AJBM Anales del Jardín Botánico de Madrid Anal. Jard. Bot. Madr. 0211-1322 1988-3196 Consejo Superior de Investigaciones Científicas ajbm.2630 10.3989/ajbm.2630 Artículos Supermatrix analyses reveal the importance of outgroup, gene and taxon sampling in Onosma (Boraginaceae) phylogenetics Análisis de supermatrices demuestran la importancia del muestreo de grupos externos, genes y taxones en la filogenética de Onosma (Boraginaceae) https://orcid.org/0000-0002-2095-3816 Aygören Uluer Deniz d.aygoren@ahievran.edu.tr Department of Plant and Animal Production, Ahi Evran University, Kirşehir, Turkey Fuertes-Aguilar Javier 25 05 2023 12 2023 80 1 e133 17 01 2022 23 02 2023 25 05 2023 14 06 2023 © 2023 CSIC 2023 This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) License. Abstract

Tribe Lithospermeae (Boraginaceae) consists of ca. 26 genera and 470 species, in which Onosma constitutes approximately one third of the species (~150). Although the tribe is strongly supported as monophyletic, both generic and species boundaries remain ambiguous. Among them, not only the phylogenetic position of Eastern Asian Onosma species, but also the taxonomic limits of the genus remain unclear. Whether Eastern Asian Onosma is monophyletic, or the genus should be widened to include Maharanga, and maybe Cystostemon, are still open questions. For these reasons, I performed 16 phylogenetic analyses with different taxon coverages, alignments, gene regions and outgroups, with up to 746 taxa of tribe Lithospermeae and with five DNA regions, using data from GenBank. The results, with the widest taxon coverage to date, show that while genus Onosma is not monophyletic in any of the analyses, the phylogenetic relationships among Onosma s.s., Eastern Asian Onosma, Maharanga and Cystostemon differ among analyses. However, the approximately unbiased (AU) test showed that the topology (((Eastern Asian Onosma+Maharanga) Cystostemon) Onosma s.s.) is overwhelmingly supported. Therefore, the current study highlights the importance of taxon, gene and outgroup sampling in Onosma phylogenetics.

 Resumen

La tribu Lithospermeae (Boraginaceae) consta de ca. 26 géneros y 470 especies, en las que Onosma constituye aproximadamente un tercio de las especies (~150). Si bien la tribu ha sido fuertemente apoyada como monofilética, los límites taxonómicos a nivel de género y especie son todavía ambiguos. Entre ellos, no solo la posición filogenética de las especies de Onosma de Asia oriental, sino también la delimitación del género son confusas. Si Onosma de Asia oriental es monofilético o si el género debe ampliarse para incluir a Maharanga y tal vez Cystostemon son todavía cuestiones por resolver. Por estas razones, realicé 16 análisis filogenéticos con diferentes coberturas de taxones, alineamientos, regiones de genes y grupos externos, con hasta 746 taxones de la tribu Lithospermeae y con cinco regiones de ADN, usando datos de GenBank. Los resultados, con la cobertura taxonómica más amplia hasta la fecha, demuestran que si bien el género Onosma no resultó monofilético en ninguno de los análisis, las relaciones filogenéticas entre Onosma s.s., Onosma de Asia oriental, Maharanga y Cystostemon difieren entre análisis. Sin embargo, el test AU (“approximately unbiased”) mostró que la topología (((Onosma-Asia Oriental+Maharanga) Cystostemon) Onosma s.s.) es ampliamente compatible. Por lo tanto, este trabajo destaca la importancia del muestreo de taxones, genes y grupos externos en la filogenética de Onosma.

 Cystostemon Lithospermeae Maharanga Onosma outgroup phylogeny taxon/gene sampling Cystostemon Lithospermeae filogenia grupo externo Maharanga muestreo de taxones/genes Onosma Introduction

The family Boraginaceae Juss. (Boraginales) is subdivided in five subfamilies, including subfamily Boraginoideae Arn. with two tribes, namely Boragineae and Lithospermeae (Chacón & al. 2019). Tribe Lithospermeae with ca. 26 genera and 470 species (Chacón & al. 2019) is characterized by several morphological traits: herbaceous habit, sympetalous corolla with generally basal and faucal scales, fruit with four nutlets with basal attachments (Cohen 2014; Weigend & al. 2016). While Lithospermeae has been strongly supported as monophyletic (Thomas & al. 2008; Cohen & Davis 2009; Cohen 2014; Selvi & al. 2017), within the tribe both generic and species boundaries are ambiguous, and the phylogenetic relationships differ among studies (Weigend & al. 2009; Cohen 2011; Cohen 2014; Chacón & al. 2019).

