- Research article
- Open Access
Complex patterns of reticulate evolution in opportunistic weeds (Potentilla L., Rosaceae), as revealed by low-copy nuclear markers
BMC Evolutionary Biology volume 20, Article number: 38 (2020)
Most cinquefoils (Potentilla L., Rosaceae) are polyploids, ranging from tetraploid (4x) to dodecaploid (12x), diploids being a rare exception. Previous studies based on ribosomal and chloroplast data indicated that Norwegian cinquefoil (P. norvegica L.) has genetic material from two separate clades within Potentilla; the Argentea and the Ivesioid clades – and thus a possible history of hybridization and polyploidization (allopolyploidy). In order to trace the putative allopolyploid origin of the species, sequence data from low-copy, biparentally inherited, nuclear markers were used. Specimens covering the circumpolar distribution of P. norvegica and its two subspecies were included, along with the morphologically similar P. intermedia. Potentilla species of low ploidy level known to belong to other relevant clades were also included.
Gene trees based on three low-copy nuclear markers, obtained by Bayesian Inference and Maximum Likelihood analyses, showed slightly different topologies. This is likely due to genomic reorganizations following genome duplication, but the gene trees were not in conflict with a species tree of presumably diploid taxa obtained by Multispecies Coalescent analysis. The results show that both P. norvegica and P. intermedia are allopolyploids with a shared evolutionary history involving at least four parental lineages, three from the Argentea clade and one from the Ivesioid clade.
This is the first time that reticulate evolution has been proven in the genus Potentilla, and shows the importance of continuing working with low-copy markers in order to properly resolve its evolutionary history. Several hybridization events between the Argentea and Ivesioid clades may have given rise to the species of Wolf’s grex Rivales. To better estimate when and where these hybridizations occurred, other Argentea, Ivesioid and Rivales species should be included in future studies.
The evolution of species is usually considered to be a slow process, working over thousands or even millions of years. Sometimes, however, new species evolve within a relatively short period of time through polyploidization. This phenomenon is common throughout the vascular plants, where genome duplications can be found from the ferns  and lycopods , to the asterids . Two main types of polyploidization are recognized; autopolyploidization, where the duplication occurs within a single species, and allopolyploidization, where the duplication occurs in combination with hybridization between two different species . A doubling of the chromosomes can make a sterile hybrid fertile [5, 6] and cause a reproductive barrier between individuals of the new genomic state and the old state [6, 7]. This may create a new, independently evolving, lineage that could thus be regarded as a new species .
The rose family (Rosaceae Juss.) is well known for its many polyploid taxa, and there seem to have been a large number of independent auto- and allopolyploidization events during its evolutionary history [9,10,11]. Chromosome counting data, summarized by Vamosi and Dickinson , suggest that around half of the family’s genera include at least one polyploid species. Some, as for instance Acaena L., Alchemilla L. and Sorbaria (Ser.) A. Braun, consist only of polyploids.
The cinquefoils, Potentilla L., is an example of a genus in Rosaceae with mixed ploidy levels. According to the Chromosome Counts Database  only a few species seem to be exclusively diploid, e.g. P. biflora Willd. ex Schltdl., P. freyniana Bornm. and P. valderia L. At the other end, P. gracilis Douglas ex Hook., P. tabernaemontani Asch. and P. indica (Jacks.) Th. Wolf have been reported to have dodecaploid (12x) populations. Furthermore, it is not uncommon for single species to have multiple ploidy levels. The genus has undergone a major recircumscription since the first molecular studies of the group were performed [14, 15]; both plastid and nuclear ribosomal markers showed that it had been polyphyletic. They strongly indicated that some previous Potentilla species are more closely related to the strawberries, Fragaria L., in the Fragariinae clade, such as those species now assigned to the genera Dasiphora Raf. and Drymocallis Fourr. In contrast, the genus Duchesnea Sm. and some species of Sibbaldia L., were instead shown to belong to Potentilla [14, 16]. However, the debate on where to draw the generic delimitation is still ongoing; as whether to include the genus Argentina Hill. and its sisters [15, 17] or not [18,19,20]. Regardless whether Argentina is included or not, the genus is still polyphyletic in certain classifications where Duchesnea (P. indica) and the genera of the North American Ivesioid clade (Horkelia Cham. & Schltdl., Horkeliella (Rydb.) Rydb. and Ivesia Torr. & A.Gray) are separated from Potentilla [17, 21, 22]. Within Potentilla in the strict sense, there are a number of well supported subclades, such as the Alba, Reptans and Ivesioid clades . The most species-rich subclade, called either “Argentea”  or “core group”  in previous studies, is, however, in itself poorly resolved [18, 20, 23].
Previous studies have found a possible connection between the Argentea and Ivesioid clades in the polyploid species P. norvegica L. This species has been shown to have different phylogenetic relationships depending on whether the analyses were based on chloroplast [15, 18, 23] or nuclear ribosomal data [14, 15, 23]; with chloroplast data the species groups with the Argentea clade, but with ribosomal data it groups with the Ivesioids. Töpel et al.  speculated that this may be due to an evolutionary history of polyploidization in combination with hybridization between these two clades. It is, however, not previously known to what extent these two processes have played a part in the formation of P. norvegica, or if the discordance between chloroplast and ribosomal data is the result of other processes, such as a single hybridization event followed by introgression .
In his monograph of Potentilla, Wolf  placed P. norvegica together with 20 other species in his “grex” Rivales. Of these, P. intermedia L. and P. supina L. have a similar circumpolar distribution as P. norvegica, while the North American species P. biennis Greene and P. rivalis Nutt. are morphologically similar to P. norvegica. Another common feature is that they are annuals or short-lived perennials [17, 25]. Potentilla norvegica was originally described by Linnaeus  as two separate species based on stem and leaflet morphology of European specimens; P. norvegica L. and P. monspeliensis L. In 1803, Michaux  described P. hirsuta Michx. based on North American specimens, but Ledebour  later synonymized P. monspeliensis and P. hirsuta under P. norvegica. Nevertheless, there is striking morphological variation within the species, and today two subscpecies are generally accepted. However, it has been unclear which subspecies name has priority. In 1904, Ascherson and Graebner  described “P. norvegica II. monspeliensis”, by some nomenclatural databases interpreted as a subspecies [30, 31]. However, Hylander  must have interpreted this as a variety. Since names only have priority at the same nomenclatural rank , he was able to list “II. monspeliensis” under P. norvegica ssp. hirsuta (Michx.) Hyl. The name that will be used in this study is therefore Potentilla norvegica ssp. hirsuta, which refers to specimens displaying the morphology first used to describe P. monspeliensis. Since P. norvegica ssp. hirsuta is the most common subspecies in North America, it is sometimes referred to as the American form, and the autonym ssp. norvegica as the European form, but there are numerous findings of ssp. hirsuta in Europe. Most floras argue for an East European origin of the species, and that ssp. hirsuta later has dispersed to Europe from North America [34,35,36,37,38]. However, no molecular phylogenetic work has been performed in order to test these hypotheses.
