Verbascum, L.

Alzahrani, Ali Mohammed, Brehm, Joana Magos, Ghazanfar, Shahina A. & Maxted, Nigel, 2024, DNA barcoding of the genus Verbascum (Scrophulariaceae) in the Arabian Peninsula, TAXON 73 (2), pp. 547-555 : 548-551

■ RESULTS

Only 236 out of 276 sequences of all 4 regions (ITS, matK, rbcL, trnL) were successfully completed in the DNA barcoding of Verbascum species from the Arabian Peninsula (Appendix 1). The combined nuclear (ITS) and chloroplastic (rbcL, matK, trnL) matrices consisted of 2564 characters, of which 375 (14.6%) were variable, and 342 (13.3%) were informative. Together, the three chloroplastic (matK, rbcL, trnL) matrices contained 2115 characters, 39 (1.8%) of which were variable, and 99 (4.6%) were informative. The ITS matrix contained 449 characters, of which 239 (53.2%) were variable, and 50 (11.1%) were informative ( Table 2 ).

There was discordance among BI and MP trees of each individual marker, which had less resolution and lower support values than those of the combined markers. The parsimony analysis of the combined data resulted in a strict consensus on the 10,000 most equally parsimonious trees, with a 939-tree length, a consistency index of 0.8807 and a retention index of 0.8724 ( Table 2 ). In a few cases, the bootstrap values obtained from the MP analyses were either unresolved or less resolved than the posterior probability values obtained from the Bayesian analyses. The Bayesian and the MP analyses of the combined chloroplast and ITS genes are provided in Fig. 1 . The phylogenetic trees from separate analyses are available in the supplementary material (suppl. Figs. S1–S 8).

The Bayesian and the MP analyses of the concatenated chloroplast and ITS genes resulted in the same topologies for the phylogenetic relationships among Verbascum species, and strongly supported the genus Verbascum as monophyletic including Rhabdotosperma (Bayesian posterior probability [PP] = 1/maximum parsimony bootstrap [PB] = 100; Fig. 1 ). In addition, the phylogenetic tree was divided into 2 major branches and 11 clades (indicated by nodes 1–2 and A–K, respectively).

In the first major branch, clades A to D are formed in a polytomy with endemic species that share certain characteristics, including clustered flowers and stellate hairs. They are found from northwest Saudi Arabia to southern Yemen. Clade A (PP = 0.98/PB = 56) comprises species found in the southwest region of the Arabian Peninsula. Clade B was supported by the Bayesian analysis (PP = 0.74) but unresolved by the MP analysis; this clade consists of a complex species with a high variation, found in the Asir Mountains of southwest Saudi Arabia and the southern region of Yemen. Clade C (PP = 0.99/PB = 85) consists of species found in variable habitats and with a wide distribution from western to northwest Saudi Arabia. Clade D was well supported by the Bayesian analysis (PP = 0.87) but not by the MP analysis; its species are endemic to the Hijaz Mountains.

Clade E was strongly supported by the Bayesian and MP analyses (PP = 1/PB = 100). It includes a species that has four stamens, solitary flowers and forked hairs (rarely stellate hairs) and is found from northwest Saudi Arabia to the eastern Mediterranean. Clade F (PP = 1/PB = 60) consists of species with set-apart distributions in the north, west and south regions of the Arabian Peninsula; this clade is sister to the remainder of major branch 1.

The Bayesian analysis strongly supported Clade G (PP = 1), whereas the MP analysis weakly supported it (PB = 62). This clade’ s species can be recognised by five stamens, often solitary or rarely dichasium inflorescence, and glandular or stellate hairs, and their distributions range from northwest Saudi Arabia to the eastern Mediterranean. Clade H (PP = 0.95/PB = 53) is sister to the previous clade; however, its species have four stamens, solitary flowers, dense glandular hairs above and stellate hairs below, and its species are endemic from west to northwest Saudi Arabia. Clade I (PP = 1/PB = 99) consists of species endemic to the foothills of the Hajar Mountains in Oman and the U.A.E., with five stamens, flower clusters, bracteoles and glandular-stellate hairs.

Clade J can be divided into two strongly supported subclades (J1 and J2); both Verbascum and Rhabdotosperma species within this clade have four stamens, solitary flowers and glandular hairs; however, their seeds have distinct appearances. Subclade J1 (PP = 1/PB = 100) consists of Rhabdotosperma species with longitudinally furrowed seeds, whereas subclade J2 (PP = 1/PB = 100) comprises Verbascum species with transversally elongated seeds; both subclades are endemic to the southwestern Arabian Peninsula.

Clade K’ s (PP = 1/PB = 100) species can be recognised by four stamens, solitary flowers and glandular-pubescent hairs; they are found in the foothills and mountains northeast of Oman’ s Hajar region. This clade is sister to all the other Verbascum and the Rhabdotosperma taxa on the second main branch.

Figures

Fig. 1. Bayesian majority-rule (50%) consensus tree of the combined chloroplast and ITS sequence data matrix. Support values on branches are Bayesian posterior probability/maximum parsimony bootstrap. Clades and species (including study samples) are colour-coded; their geographic distribution is indicated in Fig. 2.