RESUMEN
Leaf shape in plants plays important roles in water use, canopy structure, and physiological tolerances to abiotic stresses; all important traits for the future development and sustainability of grapevine cultivation. Historically, researchers have used ampelography, the study of leaf shape in grapevines, to differentiate Vitis species and cultivars based on finite leaf attributes. However, ampelographic measurements have limitations and new methods for quantifying shape are now available. We paired an analysis of finite trait attributes with a 17-point landmark survey and generalized Procrustes analysis (GPA) to reconstruct grapevine leaves digitally from five interspecific hybrid mapping families. Using the reconstructed leaves, we performed three types of quantitative trait loci (QTL) analyses to determine the genetic architecture that defines leaf shape. In the first analysis, we compared several important ampelographic measurements as finite trait QTL. In the second and third analyses, we identified significant shape variation via principal components analysis (PCA) and using a multivariate least squares interval mapping (MLSIM) approach. In total, we identified 271 significant QTL across the three measures of leaf shape and identified specific QTL hotspots in the grape genome which appear to drive major aspects of grapevine leaf shape.
RESUMEN
The genomics era brought unprecedented opportunities for genetic analysis of host resistance, but it came with the challenge that accurate and reproducible phenotypes are needed so that genomic results appropriately reflect biology. Phenotyping host resistance by natural infection in the field can produce variable results due to the uncontrolled environment, uneven distribution and genetics of the pathogen, and developmentally regulated resistance among other factors. To address these challenges, we developed highly controlled, standardized methodologies for phenotyping powdery mildew resistance in the context of a phenotyping center, receiving samples of up to 140 grapevine progeny per F1 family. We applied these methodologies to F1 families segregating for REN1- or REN2-mediated resistance and validated that some but not all bioassays identified the REN1 or REN2 locus. A point-intercept method (hyphal transects) to quantify colony density objectively at 8 or 9 days postinoculation proved to be the phenotypic response most reproducibly predicted by these resistance loci. Quantitative trait locus (QTL) mapping with genotyping-by-sequencing maps defined the REN1 and REN2 loci at relatively high resolution. In the reference PN40024 genome under each QTL, nucleotide-binding site-leucine-rich repeat candidate resistance genes were identified-one gene for REN1 and two genes for REN2. The methods described here for centralized resistance phenotyping and high-resolution genetic mapping can inform strategies for breeding resistance to powdery mildews and other pathogens on diverse, highly heterozygous hosts.
