Low correlation between genomic and morphological introgression estimates in a walnut backcross

Abstract

A first backcross population of walnuts {[Juglans hindsii (Jeps.) Jeps. x Juglans regia L.] x J. regia} was used to evaluate the correlation between morphological (statistical) and genetic distance during introgression. Five traits based on leaf morphology were identified to quantitate the morphology of the parental species, their F1 hybrids, and the backcrosses to each parent. These traits were used to evaluate the morphological similarity of first backcrosses to J. regia using Mahalanobis' distance. The amount of genomic introgression of each backcross was estimated using 59 randomly amplified polymorphic DNA (RAPD) and 41 restriction fragment-length polymorphism (RFLP) genetic markers that identify polymorphisms between J. regia and J. hindsii. A smaller scaffold set of markers was also identified using published linkage data. The correlation between the measures of morphological and genomic introgression for the first backcrosses was low (0.23) but significant. The results suggest that selection based on morphology during backcrossing will not be an effective way to recover recurrent parent genome.