Onosma L. is the largest genus of Lithospermeae with ~150 species (Chacón & al. 2017). It is distributed in north-western Africa, Europe and Asia, but the centre of diversity is in the dry, rocky and sunny habitats of Turkey and Iran (Mehrabian & al. 2011; Nasrollahi & al. 2019). While Onosma has been considered as a taxonomically difficult group; only a few morphological and molecular studies have focussed on this genus (e.g., Cecchi & al. 2011; Nasrollahi & al. 2019). Similarly, to date, its sampling in the previous studies was limited [Table 1]. For instance, while Nasrollahi & al. (2019) sampled 122 Onosma samples (87 species), they included only five outgroups in their analyses. Similarly, while Chacón & al. (2019) included 257 Lithospermeae samples (180 species and 13 subspecies), their sampling of Onosma remained insufficient with only 64 newly sequenced samples (50 species and 4 subspecies). Furthermore, some recently described taxa (e.g., Onosma atila-ocakii O. Koyuncu & Yaylaci; Koyuncu & al. 2013) have never been included in a phylogenetic study, and their phylogenetic position remain unknown. In addition, it was reported that genus Onosma is not monophyletic, with the Eastern Asian genus Maharanga recovered as closely related to Eastern Asian Onosma, namely, O. rostellata Lehm., O. paniculata Bureau & Franch., O. hookeri C.B. Clarke, O. waltonii Duthie, O. sinicum and O. pyramidalis Hook.f. (Cecchi & al. 2011; Chacón & al. 2019; Nasrollahi & al. 2019), so not only the phylogenetic position of the Eastern Asian species, but also the taxonomic limits of the genus Onosma (i.e., whether Eastern Asian Onosma+Maharanga constitutes a different genus) remain unclear.

Comparison of taxon sampling of the present study and the previous studies on genus <italic>Onosma</italic> and/or tribe Lithospermeae. The highest taxon coverage is indicated in bold. EA: Eastern Asian.
Cohen (2011) Cohen (2014) Chacón & al. (2019) Nasrollahi & al. (2019) Current study ITS Current study Total evidence
Total taxa 60 258 122 746 350
EA Onosma 5 8 21 12
Aegonychon 2 3 1
Alkanna 3 28 1 37 30
Ancistrocarya 1
Arnebia 1 2 9 39 16
Buglossoides 4 4 4 50 20
Cerinthe 1 2 6 27 9
Cystostemon 1 3 4 4
Echium 1 10 55 2 118 58
Echiostachys 1 3 3 2
Glandora 2 3 7 10 6
Halacsya 1 1 1 5 1
Huyhnia 1 2 1 3
Lithodora 2 2 4 25 16
Lithospermum 37 9 44 91 60
Lobostemon 2 3 7 5
Macrotomia 1 1 1
Maharanga 1 1 2 1 2 2
Mairetis 1 1 2 4 3
Megacaryon 2 3 2
Moltkia 1 5 6 16 9
Moltkiopsis 1 1 7 4
Neatostema 1 1 2 4 3
Onosma 1 6 64 117 278 88
Paramoltkia 1 1 1 5 2
Podonosma 1 1 2 1 3 2
Pontechium 1 2 3 3
Stenosolenium 1 1 1

For these reasons, molecular analyses of Onosma with a wide taxon coverage and more genetic regions are needed. However, such a worldwide study would be both time and money consuming. Therefore, before attempting such an expensive job, a supermatrix analysis which covers more taxa and all available published sequences would be the first step to overcome these problems. With this aim, in the current study, I performed 16 phylogenetic analyses with different taxon coverage, gene regions and outgroup(s), with up to 746 taxa and four plastid and one nuclear DNA regions, namely, rps16, trnS-G spacer, trnL gene, trnL-F spacer and nuclear ribosomal DNA (nrDNA) internal transcribed spacer (ITS). I also performed approximately unbiased (AU) (Shimodaira & Hasegawa 1999) test to evaluate the p-values for the phylogenetic relationships among Onosma, Maharanga and Cystostemon.

Material and methods Taxon sampling, alignment and phylogenetic analyses

The present study includes up to 746 taxa from 24 genera of tribe Lithospermeae (Table 1). Five loci, namely, nuclear ITS, trnL gene, trnL-F spacer, trnS-G spacer and rps16 downloaded from GenBank (Appendices 1 and 2). The number of taxa, alignment length, number of variable characters, and the number of parsimony informative sites for each region are provided in Table 2.

Total number of taxa, alignment length, total variable characters, and the number of parsimony informative sites for each region sampled in the present study.
Region Number of taxa Alignment length (bp) Total variable characters Parsimony informative characters (%)
ITS (1) 746 814 611 510 (62.7)
ITS (2) 555 810 555 489 (60.4
ITS (3) 323 802 525 447 (55.)
Plastid 339 2,753 1,073 662 (24)
Total evidence (1) 350 3,555 1,598 1,109 (31.2)
Total evidence (2) 345 3,555 1,579 1,101 (31)
Total evidence with distant OG (Vahlia) 347 3,611 1,750 1,321 (36.6)
Total evidence with closely related OG (Anchusa) 347 3,629 1,585 1,138 (31.4)