The two types of molecular data most commonly used in phylogenetic studies of plants both have the inconvenience that they are not able to detect reticulate patterns in phylogenetic trees. The chloroplast is inherited uniparentally and nuclear ribosomal markers are most often subject to concerted evolution, while low-copy nuclear markers are inherited biparentally and present in each subgenome after a polyploidization event . This means that they have the potential to retrieve polyploid signals in a single gene tree. For instance, Smedmark et al.  resolved the Colurieae clade in Rosaceae with its many polyploid species using this type of marker. However, different gene trees do not necessarily depict the same evolutionary history, due to processes such as horizontal gene transfer, deep coalescence and lineage sorting . Furthermore, since it is not possible to know beforehand which sequences are homologous, low-copy markers cannot be concatenated to form larger datasets. Therefore, when polyploidy is present, it is important to investigate several low-copy markers in order to find the species tree. In a phylogenetic gene tree covering a simple polyploidization event, the gene copies of an autopolyploid (paralogues) would be each other’s sisters, while the gene copies of an allopolyploid (homoeologues) would be sisters to their respective parental lineage. This has a number of effects on species trees, since the evolutionary history of an allopolyploid would be better represented by a reticulate pattern where lineages merge, rather than by a traditional bifurcating tree .
By using low-copy nuclear markers, this study aims to determine (1) if Potentilla norvegica and P. intermedia have an allopolyploid evolutionary history resulting from hybridization between the Argentea and Ivesioid clades; (2) if this is the case, do they share polyploidy events; and (3) if morphology and geography are concordant with intraspecies phylogeny in P. norvegica.
In addition, two GAPCP1 sequences from P. intermedia were identical to two P. norvegica sequences (P to 97E and D to 113D), while the GBSSI-1 P. intermedia sequence Kb and P. norvegica sequence 96N differed in only one base pair.
Partitioning and model suggestions
The lowest log likelihood value for the partitioning and model analyses were obtained under the AICc criterium for all markers. Partitioning schemes and their assigned models are found in Table 1.
Bayesian and ML analyses
The Bayesian phylogenetic analysis of GAPCP1 resolved Potentilla norvegica sequences in four clades (Fig. 1). Three of these clades were sisters to Argentea species (clade A1, posterior probability 1.0; A2, pp. 1.0; B, pp. 0.96) and one was sister to the Ivesioids (C, pp. 1.0). Potentilla intermedia was found in the same four clades. The A1 and A2 clades formed a polytomy together with two P. intermedia sequences (A, pp. 0.99). The node connecting the A and B clades, i.e. corresponding to the Argentea clade, was not strongly supported (pp 0.82). The Ivesioid genera (Horkelia, Horkeliella and Ivesia) in clade C were divided into two subclades (both pp. 1.0), with at least one sequence from each species in each subclade. The Maximum Likelihood analysis showed the same topology, but only clades A1 and A2 were supported (bootstrap support 100 and 96, respectively).
The Bayesian analysis of GBSSI-1 showed P. norvegica sequences in three clades (Fig. 2), of which two correspond to A2 (pp 0.97) and C (pp 1.0) in the GAPCP1 tree. There was, however, no P. norvegica homoeologue associated with the Argentea species in clade B (pp 1.0). Potentilla intermedia homoeologues were found in clades A2, B and C. Clades A2 and B were sisters with low support (pp 0.82). They formed a polytomy (pp 0.87) with the third P. norvegica clade (pp 1.0) and a small clade consisting of one P. norvegica and one P. intermedia sequence (pp. 0.93). This polytomy was in turn in a polytomy (pp 1.0) with clade C and the Argentea species from clade A1 (pp 1.0). Thus, there was no Argentea clade in this tree. Within clade C, the Ivesioid species formed one subclade (pp 0.98), in which two of the four Ivesia sequences were sisters to Horkelia (pp 0.99), while the other two were unresolved. The ML analysis showed clades A1 (bs 66), A2 (bs 78), B (bs 99) and C (bs 93), but their relative positions were not supported. The clade with only P. norvegica sequences, present in the Bayesian tree, was placed as sister to P. aurea and P. brauneana (A1) in the ML tree. Even though bootstrap support was low, we will refer to this P. norvegica clade as A1†.
The Bayesian analysis of DHAR2 (Fig. 3) also showed P. norvegica in three clades, two of them corresponding to A1 (pp 1.0) and C (pp 0.93) in the other trees, while the third had not been seen previously. This clade consisted of P. norvegica, P. intermedia and one P. heptaphylla sequence, and was supported as sister to clade C (pp 1.0), while the clade itself had low support (pp 0.86). There was no supported Argentea clade in this tree. The Ivesioids formed one subclade in clade C, where one of two Horkelia sequences and one of two Ivesia sequences were sisters (pp 1.0), while the other two were unresolved. The ML analysis showed no conflicting topology of the major clades, but there were two Ivesioid subclades (bs 83 and 100), with one Ivesia and one Horkelia sequence in each, and those were supported as sisters (bs 80). The sister clade to clade C was also supported (bs 76).
No clade was specific to, or missing, any of the two P. norvegica subspecies or seven individuals throughout all three gene trees. For instance, clade C was missing individual 97 in the GAPCP1 tree and individuals 92, 95, 97 and 112 in the DHAR2 tree, while all individuals were represented in this clade in the GBSSI-1 tree.
Five species with previously published diploid chromosome counts , P. aurea, P. chinensis, P. clusiana, P. fragarioides and P. heptaphylla, failed direct sequencing and were therefore molecularly cloned. In the GBSSI-1 and DHAR2 trees, P. aurea was sister to P. brauneana in clade A1 (pp 1.0). However, in the GAPCP1 tree two P. aurea sequences were placed in clade A1, but the other two were placed in clade A2 as sisters to P. chinensis (pp 0.82). In the GAPCP1 tree, all P. heptaphylla sequences were placed in clade B, but in the GBSSI-1 tree two sequences were found in A1 and two found in A2. In the DHAR2 tree they were even further apart, with one sequence as sister to P. chinensis in A2/B and one as sister to P. norvegica and P. intermedia in the sister clade to clade C. The sequences of P. chinensis, P. clusiana and P. fragarioides formed clades of their own.
The control ML analyses for putatively missed P. norvegica gene copies did not reveal any new clades or overlooked patterns in terms of subspecies or geographical origin. However, two excluded P. intermedia GBSSI-1 sequences were indicated to belong in clade A1. One of these was added to the dataset, but the Bayesian analysis resulted in the collapse of clades B and C, which received high support in the other trees. Similarly, one P. intermedia DHAR2 sequence was indicated to belong in clade C, but when added to the dataset it also resulted in the collapse of several clades. Both sequences were therefore excluded again from their respective datasets.
