Genome-wide association study and genomic selection for proteinogenic methionine in soybean seeds

Authors: SINGER, W - Virginia Tech; SHEA, Z - Virginia Tech; YU, D - Virginia Tech; HUANG, H - Virginia Tech; Mian, Rouf; ROSSO, M - Virginia Tech; Song, Qijian; ZHANG, BO - Virginia Tech

Submitted to: Frontiers in Plant Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/31/2022
Publication Date: 4/25/2022
Citation: Singer, W.M., Shea, Z., Yu, D., Huang, H., Mian, R.M., Rosso, M.L., Song, Q., Zhang, B. 2022. Genome-wide association study and genomic selection for proteinogenic methionine in soybean seeds. Frontiers in Plant Science. https://doi.org/10.3389/fpls.2022.859109.
DOI: https://doi.org/10.3389/fpls.2022.859109

Interpretive Summary: Soybean [Glycine max (L.) Merr.] meal is the main source of protein in poultry and livestock feeds worldwide. Soybean protein is also increasing its share as human food in recent years. However, low concentrations of an essential amino acid, methionine, limit the nutritional utility of soybean protein. The objectives of this study were to identify genomic associations and evaluate the potential for genomic selection (GS) for methionine content in soybean seeds. We performed a genome-wide association study (GWAS) that utilized 311 soybean accession from maturity groups IV and V grown in three locations in 2018 and 2019. A total of 35,570 single nucleotide polymorphisms (SNPs) were used to identify genomic associations with proteinogenic methionine content that was quantified by high-performance liquid chromatography (HPLC). Across four environments, 23 novel SNPs were identified as being associated with methionine content. The strongest associations were found on chromosomes 3, 8, and 16. Several gene models were recognized within proximity to these SNPs, such as a leucine-rich repeat protein kinase and a serine/threonine protein kinase. Identification of the linked SNPs should help soybean breeders to improve protein quality in soybean seed. GS was evaluated using k-fold cross validation within each environment with two SNP sets, the complete 35,570 set and a subset of 248 SNPs determined to be associated with methionine through GWAS. Average prediction accuracy (r2) was highest using the SNP subset ranging from 0.45-0.62, which was a significant improvement from the complete set accuracy that ranged from 0.03-0.27. The results indicate that methionine content of soybean may be improved via marker assisted breeding.

Technical Abstract: Soybean [Glycine max (L.) Merr.] seeds have an amino acid profile that provides excellent viability as a food and feed protein source. However, low concentrations of an essential amino acid, methionine, limit the nutritional utility of soybean protein. The objectives of this study were to identify genomic associations and evaluate the potential for genomic selection (GS) for methionine content in soybean seeds. We performed a genome-wide association study (GWAS) that utilized 311 soybean accession from maturity groups IV and V grown in three locations in 2018 and 2019. A total of 35,570 single nucleotide polymorphisms (SNPs) were used to identify genomic associations with proteinogenic methionine content that was quantified by high-performance liquid chromatography (HPLC). Across four environments, 23 novel SNPs were identified as being associated with methionine content. The strongest associations were found on chromosomes 3, 8, and 16. Several gene models were recognized within proximity to these SNPs, such as a leucine-rich repeat protein kinase and a serine/threonine protein kinase. Identification of the linked SNPs should help soybean breeders to improve protein quality in soybean seed. GS was evaluated using k-fold cross validation within each environment with two SNP sets, the complete 35,570 set and a subset of 248 SNPs determined to be associated with methionine through GWAS. Average prediction accuracy (r2) was highest using the SNP subset ranging from 0.45-0.62, which was a significant improvement from the complete set accuracy that ranged from 0.03-0.27. This indicates that GS utilizing a significant subset of SNPs may be a viable tool for soybean breeders seeking to improve methionine content.