First, for the ITS region, three data matrices with different taxon sampling were created: dataset 1) ITS matrix with missing data (i.e., either ITS1 or ITS2 regions, or both available for each taxon) with 746 taxa; dataset 2) ITS data matrix without missing data with 555 taxa (i.e., both the ITS1 and ITS2 regions are available for each taxon); dataset 3) ITS data matrix without missing data with 323 taxa (i.e., for this dataset, the difficult taxa to align, possibly due to misidentifications and/or ITS region problems, were removed). Second, for the plastid data only one data matrix was created: dataset 4) trnL gene, trnL-F spacer, rps16 and trnS-G spacer were included for 339 taxa. Third, for the total evidence analyses (i.e., all five concatenated loci), four matrices were created: dataset 5) total evidence matrix with 350 taxa; dataset 6) total evidence matrix with 345 taxa [five taxa with only one sequence for each region, namely Halacsya Dörfl., Macrotomia DC. ex Meisn., Stenosolenium Turcz., Lithospermum tschimganicum B.Fedtsch (= Ulugbekia/Arnebia tschimaganica) and Aegonychon Gray were excluded]. With the addition of two different outgroup taxa (i.e., out of tribe Lithospermeae), namely Vahlia Thunb. and Anchusa L., two more datasets were created: dataset 7) total evidence matrix with distant outgroups, namely two Vahlia taxa, and 345 ingroup taxa; and finally dataset 8) total evidence matrix with closely related outgroups, namely two Anchusa taxa, and 345 ingroup taxa (Table 2). Other than these last two datasets (7 and 8), no outgroup(s) out of tribe Lithospermeae were used. In terms of taxon coverage for each dataset, see Table 3. These strategies were used to detect the effect of outgroups (e.g., datasets 5, 6, 7 and 8), different alignments (e.g., dataset 1, 6 and 3), genes (e.g., datasets 3, 4 and 6) and missing data (e.g., datasets 1 and 2) on the phylogenetic relationships within Onosma (Smith 1994; Qiu & al. 2001). Sequences were aligned by Geneious Pro 4.8.4 (Kearse & al. 2012) and manually edited for 16 different Maximum likelihood (ML) and Bayesian Inference (BI) analyses (Table 2). ML analyses were conducted with RAxML version 8.2.12 (Stamatakis & al. 2014) on the CIPRES Science Gateway (Miller & al. 2011). The GTRGAMMA model was selected as the best-fit model by using the Akaike information criterion (AIC) in the software jModelTest2.1.10 (Guindon & Gascuel 2013; Darriba & al. 2012), and the “Let RAxML halt bootstrapping automatically” options selected. A rapid bootstrap analysis/search for best-scoring ML tree was performed. BI analyses were implemented using MrBayes 3.1.2 (Ronquist & Huelsenbeck 2003) on the CIPRES Science Gateway (Miller & al. 2011), for 10 million generations, sampling every 100 generations, with a random starting tree. The first 25% of trees were discarded as “Burn-in” and the remaining trees were used to build the consensus tree. Tracer v1.5 (Rambaut & Drummond 2007 was used to inspect the MCMC output and to determine convergence of the two chains. iTOL (Interactive Tree of Life) (Letunic & Bork 2016) was used to create the tree image.

Results of the phylogenetic analyses with different genetic regions and different taxon sampling. Both BS (%) and PP (0-1) support values are given for each clade. The difference between the taxon sampling of different analyses is explained in Materials and Methods. Taxon numbers are indicated within parentheses. Empty cells indicate only one sequence was sampled. C-OG, close outgroup; D-OG, distant outgroup; X, the taxon was not recovered as monophyletic. Tree topology: T1, (((Eastern Asian <italic>Onosma</italic>+<italic>Maharanga</italic>) <italic>Cystostemon</italic>) distantly related to <italic>Onosma</italic> s.s); T2, (((Eastern Asian <italic>Onosma</italic>+<italic>Maharanga</italic>) <italic>Onosma</italic> s.s) <italic>Cystostemon</italic>)); T3, (((Eastern Asian <italic>Onosma</italic>+<italic>Maharanga</italic>) <italic>Cystostemon</italic>) <italic>Onosma</italic> s.s).
Clade name ITS (323) ITS (555) ITS (746) Plastid (339) Total evidence (350) Total evidence (345) Total evidence, D-OG (347) Total evidence, C-OG (347)
Tree topology X T1 T1 X T2 T3 T3 T3
Onosma s.s. 100; 1.00 100; 1.00 100; 1.00 X 100; 1.00 100; 1.00 93; 1.00 100; 1.00
E Asian Onosma+ Maharanga 100; 1.00 100; 1.00 100; 1.00 80; 1.00 98; 1.00 97; 1.00 94; 0.96 96; 1.00
Maharanga 98; 1.00 98; 1.00 98; 1.00 98; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Asian Onosma 91; 1.00 87; 0.98 90; 0.99 X X X X X
Alkanna 84; 1.00 80; 0.95 80; 0.94 87/ 0.80 99; 1.00 99; 1.00 99; 1.00 99; 1.00
Arnebia (excluding A. hispidissima) X X X X X 87; 1.00 92; 1.00 87; 0.99
Buglossoides X X X X X 97; 1.00 87; 0.96 96; 1.00
Cerinthe 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Cystostemon 92; 1.00 92; 0.98 92 0.97 98; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Echiostachys 98; 1.00 95; 1.00 95; 0.99 62; 0.80 99; 1.00 98; 1.00 97; 1.00 98; 1.00
Echium 99; 1.00 91; 0.99 91; 0.99 X 92; 0.99 89; 0.99 92; 1.00 90; 0.97
Halacsya 100; 1.00 100; 1.00
Huynhia 85; 0.99 89; 1.00 89; 1.00 87; 0.99 86; 0.95
Lithodora I 99; 1.00 98; 1.00 99; 1.00 X 95; 1.00 100; 1.00 100; 1.00 100; 1.00
Lithodora II (including Glandora) 100; 1.00 79; 0.96 75; 0.95 X 100; 1.00 96; 1.00 93; 1.00 94; 1.00
Lithospermum 97; 1.00 93; 0.99 98; 1.00 X 100; 1.00 100; 1.00 99; 1.00 100; 1.00
Lobostemon (excluding L. trigonus) 100; 1.00 99; 1.00 98; 1.00 X 97; 1.00 97; 1.00 85; 1.00 97; 1.00
Mairetis 100; 1.00 100; 1.00 100; 1.00 98; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Megacaryon 100; 1.00 100; 1.00 100; 1.00 92; 0.86 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Moltkia 100; 1.00 100; 1.00 100; 1.00 63; 0.72 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Moltkiopsis 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Neatostema 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Paramoltkia 100; 1.00 100; 1.00 100; 1.00 X 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Podonosma 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Pontechium 100; 1.00 100; 1.00 100; 1.00 96; 1.00 100; 1.00 100; 1.00 100; 1.00 100; 1.00
Alternative topology testing

