Affiches scientifiques

The tremendous throughput of modern DNA sequencing technology has allowed the resequencing of hundreds of soybean accessions, thus providing a catalogue of millions of single nucleotide polymorphisms (SNPs) and capturing causative mutations in many genes of interest. Such whole genome sequencing (WGS) is however too costly to perform routinely on a large number of lines as required in breeding work. In contrast, thanks to its low cost, high-throughput genotyping (via GBS or SNP arrays) has made it possible to genotype large numbers of individuals at thousands of SNPs. In this work, we wanted to test if these two technologies could be exploited jointly. We performed WGS on a set of 23 Canadian soybean lines. In parallel, two collections of soybean accessions were analyzed via GBS: a set of 136 Canadian lines and a set of 89 PI lines from diverse geographical origins. For each set, we used the WGS SNP data as a reference to impute the untyped SNP loci. To examine the quality and usefulness of this imputed dataset, we analyzed GmPhyA3, a gene for which several causative mutations have been reported to affect maturity. In both collections, GBS-derived SNPs failed to capture the causative mutations. In contrast, the imputation of the previously untyped causative mutation was found to be extremely accurate as long as a given causative mutation was present in the reference panel. This work shows that it is possible to obtain exhaustive SNP datasets using imputation on the basis of affordable SNP genotyping technologies.