Utilizing plant genetic resources for pre-breeding of water-efficient sorghum: genetics of the limited transpiration trait
Date
2022
Authors
Cerimele, Gina, author
Morris, Geoffrey, advisor
Cotrufo, Francesca, committee member
McKay, John, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Shifting precipitation patterns driven by the changing climate threaten productivity of dryland agricultural systems. Increasing the efficiency of water use by crops grown in dryland regions, such as sorghum (Sorghum bicolor), is a target for plant breeding to address this issue. c variants conferring efficient water use in sorghum may be found within collections of plant genetic resources (PGR). However, tropical sorghum PGR require adaptation to the target temperate environment to begin the pre-breeding trait discovery process. The landmark Sorghum Conversion Program unlocked diverse sorghum genetics for temperate breeding by adapting tropical African lines to temperate height and maturity standards. In the U.S. Sorghum Belt, spanning South Dakota to central Texas, the limited transpiration (LT) trait could provide growers a 5% yield increase in water-limited conditions with high vapor pressure deficit (VPD) according to crop modeling. To transfer the LT trait into commercial breeding programs, an elite donor line must be developed. Characterizing the genetic architecture of LT informs markers and breeding strategy for development of an elite donor. To characterize the genetic architecture of LT, two biparental recombinant inbred line (RIL) mapping families were developed from crossing putative LT parents SC979 and BTx2752 by putative non-LT parent RTx430. For this study, the families were grown together as a mapping population in three locations (continental-humid eastern Kansas, semi-arid western Kansas, and semi-arid Colorado) in one year. The families were phenotyped for the LT trait using UAS- collected thermal imaging and canopy temperature as a proxy. The families were initially designed with the goal of controlling phenotypic covariates of canopy temperature associated with height and flowering time, like neighbor-shading and artifactual temperature inflation related to panicle imaging. To test whether the family design controlled for height and flowering time covariates, the populations were phenotyped for both traits. High broad-sense heritability (H2) > 0.86 for all traits and families across locations indicates that the traits are not fixed. However, phenotypic distributions reveal that most lines are within an agronomically-relevant range that limits confounding covariates. Using DArTseq-LD genotyping data, GWAS analyses of height and flowering time reveal putatively significant marker-trait associations (MTA) with known loci underlying height and maturity in sorghum. These results collectively indicate that, while genetic variation for height and flowering exist in the LT mapping families, the resulting phenotypes are homogeneous enough to be suitable for LT genetic mapping. To test hypotheses on the monogenic, oligogenic, or polygenic architecture of the LT trait, canopy temperature data collected by the UAS-thermal imaging missions was used. Non-zero H2 of canopy temperature in most location-timepoints indicates genetic variation is present for LT in the population. Continuous phenotypic distributions imply a quantitative architecture. GWAS analyses revealed moderate marker-trait association peaks visible within timepoints and across locations, indicating oligogenic architecture of LT. Some of those peaks also colocalize with sorghum homologs of aquaporin genes in Arabidopsis thaliana, suggesting that aquaporin variation could be a molecular basis underlying the trait. These results provide a basis for marker-assisted selection in developing an LT donor line.