Genome-wide association mapping and gene expression analysis identify OsCPS1 as a new candidate gene controlling early seedling length in rice

High seedling vigor can improve the ability to compete against weeds and flooding at the seedling stage and is essential for the direct seeding of rice. Early shoot length is an important performance index in seedling vigor evaluation. However, information on the identity of rice germplasm with high seedling vigor, and the genetic basis of seedling vigor are not well understood. In this study, we have conducted a genome-wide association study using 302 international diverse rice accessions from the Rice Diversity Panel 2. Six quantitative trait loci (QTLs) were found to associate with shoot length (SL). The locus qSL2 was further analyzed for candidate gene characterization. We identified OsCPS1, which encodes CDP synthase and functions in GA (Gibberellins) biosynthesis in rice, exhibits differential expression between long and short SL accessions. Using the Nipponbare genome sequence as the reference, we identified a 36 bp deletion in the 5' UTR of OsCPS1 in long SL accessions, which is absent in short SL accessions. GA content analysis showed that the levels of bioactive GA(1) and GA(4) are considerably higher in long SL accessions than in short SL accessions. Genome-wide gene expression analysis indicated the expression of some photosynthesis genes is higher in long SL accessions than in short SL accessions. In contrast, genes involved in ABA (Abscisic Acid)-activated signal pathway showed lower expression in long SL accessions. Population analysis across wild rice, indica and japonica, suggested that OsCPS1 may be under selection in japonica during domestication. The results suggest that OsCPS1 is a candidate gene underlying qSL2. These data provide a promising source for candidate genetic variation associated with seedling vigor, with practical applications in rice breeding.