The approximately unbiased (AU) (Shimodaira & Hasegawa 1999) test was used to evaluate the p-values for the possible (((Eastern Asian Onosma+Maharanga) Onosma s.s.) Cystostemon) and (((Eastern Asian Onosma+Maharanga) Cystostemon) Onosma s.s.) topologies resulted from the phylogenetic analyses by W-IQ-TREE (http://iqtree.cibiv.univie.ac.at/, Trifinopoulos & al. 2016), and by using the ‘-au’ option and 10,000 bootstrap replicates with the “total evidence dataset with 350 taxa”. Note that due to the possible problems with the ITS region (explained below), the AU test was performed just for the two possible topologies, indicated in Figures 2b and 2c.

Results

The genus Onosma and its subsections, namely, Onosma, Haplotricha and Heterotricha Riedl (1967) did not resulte monophyletic in any of the analyses (Table 3, Fig. 1) (Appendix 3-10). In contrast, Onosma s.s. was monophyletic (93-100% BS, 0.99-1.00 PP) in all analyses, except the plastid matrix analysis, where the clade was not well-resolved

Summary of the phylogenetic tree from ML analysis of <italic>rps</italic>16+<italic>trn</italic>S-G+<italic>trn</italic>L+<italic>trn</italic>L-F+ITS dataset showing the phylogenetic relationships within tribe Lithospermeae. Bootstrap support values and posterior probabilities are indicated above branches. Outgroups are indicated, Lithospermeae genera are colour coded.

The topologies describing the phylogenetic relationships among Eastern Asian genera Onosma, Maharanga, Onosma s.s and Cystostemon differed depending on the analysis (Table 3, Fig. 2). First, Onosma s.s. was not sister to the Eastern Asian Onosma+Maharanga clade in neither the ITS with 555 taxa nor the ITS with 746 taxa analyses (Fig. 2a) although note that the relationships among Onosma s.s., Eastern Asian Onosma, Maharanga and Cystostemon were not resolved in the ITS with 333 taxa and plastid data analyses. Second, the “total evidence tree with 350 taxa” analysis yielded a (((Eastern Asian Onosma+Maharanga) Onosma s.s.) Cystostemon) topology (Fig. 2b). However, a possible (((Eastern Asian Onosma+Maharanga) Cystostemon) Onosma s.s.) relationship was recovered for the first time here, and apart than the distant and closely related outgroup analyses, this topology was also recovered from the “total evidence tree with 345 taxa” analysis (Fig. 2c).

Phylogenetic relationships within the <italic>Onosma</italic> s.s.+Eastern Asian <italic>Onosma</italic> +<italic>Maharanga</italic>+ <italic>Cystostemon</italic> clade with different gene and different taxon sampling analyses. OG: outgroup, total evidence tree: <italic>rps</italic>16+<italic>trn</italic>S-G+<italic>trn</italic>L+<italic>trn</italic>L-F+ITS; <bold>a,</bold> ((Eastern Asian <italic>Onosma</italic>+<italic>Maharanga</italic>) <italic>Cystostemon</italic>) topology (in which <italic>Onosma</italic> s.s. was not sister to this clade) recovered from two analyses, namely, ITS with 55 taxa and ITS with 746 taxa; <bold>b,</bold> (((Eastern Asian <italic>Onosma</italic>+<italic>Maharanga</italic>) <italic>Onosma</italic> s.s.) <italic>Cystostemon</italic>) topology recovered from the total evidence analysis with 350 taxa; <bold>c,</bold> (((Eastern Asian <italic>Onosma</italic>+<italic>Maharanga</italic>) <italic>Cystostemon</italic>) <italic>Onosma</italic> s.s.) topology recovered from three analyses, namely, the total evidence analyses with 345 taxa, distant OG analysis and closely related OG analysis.

