Exploiting Wild Relatives for Genomics-assisted Breeding of Perennial Crops

Abstract

Perennial crops are vital contributors to global food production and nutrition. However, the breeding of new perennial crops is an expensive and time-consuming process due to the large size and lengthy juvenile phase of many species. Genomics provides a valuable tool for improving the efficiency of breeding by allowing progeny possessing a trait of interest to be selected at the seed or seedling stage through marker-assisted selection (MAS). The benefits of MAS to a breeder are greatest when the targeted species takes a long time to reach maturity and is expensive to grow and maintain. Thus, MAS holds particular promise in perennials since they are often costly and time-consuming to grow to maturity and evaluate. Well-characterized germplasm that breeders can tap into for improving perennials is often limited in genetic diversity. Wild relatives are a largely untapped source of desirable traits including disease resistance, fruit quality, and rootstock characteristics. This review focuses on the use of genomics-assisted breeding in perennials, especially as it relates to the introgression of useful traits from wild relatives. The identification of genetic markers predictive of beneficial phenotypes derived from wild relatives is hampered by genomic tools designed for domesticated species that are often ill-suited for use in wild relatives. There is therefore an urgent need for better genomic resources from wild relatives. A further barrier to exploiting wild diversity through genomics is the phenotyping bottleneck: well-powered genetic mapping requires accurate and cost-effective characterization of large collections of diverse wild germplasm. While genomics will always be used in combination with traditional breeding methods, it is a powerful tool for accelerating the speed and reducing the costs of breeding while harvesting the potential of wild relatives for improving perennial crops.

Keywords: association mapping; crop wild relatives; genomics-assisted breeding; linkage mapping; marker-assisted selection; perennials.

FIGURE 1

Lack of differentiation between wild and domesticated perennial species. By plotting the two major axes of variation against each other (i.e., PC1 vs. PC2) we gain an overview of the genetic relatedness among samples. The primary wild ancestors and domesticated species cannot be clearly separated. Principal component analysis (PCA) was performed using single nucleotide polymorphism (SNP) data to compare primary progenitor species and cultivated accessions of grape (A) and apple (B). Cultivated accessions, as labeled by the USDA, are indicated in blue, while the primary progenitor species are indicated in orange. Equal sample sizes were used for both species and additional samples were projected onto the PCA axes.

FIGURE 2

Schematic of breeding using marker-assisted selection (MAS). Wild relatives containing a trait of interest are crossed with a cultivated crop. In this example, the wild parent is heterozygous for a dominant Mendelian trait. With a marker associated with this trait, offspring can be screened for the trait and eliminated at the seedling stage. MAS ensures that the trait of interest is present in the progeny through several generations of backcrossing. Not shown here is that, with each generation, there is an increase in the proportion of cultivated ancestry while maintaining the desirable wild trait.

FIGURE 3

Comparison of the effectiveness of genome-wide association studies (GWAS) and linkage mapping for mapping alleles of interest in wild relatives. When an allele of interest is found only in wild germplasm it co-segregates with population structure and cannot be mapped using GWAS. Linkage mapping provides a viable alternative for mapping traits in wild relatives. However, in the F1 generation, alleles homozygous for alternative states in the wild and cultivated parent will not segregate. Thus, a backcross, or pseudo-backcross, is required to map most alleles of interest.