Fig. 3: Genetic clustering results (K = 2) from 21 de novo ddRADseq dataset assemblies for the corkscrew anemone Bartholomea annulata compiled using pyRAD v3.0.66 and Stacks v2.55.

Genetic clusters were detected using discriminant analysis of principal components (DAPC). Datasets varied by the clustering threshold parameter to assemble reads into loci (0.90 or 0.85 sequence similarity) and the maximum amount of missing data allowed before a locus could be incorporated into the final dataset (25 or 50%). Genetic cluster plots with the prefix “py” were produced in pyRAD and plots labeled with the prefix “st” were produced in Stacks. Similarity metrics for each assembly program and parameter setting were calculated by pong and represent the average pairwise percentage of individuals with identical cluster assignment probabilities across aposymbiotic, holobiont, and unmapped datasets.