Multispecies coalescent analysis
The substitution model suggested for all markers was HKY , with gamma as site heterogeneity model for GAPCP1 and GBSSI-1, and invariant sites for DHAR2. The clock model and tree prior that was best fitted to the low-copy marker only dataset was a relaxed uncorrelated lognormal clock with a birth-death process, and for the combined low-copy and chloroplast marker dataset a relaxed uncorrelated lognormal clock with a birth process. The two trees had the same topology, but some of the support values differed (Fig. 4). In both trees, the Ivesioid clade was supported (pp 1.0) and P. aurea and P. brauneana were sisters (pp 1.0), corresponding to clade A1 in the gene trees. Potentilla hirta, P. heptaphylla and P. argentea formed a polytomy (pp 0.95 in the low-copy marker dataset and pp. 0.88 in the combined dataset) corresponding to clade B, while P. chinensis of clade A2 was unresolved. The Argentea clade received low support (pp 0.82) in the low-copy marker tree and full support (pp 1.0) in the combined tree.
Most specimens studied from the collections of BG, GB, O, S and UPS were of intermediate morphology. They had, for instance, whole stipules (ssp. norvegica), but obovate leaflets and obtuse leaflet teeth (ssp. hirsuta). For European specimens, there was approximately equal occurrence of typical individuals of the two subspecies. For the North American and East Russian specimens, typical individuals showing the ssp. hirsuta morphology were more common than those showing the ssp. norvegica morphology. The few North American specimens showing the ssp. norvegica morphology were all but one (Alaska, USA) collected in the East (Ontario, Canada, to New York, USA), a pattern also seen by Rydberg .
Despite the slightly different topologies of the three single-copy nuclear markers presented in this study, it is clear that both Potentilla norvegica and P. intermedia are allopolyploids with a shared evolutionary history involving one parental lineage in the Ivesioid clade and multiple parental lineages in the Argentea clade. These results rule out a simple case of introgression, and reveal a complex reticulate evolutionary history of several hybridization events in combination with polyploidization. For P. norvegica, there was no condordance between geography and intraspecies phylogeny. Thus, on the basis of our data we see no support for species differentiation, as first suggested by Linnaeus , since the majority of the individuals studied in the herbaria were of intermediate morphological form. Neither did our molecular data support a division into subspecies, but a more extensive study involving more individuals of especially ssp. norvegica would be better able to investigate the relationship between them.
As previously shown in studies based on chloroplast and ribosomal data [14, 15, 18, 20, 23, 44], the Ivesioid clade is deeply nested in Potentilla (Figs. 1, 2 and 3). Thus, following the established practice of only recognizing monophyletic taxa, the Ivesioid genera Horkelia, Horkeliella and Ivesia should be incorporated in Potentilla. The type species of Potentilla, P. reptans, is part of the small Reptans clade, which is the sister clade to the Argentea and Ivesioid clades. If the Ivesioid genera were to be retained, the many species of the large Argentea clade would have to be reclassified, and it is probable that almost all would have to change names. However, the new evidence presented here of a hybridization event between the Argentea and Ivesioid clades indicate a close relationship between the groups, and adds a compelling argument for including the Ivesioid genera in Potentilla.
The three gene trees conform well to the backbone reference (Fig. 4), apart from some P. aurea and P. heptaphylla sequences. It is, however, clear that one P. norvegica GBSSI-1 homoeologue (subgenome-specific gene copy) is missing in clade B and one P. intermedia GBSSI-1 homoeologue is missing in clade A1† (Fig. 2). In the DHAR2 tree (Fig. 3), there is a major rearrangement in which the Ivesioid clade C is sister to what could be assumed to be parts of clade A2 or B. In addition, contrary to previous analyses based on chloroplast and nuclear ribosomal data [16, 18, 20, 23], the support for the Argentea clade was low both in the individual gene trees (Figs. 1, 2 and 3) and in the species tree based on low-copy markers only (Fig. 4). Thus, it is evident that phylogenetic relationships of low-copy nuclear genes are complicated by a number of evolutionary processes. A polyploid genome with high genetic redundancy may be subjected to large genomic alterations, such as deletions, insertions, or recombinations, to a high extent without causing fatal effects . For instance, entire homoeologues may be lost as a response to genomic reorganization after polyploidization [46, 47] or via incomplete lineage sorting during speciation after hybridization . Furthermore, if an interallelic recombination  splits a gene in two unequal parts during meiosis, the new recombinant will position itself as sister to its major donor in the gene tree, and such a process might explain the clade rearrangement seen in the DHAR2 treee.
Previous dating analyses have assigned somewhat different ages to the Potentilla crown group (excluding Argentina), either between ca 36 to 15 Mya [18, 20, 48] or between ca 56 to 32 Mya . Estimations of the Agentea-Ivesioid split also varies, with ages between 15.2–9.8 Mya [18, 48] and 36.6–18.7 Mya . There is also disagreement as to whether the Argentea crown clade is younger  or older [18, 48] than the Ivesioid crown clade, but this may be a sampling issue since undersampling of a species rich sister clade would tend to result in underestimating the age of its crown. Today, the Argentea clade consists of the majority of the Potentilla species. They have a circumpolar distribution in the Northern Hemisphere, are adapted to a variety of climates, and are of multiple ploidy levels. In contrast, the Ivesioids are limited to dry areas in western United States  and are, as far as known, tetraploid . According to Töpel et al.  they also evolved in the same area, while Dobeš and Paule  estimated an origin in East Asia both for the Potentilla crown group and the Ivesioids. However, considering the Ivesioids being geographically restricted and ecologically specialized, the Western American origin of the crown clade found by Töpel et al.  may be the most plausible. It is notable, however, that if they are indeed sister groups, their stem lineages are of the same age, and any species that would fall below the crown clades of Argentea or the Ivesioids are either unsampled or extinct.
During the Eocene (56–33.9 Mya ), before or in the early stages of the diversification of the Potentilla crown group, the North Atlantic land bridge was broken up [50, 51] and the Turgai strait still separated Asia from Europe . A land bridge over the Bering strait existed during most of the later Tertiary to mid Pliocene [51,52,53], and the original dispersal of the Ivesioid and Argentea ancestors from Asia to North America is most likely to have occurred before its breakup. Today the Bering Strait area is subject to very cold and long winters, but the clade ages suggested by Töpel et al.  indicate that the dispersal may have coincided with the Mid Miocene Climatic Optimum, when the Earth was on average 3 °C warmer than present . However, considering the current cold climate tolerance of both P. norvegica and P. intermedia [17, 38], dispersal did not necessarily have to have coincided with warmer periods. Therefore, the younger clade ages estimated by Dobeš and Paule  and Feng et al.  need not be dismissed.
Regardless of their relative ages, and judging from extant species, the Argentea clade has gone through many more speciations, polyploidizations and hybridizations than the Ivesioid clade. Nonetheless, there is an indication of an early autopolyploid event in the Ivesioids, and this is especially evident in the GAPCP1 tree (Fig. 1); the two subclades in clade C each contain one or two sequences of all Ivesioid species included.
The single P. norvegica homoeologue in clade C (Figs. 1, 2 and 3) indicates that the Argentea-Ivesioid hybridization may have happened before polyploidization and diversification of the Ivesioid crown group. This makes the hybridization event difficult to pinpoint geographically; Töpel et al.  predicted a wide climate preference for the Ivesioid ancestor, and both P. norvegica and P. intermedia have weedy growth habits and can be found all around the Northern Hemisphere. Neither is it possible to say, based on our species sample and the resolution of our gene trees, if the Argentea-Ivesioid hybridization is the oldest or the most recent. To illustrate the mode of speciation that P. norvegica and P. intermedia have gone through, one possible chain of events is shown in Fig. 5 based on our interpretation of the GAPCP1 tree.
