Genotyping-by-Sequencing in Switchgrass (Panicum virgatum): Method Development and use in Genetic Diversity Assessment

Genotyping-by-sequencing (GBS) has the potential to reduce the cost and improve the speed of genotyping projects. GBS works by end-sequencing of reduced representation libraries of multiple genotypes, which enables identification of single nucleotide polymorphisms (SNP) from aligned sequencing reads. However, genotyping the same loci among a number of individuals with sufficient depth to unequivocally make genotyping calls is a challenge in polyploid heterozygous genotypes. The objective of this study was to find a suitable enzyme and genotype multiplex level with which we can sequence an optimal number of loci with acceptable read depth. We tested two methods: (a) combination of FseI, an 8-bp recognition site and MspI, a 4-bp recognition site and (b) BsaXI, a type IIB enzyme. Using a panel of 22 diverse switchgrass genotypes and 22 progenies from a biparental population sequenced in a single flow cell on an Illumina Hi-Seq sequencer for the FseI-MspI method, we were able to identify ~20,000 SNPs, ~9,000 of which were genotyped in 75% of the genotypes. Similarly, using 73 diverse switchgrass genotypes in a single flow cell for BsaXI, about 50,000 SNPs were identified with about 7,000 present in 75% of genotypes.  We used these SNPs to characterize the genotypic diversity and population structure in switchgrass. The genotypes were broadly divided in two categories representing upland and lowland switchgrass. Furthermore, the genotypes were grouped according to their geographic origin. The sequencing results reflected results based on SSR markers but provided a better resolution of some relationships. Our results suggested that GBS can work well in switchgrass by using enzymes that reduce the number of sites to be sequenced, enabling genotypes to be called with high reliability.