While the Eastern Asian Onosma+Maharanga clade was strongly supported as monophyletic (94-100% BS, 1.00 PP) in all analyses, only in the ITS with 555 and 746 taxa analyses the Eastern Asian Onosma emerged monophyletic (87-89% BS, 0.96-1.00 PP), in the other analyses the Maharanga samples were embedded in the Eastern Asian Onosma. Similarly, Cystostemon (89-100% BS, 0.93-1.00 PP) and Maharanga (98-100% BS, 1.00 PP) were also monophyletic in all analyses. In terms of the differences between nuclear (ITS) and plastid (trnL+trnL-F+trnS-G+rps16) datasets, while Onosma s.s. (100% BS, 1.00 PP), Eastern Asian Onosma (87-91% BS, 0.92-0.98 PP) was monophyletic in all ITS analyses, the clade was not monophyletic in the plastid data analysis (Table 3). Moreover, while the Eastern Asian Onosma+Maharanga clade received a 100% BS and 1.00 PP in all ITS analyses, in the plastid data analysis this clade was supported with 80% BS and 1.00 PP (Table 3).

The AU test showed that the (((Eastern Asian Onosma+Maharanga) Onosma s.s.) Cystostemon) was rejected (p-value = -0.00133), but the (((Eastern Asian Onosma+Maharanga) Cystostemon) Onosma s.s.) topology was overwhelmingly supported (p-value = 0.999) (Table 4). In terms of outgroups (tribe Lithospermeae) Arnebia Forssk. (excluding A. hispidissima (Lehm.) A.DC.) and Buglossoides Moench. were monophyletic only in the total evidence analyses with 345 taxa (total evidence analyses with distant and closely related outgroup analyses) (Table 3). One Aegonychon sample was embedded within the Buglossoides samples, one Macrotomia sample also rendered Arnebia non-monophyletic in the remaining trees (i.e., three ITS analyses, and total evidence analysis). The remaining genera of Lithospermeae were only non-monophyletic in the plastid data analyses (unresolved) (Table 3).

Topology test results of two phylogenetic relationships (topologies) of <italic>Onosma</italic>, <italic>Maharanga</italic> and <italic>Cystostemon</italic> (p-value < 0.05 indicates statistical rejection). AU: approximately unbiased.
Topology logL AU test p-values Significantly worse
(((Eastern Asian Onosma+Maharanga) Onosma s.s.) Cystostemon) -39593.0028 -0.00133 YES
(((Eastern Asian Onosma+Maharanga) Cystostemon) Onosma s.s.) -39400.53477 0.999 NO
Discussion

To date, the current study is the first to encompass not only the widest taxon coverage in the tribe Lithospermeae, but also the Eastern Asian Onosma+Maharanga+ Cystostemon clade. The results clearly show that the phylogenetic relationships within the clade are clearly dependent on changes in taxon, gene and outgroup sampling (Table 3, Fig. 2). For instance, the “total evidence with 350 taxa” analysis yielded a topology, (((Eastern Asian Onosma+Maharanga) Onosma s.s.) Cystostemon), similar to the molecular dating analysis of total evidence tree of Chacón & al. (2019). However, this topology is rejected by the AU test (p-value = -0.00133) (Table 4, Fig. 2). On the other hand, the results of AU test overwhelmingly supported the (((Eastern Asian Onosma+Maharanga) Cystostemon) Onosma s.s.) topology (p-value = 0.999) which was recovered from the distant and closely related outgroup analyses, and the “total evidence tree with 345 taxa” analysis (Table 4, Fig. 2), a topology never reported before.

Apart from the molecular data here presented, except for the corolla morphology, Cystostemon shares great morphological similarities with Onosma and Maharanga (Cohen 2014). However, whether genus Onosma is not monophyletic, or the genus should be extended to include Maharanga and maybe Cystostemon, or the “O. rostellata and Sino-Indian Onosma +Maharanga” constitutes a different lineage are questions that remain unanswered (Cohen 2014; Chacón & al. 2019; Nasrollahi & al. 2019). The main obstacle to make a taxonomic decision is that the taxon sampling of these clades is still very limited in each of the above-mentioned studies and the present one. Any further phylogenetic analyses for taxonomic rearrangement should include more Eastern Asian Onosma, Maharanga and Cystostemon samples and perhaps more genetic regions. For instance, the genus Maharanga comprises around 10 species and Cystostemon includes 16; yet sequences from only two and four species, respectively, are currently available in GenBank. Similarly, for the Eastern Asian Onosma, except O. hookerii (seven individuals), only a few (one to three) sequences have been published. These limitations highlight the importance of taxon sampling in phylogenetic studies as have been indicated before (Zwickl & al. 2002; Heath & al. 2008).

