Phylogenomics of the carrot genus (Daucus, Apiaceae)

Authors: ARBIZU, CARLOS - University Of Wisconsin; Ruess, Holly; Senalik, Douglas; Simon, Philipp; Spooner, David

Submitted to: American Journal of Botany
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/2/2014
Publication Date: 10/1/2014
Publication URL: https://handle.nal.usda.gov/10113/60041
Citation: Arbizu, C., Ruess, H., Senalik, D., Simon, P.W., Spooner, D.M. 2014. Phylogenomics of the carrot genus (Daucus, Apiaceae). American Journal of Botany. 101(10):1666-1685.

Interpretive Summary: Cultivated carrot is the most widely grown crop of the carrot family and is cultivated on 1.1 million hectares globally of which 35,000 hectares are US grown and worth $600M annually. Cultivated carrot was selected from the common weed, Queen Anne’s Lace, which is widely distributed across temperate regions of the globe. Like many vegetables, the time frame and geographic region(s) of the first cultivation of carrots are unclear. This study uses a very powerful technique for investigating crop origins, by the use of a large number of single-copy molecular DNA sequences; here with 94 sequence regions from the nucleus. In total, 111,113 DNA sequence data points were generated. Analysis of these data provide very firm conclusions indicating that cultivated carrot is related to a group of wild carrot all possessing the same number of chromosomes (18), unlike all other wild carrots that possess 20, 22, or 44 chromosomes. The results also concur with prior studies showing that the genus Daucus, as currently classified, needs to be reevaluated, because it includes some other genera in the carrot family. The species status of two well-recognized species, technically referred to as D. broteri and D. guttatus is unresolved, with the present data supporting three species. Our study is useful in documenting relationships that will be the basis of future studies using additional DNA markers and additional collections of Daucus and related genera. These studies will guide breeders in selecting appropriate wild species to be used in their breeding programs to improve cultivated carrot.

Technical Abstract: Molecular phylogenetics of genome-scale data sets (phylogenomics) often produces phylogenetic trees with unprecedented resolution. We here explore the utility of multiple nuclear orthologs for the taxonomic resolution of a wide variety of Daucus species and outgroups. We studied the phylogeny of 89 accessions of 13 species and two subspecies of Daucus, and an additional ten accessions of related genera (Ammi, Astrodaucus, Caucalis, Margotia, Oenanthe, Orlaya, Pseudorlaya, Torilis, Turgenia) with DNA sequences of 94 nuclear orthologs of average aligned length of 1180 bp, with a concatenated database length of 111,166 bp. We analyzed our data with maximum parsimony (MP), maximum likelihood (ML), and concordance analysis. Reiterative analyses examined data of both alleles using ambiguity codes or a single allele with the highest coverage, trimmed vs. untrimmed homopolymers, pure exonic vs. pure intronic data and the use of all 94 markers vs. reduced subset of markers. Our MP and ML trees are highly resolved, with 100% bootstrap support for most of the external and many of the internal clades. They grouped multiple accessions of many different species as monophyletic with strong support, but failed to support others such as 1) the subspecies of D. carota and D. capillifolius, 2) D. sahariensis and D. syrticus, and 3) D. broteri and D. guttatus. These taxa also are hard to distinguish with morphological characters. The single allele analysis gave slightly better topological resolution in some clades, and is a logically more defendable analysis. Trimming homopolymers provided potentially more reliable data, but was time consuming failed to increase taxonomic resolution. The exonic data, as expected, had fewer parsimony-informative characters in the intronic regions as a proportion of the total characters. Similar phylogenetic results of the dominant topology can be obtained with many fewer markers. Our research highlights some difficult species groups in Daucus, and discovered misidentifications in germplasm collections. It highlights a useful subset of markers and methodological approaches for future studies of a dominant topologies in Daucus, saving time and funds. The entire database is useful for future workers wishing to explore causes of discordance in Daucus.