Root phenotyping and plant breeding of crops for enhanced ecosystem services

Abstract

Diversifying and perennializing cropping systems can increase productivity while supporting ecosystem services such as soil protection, nutrient retention, and greenhouse gas mitigation. New crops can help achieve these goals, and advanced computational tools allow plant breeders to rapidly domesticate new crops and select for many traits that support both ecosystem services and profitable production. Intermediate wheatgrass [Thinopyrum intermedium (Host.) Barkworth. & D.R. Dewey; IWG] is a cool‐season perennial grass undergoing domestication to function as a perennial grain crop. Key aboveground domestication traits have been improved to support economically viable yields using genomic selection. However, few studies have quantified belowground traits despite their potential role in conferring ecosystem services. We present a platform for using minirhizotron cameras and machine learning software to analyze rhizotron images for inclusion in genomic selection models. The strength and direction of pairwise correlations between traits were variable with correlation coefficients (r) ranging from −0.27 to 0.99. Grain yield was positively, although weakly, correlated with total root length, area, and volume (r = 0.21, 0.21, and 0.19, respectively). Estimates of narrow sense heritabilities ranged from 0.41 to 0.76 for all traits and 0.46 to 0.66 for root traits. Root trait predictions using a genomic prediction model, measured by correlating model‐predicted values and field‐observed values, ranged from 0.08 to 0.23. Aboveground traits were better predicted (0.17 < r < 0.33). Simply selecting for aboveground traits could result in populations with desirable root traits, but our results demonstrate the potential for genomic selection to aid in advancing populations with specific root traits important for ecosystem services.