The phylogenetic pattern seen in chloroplast markers of Rivales species (in the sense of Wolf ) occurring in North America suggests that other species than P. norvegica and P. intermedia may have connections to the Argentea-Ivesioid hybridization event [18, 44]. Based on chloroplast data, P. norvegica, P. newberryi, P. rivalis and P. supina resolve with the Argentea clade, while P. biennis is sister to the Ivesioid clade. For P. norvegica, it is evident that the pollen donor came from the Ivesioid clade , and therefore it is notable that P. biennis is the only Rivales species that resolves with the Ivesioid clade. Potentilla biennis, P. newberryi and P. rivalis have a limited central to western North American distribution similar to that of the Ivesioids . In addition, P. biennis and P. rivalis are morphologically similar to P. norvegica. Thus, it seems likely that the Argentea-Ivesioid hybridization event occurred in North America rather than in Asia. That would make the Eastern European origin of P. norvegica, as proposed by many floras [35, 37, 38], doubtful. It is therefore possible that the Rivales group originated following multiple hybridization events between the two clades. To better pinpoint where they occurred and which evolutionary routes that were then taken by the lineages that emerged, additional Argentea and Rivales species of various ploidy levels should be included in future analyses, such that all continents are better covered.
The four homoeologues that were found in P. norvegica had a high degree of variation. In the case of P. intermedia, this variation seemed even greater, since it is found in more subclades than P. norvegica. Both P. norvegica and P. intermedia have more than one ploidy level reported , and there are many other examples of plant populations with mixed ploidy levels [42, 55,56,57]. Sterile hybrids may still be able to produce offspring through apomixis, and this apomixis is in turn heritable . According to Asker , both P. norvegica and P. intermedia can reproduce in this manner, which could explain the existence of multiple ploidy levels and high sequence variation within the two species. In addition, several of the putatively diploid species (P. aurea, P. chinensis, P. clusiana, P. fragarioides and P. heptaphylla)  included in this study failed direct sequencing of all markers, and showed a remarkable sequence variation. Potentilla heptaphylla was resolved together with P. argentea and P. hirta in the backbone reference (pp 0.95/0.88) (Fig. 4), but was seen in three different clades (A2, B and C) in the separate gene trees. This suggests allopolyploidy rather than single gene duplications, since the gene copies were resolved as sisters to different species in the same gene tree. The ploidy level of P. aurea is difficult to determine solely from the results presented here, since it is found in clades A1 and A2 in the GAPCP1 tree, but only in clade A1 in the GBSSI-1 and DHAR2 trees. However, as seen for P. norvegica in the GBSSI-1 and DHAR2 trees, it is possible that P. aurea and P. heptaphylla have lost homoeologues too. Future studies of polyploid species in Potentilla should consider chromosome counting and flow cytometry of the specimens included in order to more securely connect the gene trees with ploidy level, in addition to recreate a more accurate, reticulate species tree.
This is the first study of species level relationships and reticulate patterns in Potentilla based on low copy nuclear markers. With this type of data it was possible to reveal a complex evolutionary history of polyploidizations and hybridizations, not only within previously identified subclades, but also between subclades. The nature of the results, and implications for the interpretation of evolutionary events and distribution patterns, demonstrate the importance of continued work with this kind of data.
The gene trees showed that P. norvegica and P. intermedia are allopolyploids with multiple parental lineages in the Argentea clade, and one in the Ivesioid clade. This close relationship between the two clades is one of several arguments for an inclusion of the genera of the Ivesioid clade (Horkelia, Horkeliella and Ivesia) in Potentilla. This inclusion would help to make Potentilla monophyletic.
Gene sequences from both Potentilla norvegica and P. intermedia are present in the same major clades. This indicates that the allopolyploidy events occurred in their common ancestral lineage.
This study shows no support for species differentiation of P. norvegica, as previously suggested, since there was no condordance between geography and intraspecies phylogeny. In addition, the majority of the preserved specimens studied were of intermediate morphological form between the two subspecies. A more extensive study including more specimens is needed in order to determine the support for recognition of the subspecies.
Hybridization between the Argentea and Ivesioid clades may have occurred several times and given rise to the species of Wolf’s grex Rivales . To better estimate when and where these hybridizations occurred, other Argentea and Rivales species of various ploidy levels should be included in future studies, such as P. rivalis and P. biennis.
To cover the circumpolar distribution of Potentilla norvegica L.,  herbarium material of one morphologically typical individual of each subspecies, ssp. norvegica and ssp. hirsuta (Michx.) Hyl., were included from Scandinavia and central Europe, in addition to two North American and one eastern Russian specimen of ssp. hirsuta. From the Argentea clade, species were selected if they had reported diploid populations , and from the Ivesioid clade the type species of Horkelia and Ivesia were selected. Low-ploidy outgroup species were selected from the Reptans, Fragarioides and Alba clades. Potentilla intermedia L. was also included since it shares several features with P. norvegica: similar morphology, weedy growth habit and assigned to grex Rivales by Wolf , and could therefore be suspected to have a similar evolutionary history as P. norvegica. All specimens included are listed in Table 2.
New primer pairs were designed for three low-copy nuclear markers (Table 3); GAPCP1 (glyceraldehyde-3-phosphate dehydrogenase) with primer sites in exons 11 and 14, GBSSI-1 (granule-bound starch synthase I) in exons 1 and 4 and DHAR2 (dehydroascorbate reductase 2) in exons 4 and 5. In order to find suitable primer placements, the 150 base pair long Illumina raw reads of a Potentilla argentea genome (putatively diploid ), were assembled using SOAPdenovo2  on the Abel cluster (hosted by the University of Oslo, Norway). Alignments of the resulting contigs to available Rosaceae sequences at GenBank were used to screen for conserved regions in the markers. Candidate sequences were blasted in Geneious version 10.2  to the Fragaria vesca genome published at Genbank  and to the P. argentea contigs to ensure that they would not amplify multiple regions. Annotation was based on the F. vesca genome (GAPCP1: XM_004306515; GBSSI-1: XM_004306569; DHAR2: XM_004307358).
The P. argentea sequences used in this study were taken from these contigs, and were therefore not produced as the rest of the sequences (see below).
