Veronica (Pseudolysimachium)

Khan, Gulzar, Mayland-Quellhorst, Eike, Kosachev, Petr A., Mandáková, Terezie, Lysak, Martin A. & Albach, Dirk C., 2024, Altai Mountains - cradle of hybrids and introgressants: A case study in Veronica subg. Pseudolysimachium (Plantaginaceae), TAXON 73 (2), pp. 530-546 : 536-539

publication ID	https://doi.org/ 10.1002/tax.13176
DOI	https://doi.org/10.5281/zenodo.14107803
persistent identifier	https://treatment.plazi.org/id/03E0551F-FFCE-990A-31F5-F90AFDA8BBF0
treatment provided by	Felipe
scientific name	Veronica (Pseudolysimachium)
status

Treatment

■ RESULTS

Species boundaries and cohesion. — The assignment test in STRUCTURE with the dataset including all 233 individuals recovered K = 5 as optimal both by Evanno Δ K and likelihood scores (suppl. Fig. S3 View Fig ), which differs from the expected K = 10 based on species assignment ( Table 1 View Table 1 ).

The result of three different groupings, K = 4, 5, 6 recovered more or less highly genetically cohesive groups including Veronica longifolia , V. pinnata , V. porphyriana and V. spicata ; however, the individuals from V. incana , in K4 showed admixture of V. longifolia and V. porphyriana , but were recovered as a separate group in K5 and K6 with high genetic portion from V. porphyriana ( Fig. 1 View Fig ). Removing the putative hybrids and including only putative purebreds (174 individuals, ~75% of total dataset), the STRUCTURE result of K = 5 again showed more or less the same results ( Fig. 2B View Fig , suppl. Fig. S3 View Fig ).

The DAPC recovered 20 principal components as the best to group our data and clearly resolved all the individuals into K = 5 ( Fig. 2C View Fig ; suppl. Fig. S4 View Fig ). The DAPC results were supported by the unrooted trees inferred by IQ-TREE ( Fig. 2A View Fig ). The morphometric analysis (suppl. Appendix S1.) also supported the results of the genetic analyses in delimiting five main groups similar to those in Fig. 1 View Fig . Genetic differentiation among species based on AMOVA was F ST = 0.16 ( Table 2 View Table 2 ).

The hierarchical STRUCTURE analysis retrieved five cohesive groups including Veronica incana , V. longifolia , V. pinnata , V. spicata and V. porphyriana . In addition to these five groups, V. arenosa merged with V. pinnata , while V. sajanensis , and V. spuria with V. spicata . However, V. reverdattoi and V. taigischensis did not fit to any clear group (suppl. Table S2 View Table 2 ).

Chromosome localization of rDNA loci and classes of putative hybrids. — Chromosome numbers and rDNA loci were determined in six species. Veronica longifolia , V. pinnata , V. porphyriana , V. ×schmakovii , and V. spicata were found to have 2 n = 2 x = 34 chromosomes and one 5S rDNA-bearing chromosome pair, whereas V. incana had a tetraploid chromosome number (2 n = 4 x = 68) with a corresponding double number of 5S loci, i.e., two 5S-bearing chromosome pairs. In contrast to the conserved number of 5S rDNA, we found a large interspecific variation in the number of 35S rDNA: four NOR-bearing chromosomes in V. porphyriana , six in V. ×schmakovii and V. spicata , seven in V. pinnata , eight in V. longifolia , and at least nine in the tetraploid V. incana . In V. incana , additional very weak signals were observed ( Fig. 3 View Fig ). Each gDNA of V. longifolia and V. porphyriana hybridized to one half of the chromosome complement of V. ×schmakovii . Therefore, GISH strongly suggests a homoploid hybrid origin of V. ×schmakovii from parental genomes closely related to V. longifolia and V. porphyriana . This is also supported by the number of 35S rDNA-bearing chromosomes in V. ×schmakovii (six), an intermediate value between V. longifolia (eight) and V. porphyriana (four).

The hypothesis of hybridization between and among species of Veronica in the Altai Mountains based on their morphotypes was mostly supported by STRUCTURE, and NewHybrids ( Fig. 4 View Fig ; suppl. Table S3). The first-level STRUCTURE analysis is a bit more inconclusive in that respect showing admixture of various levels involving both the biparental hybrids and potential triparental hybrid ( Fig. 2 View Fig ). The maximum likelihood tree based on all individuals also revealed that the putative hybrid individuals are mostly sister to one of the suggested parents (suppl. Fig. S5 View Fig ). The genetically admixed individuals were mostly the ones identified as hybrids a priori by morphological characteristics (suppl. Fig. S1 View Fig ); however, we also recovered admixture in individuals considered to be pure species, e.g., V. sajanensis , V. spuria , V. taigischensis , and V. reverdattoi . The STRUCTURE results showed that in most cases the putative hybrid individuals in all five scenarios had posterior probabilities ranging between 0.15 <Q <0.50 or 0.50 <Q <0.85, which suggests a hybrid swarm from F1 to later-generation hybrids ( Fig. 4 View Fig ; suppl. Table S3). The analysis implemented in NewHybrids to group the hybrid individuals supported the results of STRUCTURE with the exception of scenario 5 and recovered hybrids of different classes but dominated by F1 (suppl. Table S3).

Demographic history of pure species. — Using G-PhoCS to model the associated demographic history in the species of Veronica ( Fig. 5 View Fig ), we retrieved divergence times between the major clades and branches to be recent (1.14 Mya to 1.57 Mya) at 95% HPD (highest posterior density; see suppl. Table S4 for complete information). Even with the upper HPD at 95%, the divergence time was in the range of 1.35 Mya to 1.84 Mya. Among the five species, the earliest recently diverged species are V. porphyriana and V. incana (1.14 Mya), while the most divergent species is V. pinnata (1.84 Mya) at upper 95% HPD. Similarly, the lowest effective population size is that of V. pinnata (4280 individuals) followed by V. porphyriana (6150 individuals), while the largest one belongs to V. spicata (21,000 individuals) at average 95% HPD. The effective population sizes of their ancestors were also different in comparison with the extant effective population sizes of all the five species, e.g., the smallest one of V. porphyriana and V. spicata ancestors (1160 individuals), whereas the largest one for the ancestor population of all species (16,100 individuals) ( Fig. 5 View Fig ; suppl. Table S4). Our modeling of gene flow in both forward and reverse directions showed high probability of frequent gene flow from V. longifolia and V. pinnata to V. incana , V. porphyriana and V. spicata . Nevertheless, from V. incana and V. porphyriana to the other three species, the probability of gene flow was also comparatively high. Similarly, the gene flow among V. incana , V. porphyriana and V. spicata was high in both forward and reverse directions ( Fig. 5 View Fig ; suppl. Table S4).