Strong parallel evidence of selection during switchgrass sward establishment in hybrid and lowland ecotypes

Authors: TILHOU, NEAL - Former ARS Employee; KAEPPLER, SHAWN - University Of Wisconsin; CASLER, MICHAEL - Retired ARS Employee

Submitted to: Grassland Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/23/2022
Publication Date: 4/30/2022
Citation: Tilhou, N., Kaeppler, S.M., Casler, M.D. 2022. Strong parallel evidence of selection during switchgrass sward establishment in hybrid and lowland ecotypes. GRASSLAND RESEARCH. 1(1):31-42. https://doi.org/10.1002/glr2.12007.
DOI: https://doi.org/10.1002/glr2.12007

Interpretive Summary: Switchgrass is an important candidate crop for development of perennial and sustainable biomass production fields. Switchgrass production fields are normally planted with seed at a spacing of about 1-2 seeds per square inch. Once seedlings emerge from the soil, they face massive competitive pressures from their neighbors and from neighboring weed plants. Due to this competition, it has been estimated that over 90% of emerged seedlings will die within two years of the planting date. This study used whole-genome DNA sequencing to track the changes in genes associated with establishment mortality over the first three years from planting of 46 improved switchgrass lines. Highly consistent changes in discrete DNA regions were observed across most lines, strongly indicating that mortality is due, at least in part, to genetic selection. Predictive models indicated that surviving plants were more winter hardy, but not different in flowering time, than the average of the original plants. These results have significant implications for researchers and growers of all perennial grasses used for either biomass production or forage, because they indicate that significant and permanent genetic changes can occur after seeds are planted in the ground.

Technical Abstract: Switchgrass sward establishment results in up to 90% seedling mortality. To determine degree of selection during establishment, we sequenced pooled leaf samples from replicated swards of 46 half-sib families from two breeding groups (lowland and hybrid) prior to and through three years of stand establishment. We also predicted the genomewide polygenic impact of establishment selection on heading date and winter survivorship using independent training data. Tests using smoothing-spline defined windows of fixation indices (Fst) indicated strong and parallel selection in both breeding groups. Five windows exceeded a significant threshold of 99.9% in >10 families. The polygenic prediction model determined that establishment selection was at least partially associated with winter survivorship but resulted in variable heading date alterations. The genomewide variation is consistent with selection for a small number of related parental lines. Therefore, sward samples were used to train a genomic prediction model for establishment survival breeding value and rank genotypes from an independent trial. This model concluded that a small number of hybrid and coastal ecotype individuals are promising germplasm sources for improved sward survival. This study confirms prior reports of sward selection during grassland establishment and highlights the strength of pooled DNA sequencing for survival traits.