile F. 2014. Identification of loci governing eight agronomic traits using a GBS-GWAS approach and validation by QTL mapping in soybean. Plant Biotechnology Journal 13:211-221.
Torkamaneh D, Belzile F. 2015. Scanning and Filling: Ultra-Dense SNP Genotyping Combining Genotyping-By-Sequencing, SNP Array and Whole-Genome Resequencing Data. PLoS ONE 10(7): e0131533.
Torkamaneh D, Laroche J, Bastien M, Abed A, Belzile F. 2016. Fast-GBS: a new pipeline for the efficient and highly accurate calling of SNPs from genotyping-by- sequencing data. BMC Bioinformatics. 18:5; DOI 10.1186/s12859-016-1431-9.
Torkamaneh D, Laroche J, Belzile F. 2016. Genome-Wide SNP Calling from Genotyping by Sequencing (GBS) Data: A comparison of seven pipelines and two sequencing technologies. PLoS ONE 11(8): e0161333.