DNA extraction and PCR
Twenty milligrams of silica gel-dried or herbarium leaf material were extracted using the Qiagen DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions, with the exception that the samples were left overnight at 56 °C and then allowed to lyse at 65 °C for 10 min. PCR mixtures included 2.5 μl 10x buffer (Mg2+ plus, 20 mM), 2 μl dNTP (2.5 mM each), 1 μl forward and reverse primers (10 μM), 0.15 μl TaKaRa Ex Taq HotStart DNA polymerase (5 U/μl) (Takara Bio, Shiga, Japan), 1–2 μl template, and ddH2O to add up to 25 μl. The reactions were run on a PCR C1000TM Touch Thermal Cycler (Bio-Rad Laboratories, Hercules, CA, USA). For GAPCP1, the reactions were amplified through 3 min initial denaturation at 95 °C, followed by 35 cycles of denaturation for 30 s at 95 °C, annealing for 30 s at 51 °C and extension for 1 min at 72 °C. A final extension was performed for 5 min at 72 °C. For GBSSI-1 and DHAR2, the reactions were amplified through a touch-down program with 3 min initial denaturation at 94 °C, followed by 10 cycles of denaturation for 45 s at 94 °C, annealing for 30 s starting at 55 °C and then 0.5 °C lower for each cycle, and 60s extension at 72 °C. Thirty-five cycles with a constant annealing temperature at 49 °C followed, and a final extension for 7 min at 72 °C. The reactions were checked on a 1% agarose GelRed-stained (Biotium Inc., Freemont, CA, USA) gel under UV light.
Cloning of PCR products was performed on polyploids and specimens failing direct sequencing, using the StrataClone PCR Cloning kit (Agilent Technologies, Santa Clara, CA, USA) following the manufacturer’s instructions, with the exceptions that 40 and 80 μl of the transformation mixture were plated and that the reaction mixture was halved for species of lower ploidy level (4x). PCR reactions were performed on positive transformants with primers M13–20 and M13 reverse (as found in the manual) together with Ex-Taq HS polymerase as described above. Amplification started with an initial denaturation for 10 min at 94 °C, followed by 35 cycles of denaturation for 45 s at 94 °C, annealing for 45 s at 55 °C and extension for 3 min at 72 °C. A final extension was performed for 10 min at 72 °C. PCR products were then checked on a 1% agarose gel.
Purification and sequencing
All PCR products were purified using the Exo-Sap method . The number of clones sequenced corresponded to 95% probability of finding all gene copies, that is at least 6 clones for tetraploids, 11 clones for hexaploids, 16 clones for octoploids and 21 clones for decaploids (Lundberg et al., unpublished). Two species of Ivesia have been reported to be tetraploids , and therefore the Ivesioid species included in this study were also treated as such. The samples were prepared using a BigDye Terminator Cycle sequencing kit (Applied Biosystems, Waltham, MA, USA) and run on an ABI 3730XL DNA Analyser (Applied Biosystems). For DHAR2, some samples were sent to Macrogen Sequencing Service (Amsterdam, The Netherlands) after purification. All other molecular labwork was carried out at the Biodiversity Laboratories (DNA Section) at the University of Bergen.
Sequence treatment and alignment
For each marker, forward and reverse reads for each specimen or clone were assembled using PreGap4 and Gap4 of the Staden Package . Automatic alignment of each cloned species separately (and specimen, in the case of P. norvegica) was performed in AliView v. 1.18  using MUSCLE . Putative PCR errors were corrected and identical sequences were removed. An alignment with all P. norvegica specimens was then performed, in order to remove identical sequences shared between individuals.
To detect PCR recombinants, the alignments of cloned specimens were loaded into SplitsTree v. 4.14.4 . Sequences identified as putative PCR recombinants had no, or very short, individual edges and long, parallel, connecting edges to their parental sequences . All remaining sequences were automatically aligned together in AliView followed by manual adjustments.
Substitution model testing was performed on each marker with PartitionFinder2 , with GAPCP1 and GBSSI-1 divided into subsets of introns and the three codon positions, under the BIC and AICc criteria for the models available in MrBayes. DHAR2 was not divided into subsets, since the amplified region almost exclusively consists of the intron between exons 4 and 5.
Indels found in two or more sequences were manually coded according to the Simple Indel Coding method as present (1), absent (0) or inapplicable (N) .
Bayesian Inference analyses were run for each marker separately in MrBayes v. 3.2.6 [75, 76], using the Metropolis Coupled Markov Chain Monte Carlo algorithm , including one cold chain and three heated chains for each of two runs. Division of the alignments into subsets and assignment of models were coded according to the results from PartitionFinder2 (Table 1). The Mk model  was applied for the indels, where the likelihood prior Coding and rate prior were set to variable. The analyses were run for 5 million generations for GAPCP1 and GBSSI-1, and 7.5 million generations for DHAR2, with sampling from the chain every 1000th generation and with a burnin of 20%. An analysis was accepted if the standard deviation of split frequencies was below 0.01, the chain swap was between 20 and 80%  (McGuire et al. 2007), no trend was seen in the overlay plot and the Potential Scale Reduction Factor  values had reached 1.0 for all parameters. A clade was fully accepted if its Bayesian posterior probability was 0.95 or higher. In order for the DHAR2 analysis to converge, 13 P. norvegica sequences and one P. intermedia sequence that were suspected to cause problems had to be removed. These were identified by inspecting the whole dataset in SplitsTree. PartitionFinder2 and MrBayes were run at the CIPRES Science Gateway .
Maximum Likelihood analyses were performed in RAxML version 7.2.8 [82, 83]. under the GTR + G (nucleotides, DNA)  and Mk (indels, MULTI)  models with 1000 rapid bootstrap replicates . A clade was fully accepted if its Bootstrap support was 75 or higher.
Rooting and tree graphics
The resulting consensus trees from the BI and ML analyses were inspected using FigTree version 1.4.1  and rooted on P. biflora and P. clusiana Jacq. of the Alba clade. The Alba clade is the sister clade to the rest of the species included in this study [18, 23]. All branches with posterior probabilities below 0.8 were collapsed in Mesquite version 3.10 . The layouts were further edited using GIMP version 2.8.10 (www.gimp.org) and Inkscape version 0.48 (www.inkscape.org).
To ensure that no gene copies were incorrectly discarded as PCR recombinants, all unique sequences of the Ivesioids (Horkelia, Horkeliella and Ivesia), P. intermedia and P. norvegica were subjected to an ML analysis each (without coded indels), together with a reduced dataset of the species representing the larger clades seen in the gene trees.
Multispecies coalescent analysis
Due to initial results from the BI and ML analyses showing somewhat different topologies for the different markers, some species were subjected to a Multispecies Coalescent analysis  in BEAST v. 1.8.0  at CIPRES , in order to create a species tree as a backbone reference. Two datasets were created, one with the three low-copy markers only, and one with the low-copy markers in combination with three chloroplast regions from previous studies (trnL-F, trnC-ycf6 and trnS-ycf9) (Table 4) [18, 44, 90]. Substitution model testing was performed in PartitonFinder2 on each region, not accounting for codon positions. Two clock models were tested; strict and relaxed uncorrelated log normal . For each of these, two tree priors were tested; a birth-death process  and a birth process . The analysis of the dataset with low-copy markers only was run for 50 million generations with sampling every 1000th generation, and the combined dataset for 150 million generations with sampling every 1000th generation. To test the fit of the models to the data, path sampling and stepping-stone sampling [94, 95] were performed with 50 steps, each with a length of 1 million iterations for the low-copy marker dataset, and 150 steps with a length of 1 million iterations for the combined dataset. Log marginal likelihood differences larger than three were considered significant . The analysis with the models best fit to the data was run two independent times, and the results were inspected using Tracer v. 1.7.1 . In order to test if the prior, rather than the data, was driving the results, an additional run with sampling from prior only was performed. The tree files were then combined using TreeAnnotator of the BEAST package with a burnin of 20% of each run.
