Phenome-to-genome insights for evaluating root system architecture in field studies of maize

Understanding the genetic basis of root system architecture (RSA) in crops requires innovative approaches that enable both high-throughput and precise phenotyping in field conditions. In this study, we evaluated multiple phenotyping and analytical frameworks for quantifying RSA in mature, field-grown maize in three field experiments. We used forward and reverse genetic approaches to evaluate >1700 maize root crowns, including a diversity panel, a biparental mapping population, and maize mutant and wild-type alleles at two known RSA genes, DEEPER ROOTING 1 (DRO1) and Rootless1 (Rt1). We show the utility of increasing the dimensionality of traditional two-dimensional (2D) techniques, referred to as the "2D multi-view" method, to improve the capture of whole root system information for mapping genetic variation influencing RSA. Comparison of univariate and multivariate genome-wide association study (GWAS) approaches revealed that multivariate traits were effective at dissecting complex RSA phenotypes and identifying pleiotropic quantitative trait loci (QTLs). Overall, three-dimensional (3D) root models generated from X-ray computed tomography and digital phenotyping captured a larger proportion of RSA trait variations compared to other methods of root phenotyping, as evidenced by both genome-wide and single-gene analyses. Among the individual root traits, root pulling force emerged as a highly heritable estimate of RSA that identified the largest number of shared QTLs with 3D phenotypes. Our study shows that integrating complementary phenotyping technologies helps to provide a more comprehensive understanding of the genetic architecture of RSA in field-grown maize.