Ploidy variation and hybridization
Using flow cytometry to estimate genome size on a subset of our samples, we identified extensive ploidy variation associated with the three genetically differentiated groups. Within P. vaginatum , the fine-textured ecotype had an average estimated genome size consistent with diploidy (1.01, when 1.0 is designated as the size of the diploidP. vaginatum reference genome) (Table 1). In contrast, the relative genome size estimates for coarse-textured samples were consistent with a mix of diploids (1.04) and triploids (1.56). ForPaspalum distichum , samples had higher average relative genome size estimates overall, consistent with predictions that this species is predominantly hexaploid and pentaploid (Echarte, Clausen, & Sala, 1992). Unlike P. vaginatum samples, P. distichum genome size estimates did not form obvious multiples of the P. vaginatumreference genome. Instead, the difference in genome size between putative pentaploids and hexaploids (2.25 and 2.69, respectively) suggests a P. distichum basal genome size that is ~11% smaller than P. vaginatum . When P. distichum relative genome sizes were recalculated assuming a basal genome size of 0.89 instead of 1.0 as the point of reference, the values closely match numerical expectations for pentaploid and hexaploid genomes (2.53 and 3.03, respectively) (Table 1). Together these results suggest that P. vaginatum is composed primarily of diploid and triploid genotypes, that P. distichum is largely composed of pentaploid and hexaploid genotypes, and that the basal genome size of the latter species is likely ~11% smaller than theP. vaginatum reference genome.
By using genome-wide heterozygosity calls from the SNP dataset (F-values generated by plink) in conjunction with the flow cytometry results above, we were able to estimate ploidy for all genotypes (see Gompert & Mock, 2017). Specifically, individuals containing multiple subgenomes (e.g. allopolyploids) are expected to show an increase in heterozygosity with each additional subgenome. The highest genome-wide heterozygosity values (corresponding to more negative F-values) occurred in P. distichum and were consistent with a combination of allohexaploid and allopentaploid genotypes (average F = -1.323 and -1.241 respectively) (Table 1). Conversely, the fine-textured cluster of P. vaginatum(composed entirely of diploids) had the lowest genome-wide heterozygosity calls (average F = 0.541). For the coarse-texturedP. vaginatum group, which flow cytometry data indicated to be a combination of diploids and triploids, F-values were consistent with a slight increase in heterozygosity in putative triploids relative to diploids (average F = -0.520 and -0.279 respectively). Unexpectedly, there was a large difference in F-value between diploid coarse- and fine-textured accessions (-0.278 and 0.541); the greater heterozygosity in the coarse-textured genotypes, despite a lack of difference in genome size, is potentially consistent with hybridization in the origin of the diploid coarse-textured ecotype. Collectively these findings support the inference that P. vaginatum not only falls into two genetically distinct subgroups according to ecotype, but that the coarse-textured ecotype can be further subdivided by ploidy variation.