Another unresolved question addressed here is whether the Eastern Asian Onosma species are monophyletic or not. On one hand, ITS analysis of Nasrollahi & al. (2019) and plastid/ITS analyses of Chacón & al. (2019) revealed a non-monophyletic Eastern Asian Onosma (Maharanga samples embedded in the clade); on the other hand, the ITS analysis of Cecchi & al. (2011) and molecular dating analysis of Chacón & al. (2019) with ITS+ trnL-F+rps16+trnS-trnG data yielded a monophyletic Eastern Asian Onosma clade. The results are equivocal in the present study (Table 3, Fig. 2). Eastern Asian Onosma may indeed include Maharanga, because the two genera share great morphological similarities other than corolla and anther morphology (Zhu & al. 1995); however, these results may be a consequence of sampling gaps, or the known problems associated to the ITS region, such as, concerted evolution, paralogy, gene duplication and incomplete lineage sorting (Álvarez & Wendel 2003). Although these issues have never been reported for Boraginaceae (Cecchi & al. 2011); it is still possible that ITS data problems may be responsible for these peculiar phylogenetic relationships. Similarly, as in the previous studies in tribe Lithospermeae (e.g., Cecchi & Selvi 2009; Weigend & al. 2009; Cecchi & al. 2014; Coppi & al. 2015), the results of this study have possibly indicated that the relatively low number of parsimony-informative characters of the plastid data matrix (24%) caused the phylogenetic uncertainties, and therefore, limited-plastid data may not be informative enough to solve the genus Onosma and tribe Lithospermeae phylogenetic relationships.This was already pointed out by Nasrollahi & al. (2019) who suggested to use “fast evolving genes” for future phylogenetic studies of Onosma. Unfortunately the results of the present study hint that even these types of genes may not be enough to solve the phylogenetic inconsistency, due to the reason of the effect of taxon sampling on the group phylogeny and possible complex evolutionary histories of both clades (Nasrollahi & al. 2019). Besides, morphological characters may also not be helpful to answer the phylogenetic questions within the genus (i.e., similar morphologies of Onosma, Maharanga and Cystostemon), even when they are supported by molecular data. Thus, both the data and taxon sampling (i.e., both the ingroup and outgroup) hold great importance in Onosma phylogenetic analyses.

While Nasrollahi & al. (2019) reported a monophyletic position for Onosma hookeri and to date only one O. sinicum Diels sample was included in the previous studies (e.g., Chacón & al. 2019), the analyses of the current study showed that neither O. sinicum nor O. hookeri are monophyletic (Fig. 1). Indeed, misidentifications are common in Boraginales (Dr. Aslı Doğru Koca, pers. comm.) and the amount of wrong identifications in voucher specimens is reported to be very high in public molecular repositories (Nilsson & al. 2006; Wu & al. 2021). In the current study, I did not filter GenBank sequences by their reliable IDs or deposited voucher numbers. Indeed, this could be one of the reasons for the non-monophyly of these taxa. For example, a quick survey has shown that among the 746 ITS sequences included in the current study, only 489 of them (~66%) have their voucher numbers included in GenBank (results not shown). Although a voucher specimen will not guarantee a correct identification it provides scientific reliability and possibility of verification in the future (Wu & al. 2019). Therefore, future studies should be mindful about possible wrong identifications while adding GenBank sequences in their datasets.

The monophyly of most outgroup genera was well supported (Table 3). However, future studies aiming to evaluate the phylogenetic relationships within tribe Lithospermeae should be aware of including a more thorough sampling, particularly from genera Buglossoides, Arnebia, Halacsya, Macrotomia, Stenosolenium and Aegonychon.

The causes for phylogenetic incongruency reported for many angiosperm clades such as, Fabales (Aygoren Uluer & al. 2020a), Saxifragales (Jian & al. 2008), tribe Lithospermeae (Cohen 2011) and tribe Mirbelieae (Barret & al. 2021), can be explained by several non-exclusive reasons, such as, rapid radiation of the clade (e.g., Cohen 2011), inadequate data (Zeng & al. 2017), long branch attraction (LBA) (e.g., Qiu & al. 2001), conflicting gene trees (e.g., Yang & al. 2013), effect of outgroup sampling (e.g., Smith 1994; Aygoren Uluer & al. 2020b), reticulate events (e.g., hybridization, introgression and horizontal gene transfer, incomplete lineage sorting) (e.g., Solís-Lemus & Ané 2016) and/or stochastic noise.Further studies with a comprehensive ingroup and in outgroup sampling may answer possible causes for the phylogenetic problems among Onosma, Cystostemon and Maharanga. Therefore, future studies should increase the amount of data, if not the use of whole genome sequencing, the use of wide-genome RAD-seq data. Certainly, an increased taxon sampling of Turkish Onosma species, Eastern Asian Onosma species, Cystostemon and Maharanga should be a priority in future molecular studies.