Potentilla norvegica specimens were inspected at, or on loan from, the herbaria of Stockholm (S), Uppsala (UPS) and Gothenburg (GB) in Sweden, and the herbaria of Bergen (BG) and Oslo (O) in Norway. They were used to study the defining characters of the two P. norvegica subspecies (ssp. norvegica and ssp. hirsuta); leaflet form, leaflet dentation and stipule dentation [29, 38, 98] (Fig. 6 and Table 5).
Availability of data and materials
Most specimens included are deposited at herbaria (see Table 1). Vouchers are missing for a few of the specimens obtained from botanical gardens. The DNA sequences are deposited at GenBank under the accession numbers [MN346707-MN346962].
Werth CR, Guttman SI, Eshbaugh WH. Electrophoretic evidence of reticulate evolution in the Appalachian Asplenium complex. Syst Bot. 1985;10(2):184–92.
Takamiya M, Tanaka R. Polyploid cytotypes and their habitat preferences in Lycopodium clavatum. Bot Mag Tokyo. 1982;95:419–34.
Ownbey M. Natural hybridization and amphiploidy in the genus Tragopogon. Am J Bot. 1950;37(7):487–99.
Kihara H, Ono T. Chromosomenzahlen und systematische Gruppierung der Rumex-Arten. Z Zellforsch Mik Ana. 1926;4:475–81.
Stebbins GL. Variation and evolution in plants. Newy York: Columbia University Press; 1950.
Grant V. Plant speciation. 1st ed. New York: Columbia University Press; 1971.
Rieseberg LH, Willis JH. Plant speciation. Science. 2007;317(5840):910–4.
de Queiroz K. A unified concept of species and its consequences for the future of taxonomy. Proc Calif Acad Sci. 2005;56(18):196–215.
Evans RC, Campbell CS. The origin of the apple subfamily (Maliodeae; Rosaceae) is clarified by DNA sequence data from duplicated GBSSI genes. Am J Bot. 2002;89(9):1478–84.
Lundberg M, Töpel M, Eriksen B, Nylander JAA, Eriksson T. Allopolyploidy in Fragariinae (Rosaceae): comparing four DNA sequence regions, with comments on classification. Mol Phylogenet Evol. 2009;51:269–80.
Kessler M, Kühn A, Solís Neffa VG, Hensen I. Complex geographical distribution of ploidy levels in Polylepis australis (Rosaceae), an endemic tree line species in Argentina. Int J Plant Sci. 2014;175(8):955–61.
Vamosi JC, Dickinson TA. Polyploidy and diversification: a phylogenetic investigation in Rosaceae. Int J Plant Sci. 2006;167(2):349–58.
Rice A, Glick L, Abadi S, Einhorn M, Kopelman NM, Salman-Minkov A, Mayzel J, Chay O, Mayrose I. The Chromosome Counts Database (CCDB) – a community resource of plant chromosome numbers. New Phytol. 2014;206(1):19–26.
Eriksson T, Donoghue MJ, Hibbs MS. Phylogenetic analysis of Potentilla using DNA sequences of nuclear ribosomal internal transcribed spacers (ITS), and implications for the classification of Rosoideae (Rosaceae). Pl Syst Evol. 1998;211:155–79.
Eriksson T, Hibbs MS, Yoder AD, Delwiche CF, Donoghue MJ. The phylogeny of Rosoideae (Rosaceae) based on sequences of the internal transcribed spacers (ITS) of nuclear ribosomal DNA and the trnL/F region of chloroplast DNA. Int J Plant Sci. 2003;164(2):197–211.
Eriksson T, Lundberg M, Töpel M, Östensson P, Smedmark JEE. Sibbaldia: a molecular phylogenetic study of a remarkably polyphyletic genus in Rosaceae. Pl Syst Evol. 2015;301:171–84.
Ertter B, Elven R, Reveal JL, Murray DF. Potentilla. In: Flora of North America editorial committee, editors. 1993+. Flora of North America north of Mexico. 20+ vols, vol. 9. New York: Oxford University Press; 2014. p. 121–218.
Dobeš C, Paule J. A comprehensive chloroplast DNA-based phylogeny of the genus Potentilla (Rosaceae): implications for its geographic origin, phylogeography and generic circumscription. Mol Phylogenet Evol. 2010;56:156–75.
Soják J. Argentina Hill., a genus distinct from Potentilla (Rosaceae). Thaiszia - J Bot. 2010;20:91–7.
Feng T, Moore MJ, Li J, Wang H. Phylogenetic study of the tribe Potentilleae (Rosaceae), with further insight into the disintegration of Sibbaldia. J Syst Evol. 2017;55(3):177–91.
Ertter B, Reveal JL. Ivesia, Horkelia & Horkeliella. In: Flora of North America editorial committee, editor. editors. 1993+ Flora of North America north of Mexico. 20+ vols, vol. 9. New York: Oxford University Press; 2014a. p. 219–72.
Ertter B, Reveal JL. Duchesnea. In: Flora of North America editorial committee, editor. 1993+ Flora of North America north of Mexico. 20+ vols, vol. 9. New York: Oxford University Press; 2014b. p. 275–6.
Töpel M, Lundberg M, Eriksson T, Eriksen B. Molecular data and ploidal levels indicate several putative allopolyploidization events in the genus Potentilla (Rosaceae). PLoS Curr. 2011;3:RRN1237.
Wendel JF, Doyle JJ. Phylogenetic incongruence: Window into genome history and molecular evolution. In: Soltis P, Soltis D, Doyle J, editors. Molecular systematics of plants. II. Dodrecht: Kluwer Academic Press; 1998. p. 265–96.
Wolf T. Monographie der Gattung Potentilla. Bibliotheca Botanica. 1908;16(71):1–714.
Linnaeus C. Species plantarum. Stockholm: Salvius; 1753. Vol. 1, p. 1109.
Michaux A. Flora Boreali Americana. Paris. 1803;1:303.
Ledebour CF. Flora Rossica, vol. 2. Stuttgart: Sumptibus Librariae E. Schweizerbart; 1844. p. 36.
Ascherson P, Graebner P. Synopsis der Mitteleuropäischen Flora. Leipzig: Engelmann; 1904. Vol. 6 pt.1 p. 748.
IPNI - International Plant Names Index. The Royal Botanic Gardens, Kew, Harvard University Herbaria & Libraries and Australian National Botanic Gardens. http://www.ipni.org. Accessed 14 Sept 2018.
Tropicos. Missouri Botanical Garden. http://www.tropicos.org/Name/27802541. Accessed 15 Sept 2018.
Hylander N. Studien über nordische Gefässpflanzen. Uppsala Univ Årsskr. 1945;1945(7):203.
