Genome-wide association study of the seed transmission rate of soybean mosaic virus and associated traits using two diverse population panels

Qiong Liu, Houston A. Hobbs, Leslie L. Domier

Theoretical and Applied Genetics; December 2019, Volume 132, Issue 12, pp 3413–3424

Figure: Symptom of soybean mosaic virus disease

Key message

Genome-wide association analyses identified candidates for genes involved in restricting virus movement into embryonic tissues, suppressing virus-induced seed coat mottling and preserving yield in soybean plants infected with soybean mosaic virus.

Abstract

Soybean mosaic virus (SMV) causes significant reductions in soybean yield and seed quality. Because seedborne infections can serve as primary sources of inoculum for SMV infections, resistance to SMV seed transmission provides a means to limit the impacts of SMV. In this study, two diverse population panels, Pop1 and Pop2, composed of 409 and 199 soybean plant introductions, respectively, were evaluated for SMV seed transmission rate, seed coat mottling, and seed yield from SMV-infected plants. The phenotypic data and genotypic data from the SoySNP50K dataset were analyzed using GAPIT and rrBLUP. For SMV seed transmission rate, a single locus was identified on chromosome 9 in Pop1. For SMV-induced seed coat mottling, loci were identified on chromosome 9 in Pop1 and on chromosome 3 in Pop2. For seed yield from SMV-infected plants, a single locus was identified on chromosome 3 in Pop2 that was within the map interval of a previously described quantitative trait locus for seed number. The high linkage disequilibrium regions surrounding the markers on chromosomes 3 and 9 contained a predicted nonsense-mediated RNA decay gene, multiple pectin methylesterase inhibitor genes (involved in restricting virus movement), two chalcone synthase genes, and a homolog of the yeast Rtf1 gene (involved in RNA-mediated transcriptional gene silencing). The results of this study provided additional insight into the genetic architecture of these three important traits, suggested candidate genes for downstream functional validation, and suggested that genomic prediction would outperform marker-assisted selection for two of the four trait–marker associations.

See https://link.springer.com/article/10.1007/s00122-019-03434-w