Assessment and Accessibility of Phenotypic and Genotypic Diversity of Carrot (Daucus carota L. var. sativus) Cultivars Commercially Available in the United States

Abstract

Increased use of intellectual property rights over plant germplasm has led to a complicated landscape for exchange among plant breeders. Our goal was to examine phenotypic and genotypic diversity present in commercially available carrot (Daucus carota L. var. sativus) germplasm in relation to the freedom to operate—the ability for plant breeders to access and use crop genetic diversity. A collection of 140 commercially available carrot cultivars were grown in replicated field trials in the Madison, WI area in 2013 and 2014. Phenotypic measurements were recorded for leaf and root characteristics. Illumina sequencing was used to conduct genotyping by sequencing analysis on all cultivars to understand the range of genetic diversity present. Additionally, the intellectual property rights associated with each cultivar was noted to determine the freedom to operate. We found that although one-third of the commercially available US carrot cultivars in our study are restricted through some form of intellectual property rights, the genetic and phenotypic variability of the protected cultivars does not represent a completely separate group from the available material. Phenotypic analyses including ANOVA and principal components analysis, suggest that many of the traits differed significantly based on market class, but not by whether the cultivar had freedom to operate. The principal components and Fst analyses on the genotyping by sequencing data revealed that carrot market classes (Fst = 0.065) and freedom to operate classes (Fst = 0.023) were not genetically distinct, and that principle components 1 and 2 account for only 10.1% of the total genotypic variation, implying that cultivated carrot germplasm in the US forms an unstructured population. Our findings suggest that the genetic diversity present in carrot cultivars that have freedom to operate is potentially large enough to support carrot breeding efforts in most market classes given present levels of intellectual property protection.

Comparison of Representative and Custom Methods of Generating Core Subsets of a Carrot Germplasm Collection

“…Among cultivated carrot accessions, there is only a weak genetic substructure (Luby et al, 2016;Ellison et al, 2018). In this study, we evaluated representative and custom methods of generating core sets of material using 433 accessions from the USDA National Plant Germplasm System carrot collection.…”

“…To develop the genomic breeding value (GBV) and genomic training population (GTP) cores a dataset of 145 commercially available carrot cultivars collected in 2013 (Luby et al, 2016) and 273 open-pollinated cultivars collected before 1985 (Theisen, 2016) was used as a training population for genomic prediction. Details regarding data collection and calculations of least squares mean phenotypes for these two collections can be found in their respective publications.…”

“…Passport and phenotypic data for all accessions used in this study can be found in Supplemental Table S1. The SNPS were filtered so that the final dataset had £10% missing data at each marker and a minor allele frequency >5%, leaving 19,944 SNPs in 749 genotypes, which included samples from the National Plant Germplasm System, commercially available carrot cultivars (Luby et al, 2016), and open-pollinated cultivars (Theisen, 2016). Individual samples of the same accession were merged before SNPs were called.…”

“…Using available phenotypic and genotypic information for collections of open-pollinated and commercially available cultivars (Luby et al, 2016;Theisen, 2016), canopy height and flavor were predicted in the 433 germplasm accessions. Using available phenotypic and genotypic information for collections of open-pollinated and commercially available cultivars (Luby et al, 2016;Theisen, 2016), canopy height and flavor were predicted in the 433 germplasm accessions.…”

“…In the absence of robust evaluation data for entries in a germplasm collection itself, existing data for traits of interest on related material can be used to estimate trait values in the collection based on the genetic relatedness between phenotyped and unphenotyped individuals. Using available phenotypic and genotypic information for collections of open-pollinated and commercially available cultivars (Luby et al, 2016;Theisen, 2016), canopy height and flavor were predicted in the 433 germplasm accessions. The additive relationship matrix was estimated as A = WW¢/c where W ik = X ik + (1 -2p k ), where X ik is the genotype for accession i at marker k (k = 1, …, N k , coded as {−1, 0, 1}), and p k is the frequency of the allele at marker k. The expectation maximization imputation algorithm was used to estimate missing markers (Poland et al, 2012).…”