Turland NJ, Wiersema JH, Barrie FR, Greuter W, Hawksworth DL, Herendeen PS, Knapp S, Kusber W-H, Li D, Marhold K, May TW, McNeill J, Monro AM, Prado J, Price MJ, Smith GF. International Code of Nomenclature for algae, fungi, and plants (Shenzhen Code) adopted by the Nineteenth International Botanical Congress Shenzhen, China, July 2017. Regnum Vegetabile 159. Glashütten: Koeltz Botanical Books; 2018.
Hitchcock CL, Conquist A. Vascular plants of the Pacific northwest. Part 2: Salicaceae to Saxifragaceae. Seattle: University of Washington Press; 1968.
Tutin TG, Heywood VH, Burges NA, Valentine DH, Walters SM, Webb DA. Flora Europaea. Rosaceae to Umbelliferae, vol. 2. London: Cambridge University Press; 1968. p. 42.
Hultén E. The circumpolar plants. II – Dicotyledons. Kungliga Vetenskapsakademiens Handlingar. Fjärde serien, band 13, nr. 1. Stockholm: Almqvist & Wiksell; 1971.
Kurtto A, Lampinen R, Junikka L. Atlas Florae Europaeae. Distribution of vascular plants in Europe, vol. 13. Helsinki: Societas Biologica Fennica Vanamo; 2004.
Lid J, Lid DT. Potentilla L. In: Elven R, editor. Norsk flora. 7th ed. Oslo: Det Norske Samlaget; 2013. p. 413–23.
Small RL, Cronn RC, Wendel JF. Use of nuclear genes for phylogeny reconstruction in plants. Aust Syst Bot. 2004;17:145–70.
Smedmark J, Eriksson T, Bremer B. Allopolyploid evolution in Geinae (Colurieae: Rosaceae) – building reticulate species trees from bifurcating gene trees. Org Divers Evol. 2005;5:275–83.
Maddison WP. Gene trees in species trees. Syst Biol. 1997;46(3):523–36.
Hasegawa M, Kishino H, Yano T. Dating of human-ape splitting by a molecular clock of mitochondrial DNA. J Mol Evol. 1985;22(2):160–74.
Rydberg KA. A monograph of the North American Potentilleae, vol. 2. Lancaster: The New Era Printing Company; 1898. p. 47.
Töpel M, Antonelli A, Yesson C, Eriksen B. Past climate change and plant evolution in Western North America: a case study in Rosaceae. PLoS One. 2012;7(12):e50358.
Comai L. The advantages and disadvantages of being polyploid. Nat Rev Genet. 2005;6:836–46.
Leitch IJ, Bennett MD. Genome downsizing in polyploids. Biol J Linn Soc. 2004;82:651–63.
Bento M, Gustafsson JP, Viegas W, Silva M. Size matters in Triticae polyploids: larger genomes have higher remodeling. Genome. 2011;54:175–83.
Xiang Y, Huang C, Hu Y, Wen J, Li S, Tingshuang Y, Chen H, Xiang J, Ma H. Evolution of Rosaceae fruit types based on nuclear phylogeny in the context of geological times and genome duplication. Mol Biol Evol. 2017;34(2):262–81.
Cohen KM, Finney SC, Gibbard PL, Fan J. The ICS international Chronostratigraphic chart. Episodes. 2013;36(3):199–204.
Dietz RS, Holden JC. Reconstruction of Pangaea: breakup and dispersion of continents, Permian to present. J Geophys Res. 1970;75(26):4939–56.
Tiffney BH. The Eocene North Atlantic land bridge: its importance in tertiary and modern phytogeography of the northern hemisphere. J Arnold Arboretum. 1985;66:243–73.
Hopkins DM. Cenozoic history of the Bering land bridge. Science. 1959;129(3362):1519–28.
Tiffney BH, Manchester SR. The use of geological and paleontological evidence in evaluating plant phylogeographic hypotheses in the northern hemisphere tertiary. Int J Plant Sci. 2001;162(6):3–17.
You Y, Huber M, Müller RD, Poulsen CJ, Ribbe J. Simulation of the middle Miocene climate optimum. Geophys Res Lett. 2009;36:L04702.
Eriksen B. Mating systems in two species of Potentilla from Alaska. Folia Geobot Phytotx. 1996;31:333–44.
Renny-Byfield S, Ainouche M, Leitch IJ, Lim Y, Le Comber SC, Leitch AR. Flow cytometry and GISH reveal mixed ploidy populations and Spartina nonaploids with genomes of S. alterniflora and S. maritima origin. Ann Bot-London. 2010;105(4):527–33.
Čertner M, Fenclová E, Kúr P, Kolář F, Koutecký P, Krahulcova A, Suda J. Evolutionary dynamics of mixed-ploidy populations in an annual herb: dispersal, local persistence and recurrent origins of polyploids. Ann Bot-London. 2017;120:303–15.
Müntzing A. Heteroploidy and polymorphism in some apomictic species of Potentilla. Hereditas. 1958;44(2–3):280–329.
Asker S. Apomictic biotypes in Potentilla intermedia and P. norvegica. Hereditas. 1970;66(1):101–7.
GBIF. Global Biodiversity Information Facility Home Page. 2018. https://www.gbif.org. Accessed 2 July 2018.
IPCN - Index to Plant Chromosome Numbers. Goldblatt, P., Johnson, D. E. editors. Missouri Botanical Garden, St. Louis. 1979--. http://www.tropicos.org/Project/IPCN. Accessed 7 May 2017.
Le Dantec L, Cardinet G, Bonet J, Fouché M, Boudehri K, Monfort A, Poëssel J, Moing A, Dirlewanger E. Development and mapping of peach candidate genes involved in fruit quality and their transferability and potential use in other Rosaceae species. Tree Genet Genomes. 2010;6:995–1012.
Paule J, Sharbel TF, Dobeš C. Apomictic and sexual lineages of the Potentilla argentea L. group (Rosaceae): Cytotype and molecular genetic differentiation. Taxon. 2011;60(3):721–32.
Luo R, Liu B, Xie Y, Li Z, Huang W, Yuan J, He G, Chen Y, Pan Q, Liu Y, Tang J, Wu G, Zhang H, Shi Y, Liu Y, Yu C, Wang B, Lu Y, Han C, Cheung DW, Yiu S, Peng S, Xiaoqian Z, Liu G, Liao X, Li Y, Yang H, Wang J, Lam T, Wang J. SOAPdenovo2: an empirically improved memory-efficient short-read de novo assembler. GigaScience. 2012;1:18.
Markowitz S, Duran C, Thierer T, Ashton B, Mentjies P, Drummond A. Geneious basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics. 2012;28(12):1647–9.
Shulaev V, Sargent DJ, Crowhurst RN, Mockler TC, Folkerts O, Delcher AL, Jaiswal P, Mockaitis K, Liston A, Mane SP, Burns P, Davis TM, Slovin JP, Bassil N, Hellens RP, Evans C, Harkins T, Kodira C, Desany B, Crasta OR, Jensen RV, Allan AC, Michael TP, Setubal JC, Celton J, DJG R, Williams KP, Holt SH, Ruiz Rojas JJ, Chatterjee M, Liu B, Silva H, Meisel L, Adato A, Filichkin SA, Troggio M, Viola R, Ashman T, Wang H, Dharmawardhana P, Elser J, Raja R, Priest HD, Bryant DW Jr, Fox SE, Givan SA, Wilhelm LJ, Naithani S, Christoffels AY, Salama DY, Carter J, Lopez Girona E, Zdepski A, Wang W, Kerstetter RA, Schwab W, Korban SS, Davik J, Monfort A, Denoyes-Rothan B, Arus P, Mittler R, Flinn B, Aharoni A, Bennetzen JL, Salzberg SL, Dickerman AW, Velasco R, Borodovsky M, Veilleux RE, Folta KM. The genome of woodland strawberry (Fragaria vesca). Nat Genet. 2011;43(2):109–16.