 References Álvarez I. Wendel J.F. 2003 Ribosomal ITS sequences and plant phylogenetic inference Molecular Phylogenetics and Evolution 29 417 434 Aygoren Uluer D. Hawkins J.A. Forest F. 2020a Interfamilial relationships in order Fabales: new insights from the nuclear regions sqd 1 and 26S rDNA Plant Systematics and Evolution 306 1 14 Aygoren Uluer D. Forest F. Hawkins J.A. 2020b Supermatrix analyses and molecular clock rooting of Fabales: Exploring the effects of outgroup choice and long branch attraction on topology Botany 98 231 247 Barrett R.L. Clugston J.A. Cook L.G. Crisp M.D. Jobson P.C. Lepschi B.J. Renner M.A. Weston P.H. 2021 Understanding Diversity and Systematics in Australian Fabaceae Tribe Mirbelieae Diversity 13 391 Cecchi L. Selvi F. 2009 Phylogenetic relationships of the monotypic genera Halacsya and Paramoltkia and the origins of serpentine adaptation in circum-mediterranean Lithospermeae (Boraginaceae): insights from ITS and matK DNA sequences Taxon 58 700 714 Cecchi L. Coppi A. Selvi F. 2011 Evolutionary dynamics of serpentine adaptation in Onosma (Boraginaceae) as revealed by ITS sequence data Plant Systematics and Evolution 297 185 199 Cecchi L. Coppi A. Hilger H.H. Selvi F. 2014 Non-monophyly of Buglossoides (Boraginaceae: Lithospermeae): Phylogenetic and morphological evidence for the expansion of Glandora and reappraisal of Aegonychon Taxon 63 1065 1078 Chacón J. Luebert F. Weigend M. 2017 Biogeographic events are not correlated with diaspore dispersal modes in Boraginaceae Frontiers in Ecology and Evolution 5 26 Chacón J. Luebert F. Selvi F. Cecchi L. Weigend M. 2019 Phylogeny and historical biogeography of Lithospermeae (Boraginaceae): Disentangling the possible causes of Miocene diversifications Molecular Phylogenetics and Evolution 14 106626 Cohen J.I. Davis J.I. 2009 Nomenclatural changes in Lithospermum (Boraginaceae) and related taxa following a reassessment of phylogenetic relationships Brittonia 61 101 111 Cohen J.I. 2011 A phylogenetic analysis of morphological and molecular characters of Lithospermum L. (Boraginaceae) and related taxa: evolutionary relationships and character evolution Cladistics 27 559 580 Cohen J.I. 2014 A phylogenetic analysis of morphological and molecular characters of Boraginaceae: evolutionary relationships, taxonomy, and patterns of character evolution Cladistics 30 139 169 Coppi A. Cecchi L. Nocentini D. Selvi F. 2015 Arnebia purpurea: a new member of formerly monotypic genus Huynhia (Boraginaceae-Lithospermeae) Phytotaxa 204 123 136 Darriba D. Taboada G.L. Doallo R. Posada D. 2012 jModelTest 2: more models, new heuristics and parallel computing Nature Methods 9 772 Guindon S. Gascuel O. 2003 A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood Systematic Biology 52 696 704 Heath T.A. Hedtke S.M. Hillis D.M. 2008 Taxon sampling and the accuracy of phylogenetic analyses Journal of Systematics and Evolution 46 239 257 Jian S. Soltis P.S. Gitzendanner M.A. Moore M.J. Li R. Hendry T.A. Qiu Y.L. Dhingra A. Bell C.D. Soltis D.E. 2008 Resolving an ancient, rapid radiation in Saxifragales Systematic Biology 57 38 57 Kearse M. Moir R. Wilson A. Stones-Havas S. Cheung M. Sturrock S. Buxton S. Cooper A. Markowitz S. Duran C. Thierer T. 2012 Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data Bioinformatics 28 1647 1649 Koyuncu O. Yaylacı Ö.K. Özgişi K. Sezer O. Öztürk D. 2013 A new Onosma (Boraginaceae) species from central Anatolia, Turkey Plant Systematics and Evolution 299 1839 1847 Letunic I. Bork P. 2016 Interactive tree of life (iTOL) v3: an online tool for the display and annotation of phylogenetic and other trees Nucleic Acids Research 44 W242 W245 Mehrabian A.R. Sheidai M. Noormohammadi Z. Asrei Y. Mozafarian V. 2011 Inter-simple sequence repeats (ISSR) and morphological diversity in Onosma L. (Boraginaceae) species in Iran African Journal of Biotechnology 10 10831 10838 Miller M.A. Pfeiffer W. Schwartz T. 2011 The CIPRES science gateway: a community resource for phylogenetic analyses Proceedings of the Gateway Computing Environments Workshop GCE New Orleans, Louisiana Nasrollahi F. Kazempour-Osaloo