Location: Crop Improvement and Genetics Research
Project Number: 2030-21000-023-00-D
Project Type: In-House Appropriated
Start Date: Oct 26, 2015
End Date: Mar 11, 2018
Switchgrass (Panicum virgatum L.) and big bluestem (Andropogon gerardii) have value in pure stands or in polycultures for bioenergy production in marginal environments. With recent advances in inexpensive genotyping methods based on next generation sequencing, breeding progress can be significantly improved in these two species through the greater awareness of kinship that dense genotyping can provide. The plant breeding program of ARS scientists in Lincoln, Nebraska, currently does not have bioinformatics or molecular support to take advantage of dense genotyping while scientists at the ARS, Western Regional Research Center in Albany, California, have developed this technology for switchgrass and are now working to improve it. However, the Albany group does not have the resources to apply the technology for breeding purposes on its own. The envisioned project would expand on existing collaboration between these two groups in the area of breeding and would focus on predicting performance, identifying Quantitative Trait Loci (QTLs), and providing estimates of relatedness among individuals. Objective 1. Implement a genotyping by sequencing (GBS) platform for switchgrass and big bluestem. Subobjective 1.A: Improve sample collection, library construction and sequencing procedures. Subobjective 1.B.1: Create a private SNP database with annotation and allow authorized access for use in statistical software packages as part of a larger data management plan. Subobjective 1.B.2: Establish a functional database based on existing small grains databases but modified for big bluestem and switchgrass. Objective 2. Collaborate and provide genotypic data to ARS Lincoln, Nebraska, forage breeders for more accurate control of breeding and selection in support of breeding objectives.
Objective 1 Hypothesis: 1.A.1 Genotyping procedures appropriate for pedigree reconstruction require fewer markers than those used for genomic selection. Hypothesis: 1.A.2. SNP calling biases can be detected in GBS data using different bioinformatic pipelines and the best available reference genome. Initial genotyping by sequencing work on big bluestem will focus on two populations with high biomass yields and will employ library construction techniques to reduce the portion of genome sequenced and to increase sequence coverage per individual locus. For switchgrass, we will genotype 14 full-sib families using similar sequencing approaches on 1120 individuals from tetraploid lowland by tetraploid upland parents that are also being phenotyped. By including reference genotypes, different bioinformatic approaches will be tested to determine those that work best with switchgrass and big bluestem in terms of accuracy of identifying variants and of discriminating allelic variation from repetitive sequences. In order to efficiently store and retrieve genotypic information from populations, we will create databases dedicated to switchgrass and big bluestem genetic and phenotypic varation based on those used by breeders of small grains. An alternate route to unravel the genomic complexity of big bluestem would be to build genetic resources such as mapping populations and apply the latest whole genome sequencing technology to assemble larger contiguous sequences that would be the basis of a draft genome sequence. Objective 2 Hypothesis: 2.A. QTL for seed dormancy, cell wall properties, winter hardiness and other traits can be detected in switchgrass by combined analysis of multiple families. Hypothesis: 2.B. Breeding values of parents can be estimated more accurately using genotypic data. This plan proposes to work collaboratively to genetically dissect yield potential, seed dormancy, cell wall quality, and winter hardiness traits. This will be achieved by analysis of SNP marker-trait associations. For big bluestem, molecular relatedness and pedigree reconstruction methods will be used for breeding value prediction using BLUP to exploit genetic correlation among relatives and account for genotype by environment interactions. All marker information will be included in genomic selections, which will be conducted using methods such as ridge regression to estimate genetic parameters. Molecular relatedness based on SNPs will be determined by the number of shared alleles across loci. We expect that insights gained from switchgrass QTL mapping can be applied to accelerate some aspects of big bluestem mapping, but until a genome sequence is obtained, genetic mapping in the latter will lag behind. If these approaches for big bluestem are not possible, a secondary goal will be to create a full-sib population for traditional linkage mapping. If we find that the populations we are using are too small and do not provide the statistical power necessary to identify marker-trait associations with confidence, we will plant remnant seed from some of the families to increase their sizes or focus instead on using the relationship matrix to predict breeding values.