Dugan K, Lawrence H, Hares D, Fisher C, Budowle B. An improved method for post-PCR purification for mtDNA sequence analysis. J Forensic Sci. 2002;47(4):1–8.
Staden R. The Staden sequence analysis package. Mol Biotechnol. 1996;5:233.
Larsson A. AliView: a fast and lightweight alignment viewer and editor for large datasets. Bioinformatics. 2014;30(22):3276–8.
Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32(5):1792–7.
Huson DH, Bryant D. Application of phylogenetic networks in evolutionary studies. Mol Biol Evol. 2006;23(2):254–67.
Marcussen T, Heier L, Brysting AK, Oxelman B, Jakobsen KS. From gene trees to a dated allopolyploid network: insights from the angiosperm genus Viola (Violaceae). Syst Biol. 2015;64(1):84–101.
Lanfear R, Frandsen PB, Wright AM, Senfeld T, Calcott B. PartitionFinder 2: new methods for selecting partitioned models of evolution for molecular and morphological phylogenetic analyses. Mol Biol Evol. 2016;34(3):772–3.
Simmons MP, Ochoterena H. Gaps as characters in sequence-based phylogenetic analyses. Syst Biol. 2000;49(2):369–81.
Huelsenbeck JP, Ronquist F. MRBAYES: Bayesian inference of phylogeny. Bioinformatics. 2001;17:754–5.
Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, Larget B, Liu L, Suchard MA, Huelsenbeck JP. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012;61(3):539–42.
Altekar G, Dwarkadas S, Huelsenbeck JP, Ronquist F. Parallel Metropolis-coupled Markov chain Monte Carlo for Bayesian phylogenetic inference. Bioinformatics. 2004;20:407–15.
Lewis PO. A likelihood approach to estimating phylogeny from discrete morphological character data. Syst Biol. 2001;50(6):913–25.
McGuire JA, Witt CC, Altshuler DL, Remsen JV. Phylogenetic systematics and biogeography of humming-birds: Bayesian and maximum likelihood analyses of partitioned data and selection of an appropriate partitioning strategy. Syst Biol. 2007;56:837–56.
Gelman A, Rubin DB. Inference from iterative simulation using multiple sequences. Stat Sci. 1992;7(4):457–72.
Miller MA, Pfeiffer W, Schwartz T. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Proceedings of the Gateway Computing Environments Workshop (GCE). New Orleans: Heidelberg Institute for Theoretical Studies; 2010. p. 1–8.
Stamatakis A. RaxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics. 2006;22(21):2688–90.
Stamatakis A. RaxML version 7.2.8. 2010. https://sco.h-its.org/exelixis/web/software/raxml/.
Tavaré S. Some probabilistic and statistical problems in the analysis of DNA sequences. Lectures on mathematics in the life sciences. Am Mathematical Soc. 1986;17:57–86.
Stamatakis A, Hoover P, Rougemont J. A rapid bootstrap algorithm for the RaxML web servers. Syst Biol. 2008;57(5):758–71.
Rambaut A. FigTree: Tree figure drawing tool version 1.4.2. http://tree.bio.ed.ac.uk/software/figtree: Institute of Evolutionary Biology, University of Edinburgh; 2014.
Maddison WP, Maddison DR. Mesquite: a modular system for evolutionary analysis. Version 3.10. 2016. http://mesquiteproject.org.
Heled J, Drummond AJ. Bayesian inference of species trees from multilocus data. Mol Biol Evol. 2009;27(3):570–80.
Drummond AJ, Suchard MA, Xie D, Rambaut A. Bayesian phylogenetics with BEAUTi and the BEAST 1.7. Mol Biol Evol. 2012;29:1969–73.
Koski MH, Ashman T. Macroevolutionary patterns of ultraviolet floral pigmentation explained by geography and associated bioclimatic factors. New Phytol. 2016;211(2):708–18.
Drummond AJ, Ho SYW, Phillips MJ, Rambaut A. Relaxed phylogenetics and dating with confidence. PLoS Biol. 2006;4(5):e88.
Kendall DG. On the generalized “birth-and-death” process. Ann Math Stat. 1948;19:1–15.
Yule GU. A mathematical theory of evolution, based on the conclusions of Dr. J. C. Willis, F.R.S. Philos T Roy Soc B. 1925;213:21–87.
Baele G, Lemey P, Bedford T, Rambaut A, Suchard MA, Alekseyenko AV. Improving the accuracy of demographic and molecular clock model comparison while accommodating phylogenetic uncertainty. Mol Biol Evol. 2012;29:2157–67.
Baele G, Li WLS, Drummond AJ, Suchard MA, Lemey P. Accurate model selection of relaxed molecular clocks in Bayesian phylogenetics. Mol Biol Evol. 2013;30:239–43.
Kass RE, Raftery AE. Bayes factors. J Am Stat Assoc. 1995;90:773–95.
Rambaut A, Drummond AJ, Xie D, Baele G, Suchard MA. Posterior summarisation in Bayesian phylogenetics using tracer 1.7. Syst Biol. 2018;67(5):901–4.
Mossberg B, Stenberg L. Gyldendals store nordiske flora. 2nd ed: Hung Hing: Gyldendal; 2014.
We are grateful to the curators and other staff at the herbaria at the University Museum of Bergen (BG), University of Gothenburg (GB), University of Oslo (O), Swedish Museum of Natural History (S), Uppsala University (UPS), Natural History Museum, Vienna (W) and University of Vienna (WU) for assistance during visits, field trips and with loans; and the technicians at the Biodiversity Laboratory at the University of Bergen for help with the labwork. We also thank the other members of the Rosaceae research group at UiB for fruitful discussions, especially Ingrid Toresen who produced several of the GAPCP1 sequences. The unpublished raw reads of the Potentilla argentea genome were kindly provided by Malene Østreng Nygård and Mika Bendiksby at the Norwegian University of Science and Technology, Trondheim.
This work was supported by the University Museum of Bergen and the Olaf Grolle Olsen fund at the University of Bergen to NLP, TE and JEES. The funding sources had no participation in the design or execution of this study.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Persson, N.L., Eriksson, T. & Smedmark, J.E.E. Complex patterns of reticulate evolution in opportunistic weeds (Potentilla L., Rosaceae), as revealed by low-copy nuclear markers. BMC Evol Biol 20, 38 (2020). https://doi.org/10.1186/s12862-020-1597-7
- Low-copy nuclear markers
- Molecular cloning
- Molecular phylogeny
- Reticulate evolution