S. Saadati N. Mozaffarian V. Zare-Maivan H. 2019 Molecular phylogeny and divergence times of Onosma (Boraginaceae ss) based on nrDNA ITS and plastid rpl32-trnL (UAG) and trnH-psbA sequences Nordic Journal of Botany 37 e02060 Nilsson R.H. Ryberg M. Kristiansson E. Abarenkov K. Larsson K.H. Kõljalg U. 2006 Taxonomic reliability of DNA sequences in public sequence databases: a fungal perspective PloS ONE 1 e59 Qiu Y.L. Lee J. Whitlock B.A. Bernasconi-Quadroni F. Dombrovska O. 2001 Was the ANITA rooting of the angiosperm phylogeny affected by long-branch attraction? Molecular Biology and Evolution 18 1745 1753 Rambaut A. Drummond A.J. 2007 Tracer 1. 4 https://beast.community/tracer 23 Mar. 2022 Ronquist F. Huelsenbeck J.P. 2003 MrBayes 3: Bayesian phylogenetic inference under mixed models Bioinformatics 19 1572 1574 Selvi F. Cecchi L. Hilger H.H. Coppi A. 2017 A reappraisal of the genus Megacaryon (Boraginaceae, Lithospermeae) based on molecular, morphological, and karyological evidence Systematics and Biodiversity 15 552 563 Shimodaira H. Hasegawa M. 1999 Multiple comparisons of log-likelihoods with applications to phylogenetic inference Molecular Biology and Evolution 16 1114 1114 Smith A.B. 1994 Rooting molecular trees: problems and strategies Biological Journal of the Linnean Society 51 279 292 Solís-Lemus C. Ané C. 2016 Inferring phylogenetic networks with maximum pseudolikelihood under incomplete lineage sorting PLoS Genetics 12 e1005896 Stamatakis A. 2014 RAxML version 8: A tool for phylogenetic analysis and post-analysis of large phylogenies Bioinformatics 30 1312 1313 Thomas D.C. Weigend M. Hilger H.H. 2008 Phylogeny and systematics of Lithodora (Boraginaceae-Lithospermeae) and its affinities to the monotypic genera Mairetis, Halacsya and Paramoltkia based on ITS1 and trnLUAA-sequence data and morphology Taxon 57 79 97 Trifinopoulos J. Nguyen L.T. von Haeseler A. Minh B.Q. 2016 W-IQ-TREE: a fast online phylogenetic tool for maximum likelihood analysis Nucleic Acids Research 44 W232 W235 Weigend M. Gottschling M. Selvi F. Hilger H.H. 2009 Marbleseeds are gromwells-Systematics and evolution of Lithospermum and allies (Boraginaceae tribe Lithospermeae) based on molecular and morphological data Molecular Phylogenetics and Evolution 52 755 768 Wu H.Y. Chan K.T. But G.W.C. Shaw P.C. 2021 Assessing the reliability of medicinal Dendrobium sequences in GenBank for botanical species identification Scientific Reports 11 1 9 Yang H.M. Zhang Y.X. Yang J.B. Li D.Z. 2013 The monophyly of Chimonocalamus and conflicting gene trees in Arundinarieae (Poaceae: Bambusoideae) inferred from four plastid and two nuclear markers Molecular Phylogenetics and Evolution 68 340 356 Zeng L. Zhang N. Zhang Q. Endress P.K. Huang J. Ma H. 2017 Resolution of deep eudicot phylogeny and their temporal diversification using nuclear genes from transcriptomic and genomic datasets New Phytologist 214 1338 1354 Zhu G.L. Harald R. Rudolf K. 1995 Boraginaceae Wu Z. Y. Raven P. H. Flora of China 16 329 427 Science Press Beijing Missouri Botanical Garden Press St. Louis Zwickl D.J. Hillis D.M. 2002 Increased taxon sampling greatly reduces phylogenetic error Systematic Biology 51 588 598 <p><bold>Appendix 1.</bold> Alignment of ITS (746 taxa). <bold>Appendix 2.</bold> Alignment of total evidence analysis with 350 taxa. <bold>Appendix 3.</bold> ITS (323 taxa) RAxML result. <bold>Appendix 4.</bold> ITS (555 taxa) RAxML result. <bold>Appendix 5.</bold> ITS (746 taxa) RAxML result. <bold>Appendix 6.</bold> Plastid (339 taxa) RAxML result. <bold>Appendix 7.</bold> Total evidence (350 taxa) RAxML result. <bold>Appendix 8.</bold> Total evidence (345 taxa) RAxML result. <bold>Appendix 9.</bold> Total evidence with distant outgroup (347 taxa) RAxML result. <bold>Appendix 10.</bold> Total evidence with closely related outgroup (347 taxa)RAxML result. All appendices files are available in the Mendeley data repository at <ext-link ext-link-type="uri" xlink:href="https://data.mendeley.com/datasets/f9ghrnzghx/1">https://data.mendeley.com/datasets/f9ghrnzghx/1</ext-link></p> </app> </app-group> </back> </article>