Staying on top of all the latest ag technology, field data and research can be overwhelming. That’s why the Soybean Research & Information Initiative, formerly the Plant Health Initiative, continually provides you with access to expert information and news about soybean pests, diseases, and agronomics. The aim of this check off-funded website is to communicate the on-going progress and current understanding coming from your active and wide-ranging U.S. soybean research programs. Please visit often!

RESEARCH HIGHLIGHTS

Back
Identifying High-Yield Genotypes in the USDA Soybean Germplasm Collection
Tue, Nov 1, 2016
Identifying High-Yield Genotypes in the USDA Soybean Germplasm  Collection
by George Graef, Soybean Breeder, University of Nebraska

The USDA Soybean Germplasm Collection contains over 21,000 accessions (plant or seed samples) including wild relatives, landraces, and soybean cultivars from around the world. The majority of unimproved accessions come from China, where soybean was domesticated, as well as Japan and Korea, other areas of ancient cultivation. 

Domestication of soybean has resulted in a loss of genetic diversity, with landraces retaining only about 63% of the diversity found in the wild Glycine soja. Furthermore, 86% of the parentage of U.S. commercial soybean cultivars released between 1947 and 1988 are accounted for by only 17 ancestral plant introductions.
 
Because the genetic diversity in current soybean cultivars is limited, we need to more effectively use the vast diversity that does exist. The goal of this three-year, multi-state project, funded by the North Central Soybean Research Program, is to determine how we can effectively and efficiently select from among the large number of accessions in the USDA Soybean Germplasm Collection and identify the best lines to use in our breeding programs.

It is not feasible to conduct extensive multi-location yield testing on every accession in the collection. Therefore, our approach is to use the soybean genotype database (SoySNP50K) which includes comprehensive DNA genotype information for every accession in the USDA collection. This database, recently constructed with funding by the United Soybean Board, is a tremendous resource that we can use to relate the genotype (genetic makeup) to the phenotype (the observable traits) of the plant.

With so many soybean accessions to choose from, how can we make good decisions on which ones to use to improve yield? That becomes a sampling question, and we employed some statistical methods to sample the collection in three different ways, using the 50K SNP genotype information to assess the genotypic diversity among accessions. Two different sampling methods were aimed at maximizing the diversity among the selections in the sample. The third method was a random sample for comparison. We developed three sets of soybean accessions selected from the USDA collection based on the three selection methods and conducted extensive multi-environment yield trials on those selections during 2015 and 2016.  We grew 505 total accessions of MGI-MGIV in NE, MN, IA, IL, MO and included a MG III and a MG IV location in Kansas in 2016.

Our 2015 multi-location yield tests provided very good data, especially considering the unadapted nature of most of the soybean lines we are testing, and the variation in environments for 2015 across the region. We are just completing harvest for the 2016 season, and in general conditions were excellent. We now have high-quality phenotype data for yield and other agronomic traits from 14 environments over 2 years. We can now relate the phenotype information (yield, plant height, maturity, lodging, seed weight, and seed composition) to the genotype information in the database. 

The amount of yield and agronomic data we now have exceeds the quality of any yield plot data currently available on this number of accessions in the collection. Through this coordinated effort, we have 28 replications worth of data from 14 environments in five major soybean producing states in the north-central region. That is valuable, and very important to get good estimates of yield and other traits across environments to relate to the genotype information and develop good prediction models.

The prediction models we develop based on this sample of genotype and phenotype information will be used to go back to the entire collection and resample new, untested soybean genotypes in MG I, II, III, and IV, that are predicted to have superior genotypes for improving yield. The next phase of the project will evaluate those reselected lines from the collection to validate the models and provide superior soybeans for use in public and commercial breeding programs. 

The final information from this project will be publically available, so both public and industry research and development programs will be able to use it. This offers the potential to greatly expand the diversity of the commercial soybean germplasm base, which may provide more resilience in varying climate conditions as well as allow for longer term gains in yield compared to those without more genetic diversity. 

To assist farmers in choosing from current commercial soybean varieties, unbiased comparisons of a large number of soybean varieties across many environments are conducted each year by land-grant universities and Extension programs. Variety trial results are generally available in November each year.
View all north-central University and Extension soybean variety trials>