Although polyploidy is widespread across the plant Tree of Life, its long-term evolutionary significance is still poorly understood. Here we examine the effects of polyploidy in explaining the large-scale evolutionary patterns within angiosperms by focusing on a single family exhibiting extensive inter-specific variation in chromosome numbers. We inferred ploidy from haploid chromosome numbers for 80% of species in the most comprehensive species-level chronogram for the Brassicaceae. After evaluating a total of 94 phylogenetic models of diversification, we found that ploidy influences diversification rates across the Brassicaceae. We also found that despite diversifying at a similar rate to diploids, polyploids have played a significant role in driving present-day differences in species richness among clades. Overall, in addition to highlighting the complexity in the evolutionary consequences of polyploidy, our results suggest that rare successful polyploids persist while significantly contributing to the long-term evolution of clades. Our findings further indicate that polyploidy has played a major role in driving the long-term evolution of the Brassicaceae and highlight the potential of polyploidy in shaping present-day diversity patterns across the plant Tree of Life.

All references in accordance to the Main Text

Data S1. The Brassicaceae phylogeny was extracted using the extract.clade function in the ape R package version 5.2 [31] from Smith and Brown [22]. 

Data S2–S3: We used ChromEvol to infer ploidy for the maximum number of species included in our tree (1,667 species). First, we estimated the median chromosome number across all the available counts for each species (see also [34]). Second, we extracted all possible subtrees from the Brassicaceae phylogeny using the subtrees function implemented in the ape R package. We then selected subtrees with (i) sizes between 25 and 550 species, and (ii) a maximum of 70% of missing data (i.e., species in tree lacking chromosome counts). This produced 342 subtrees that we used to run the Ploidy Inference Pipeline (PIP) and infer ploidy. The newly developed R function used to partition the Brassicaceae phylogeny into optimal subtrees for ChromEvol analyses is provided in Data S3. 

Data S4: Brassicaceae phylogeny used to analyzed the phylogenetic signal of ploidal level variation across the Brassicaceae phylogeny using Pagel’s Lambda [38] and the D-statistics [39]. The same tree was used to fit Binary State Speciation and Extinction models (BiSSE [28,29]) and Hidden State Speciation and Extinction models (HiSSE [27]) using the diversitree R package version 0.9-10 [42] and hisse R package version 1.8.9 [27,43]. 

Data S5: We analyzed the role of polyploid species in influencing species richness among Brassicaceae genera using phylogenetic regressions [41] and phylogenetic path analyses [30,46]. We first constructed a generic-level phylogeny for the family by pruning from each genus in tree, all species except for one. The resulting generic-level tree (one species per genus) is provided in Data S5.

We recommed using R (https://www.r-project.org) and figtree (http://tree.bio.ed.ac.uk/software/figtree/) for accessing these files.