10.1002/csc2.20333 Unimputed Variants

Sixty-five of the accessions included in this study had been genotyped as part of a project to characterize carrot germplasm and develop genetic resources for cultivar improvement.

These 65 accessions were genotyped using genotyping-by-sequencing (GBS), with methods previously described by Ellison et al. (2018). Single nucleotide polymorphism variants were identified via the TASSEL-GBS pipeline version 5.2.26 as described by Glaubitz et al. (2014) and filtered with vcftools (Danecek et al., 2011) to include only those markers with <0.50 missing data and a minor-allele frequency of >0.05, resulting in a set of 110,155* Single nucleotide polymorphisms available for initial genomic analyses. Missing genotype data was imputed with Beagle 4.1 (Browning & Browning, 2007) using default settings.

The GBS data was used to assess the relatedness of these accessions by discriminant analyses of principal components using the find.clusters function of the R package (R-core team, 2016) adegenet v2.1.3 (Jombart et al., 2008). Measurements of the fixation index (Fst) between groups of accessions was assessed with the R packages adegenet (Jombart et al., 2008) and hiefstat v0.04-30 (Goudet, 2005). The relatedness of the accessions was also assessed via kinship analysis completed and plotted using default settings in the R package GAPIT v3.0 (Wang et al., 2018) applying the VanRaden algorithm. The results from the kinship analysis were used to estimate narrow-sense heritability (h2) in the R package heritability v1.3 for a subset of the abiotic stress tolerance traits.

* Erratum: Described as 110,156 in the publication