Using species distribution models to locate the potential cradles of the allopolyploid Gypsophila bermejoi G. López (Caryophyllaceae)

Miguel de Luis; Julio Álvarez-Jiménez; Francisco Javier Rejos; Carmen Bartolomé

doi:10.1371/journal.pone.0232736

. 2020 May 19;15(5):e0232736. doi: 10.1371/journal.pone.0232736

Using species distribution models to locate the potential cradles of the allopolyploid Gypsophila bermejoi G. López (Caryophyllaceae)

Miguel de Luis ¹, Julio Álvarez-Jiménez ¹, Francisco Javier Rejos ¹, Carmen Bartolomé ^1,^*

Editor: Daniel de Paiva Silva²

PMCID: PMC7237017 PMID: 32428047

Abstract

Polyploidy has been an influential force in plant evolution, playing a crucial role in diversification. Differences in polyploid and diploid distributions have been long noted, with polyploid taxa especially abundant in harsh environments. These plants have higher photosynthetic rates and/or higher tolerance to water deficits. Moreover, there is data pointing to an increase in the rate of unreduced gamete formation by plants under conditions of stress. Accordingly, a higher frequency of polyploid individuals would be expected in populations living under extreme environments, a phenomenon that may be relevant when considering the origin of allopolyploid species. Hybridization between distinct autopolyploids is known to produce allopolyploids and hence, a high frequency of compatible autopolyploids in an area could enhance the formation of stable populations of the corresponding allopolyploid hybrid. Here we consider the allopolyploid species Gypsophila bermejoi G. López and its parental taxa G. struthium L. subsp. struthium and G. tomentosa L. We have used Species Distribution Models to locate areas with low bioclimatic suitability for both parental taxa during the Last Glacial Maximum (LGM), hypothesizing that the rate of tetraploid hybrid formation would be higher than expected where low suitability areas of both parental species overlap. We selected those areas taking into account the strict gypsophyllic nature of these taxa. There is data pointing to a post-glacial origin of the current G. bermejoi populations and according to our hypothesis, such locations could be centers for hybrid tetraploid formation or potential cradles of this species. Indeed, potential Mid-Holocene cradles were also identified in this manner. The evolution of bioclimatic suitability in both LGM and Mid-Holocene cradles was studied to assess the possible survival of the hybrids, and the current distribution of G. bermejoi proved to be consistent with our hypothesis.

1 Introduction

Polyploidy or Whole Genome Duplication (WGD) is an ubiquitous and influential force in plant evolution, both on macro- and micro-evolutionary timescales. [1, 2] WGD could have played a crucial role in angiosperm evolution and in the diversification of their main lineages. Indeed, some ancient WGD events are associated with high rates of diversification, a remarkable example of which is related to the origin of Asteraceae, a family of angiosperms containing one of the largest number of species [3–6].

Polyploidy can arise in two ways. Allopolyploidy takes place when the genomes are derived from different species, individuals that can be produced by interspecific hybridization. By contrast, autopolyploidy is the result of intraspecific WGD events and it has long been considered to be less common, although it is true that autopolyploids are more difficult to detect. Moreover, if the rate of tetraploid formation were comparable to the mutation rate, many more autopolyploids would be expected in nature [7,8].

Differences in the distribution of polyploids and diploids have long been noted [9]. Polyploid taxa are frequently found in plant communities worldwide and they seem especially abundant in challenging environments, such as at high latitudes, at high elevations or in arctic areas [9–14]. Such distributions are consistent with observations regarding the physiological changes in polyploid plants, and when changes in body size follow genome duplication, it is common to find that “morphological differences could result in different adaptations of cytotypes” (emphasis ours) [15,16]. Other central processes in plant physiology may also be affected by genome duplication and there is evidence that polyploids display a higher tolerance to water deficit [17] or higher photosynthetic rates [18]. One recurrent observation in natural polyploid plants is that they usually show improved growth and better adaptation to harsh environments, consequently displaying evolutionary advantages [19]. These observations explain the higher frequency of polyploids proposed in extreme environments [20].It is known that the rate of unreduced gamete formation increases under conditions of stress in plants [21, 22], such as extreme cold, highly fluctuating temperatures, nutritional and water stress, amongst others [21]. In areas where such challenging conditions prevail, small populations would be expected and this circumstance could further enhance the establishment of polyploids [23]. Interestingly, polyploids often have a higher rate of self-fertilization, a higher capacity to propagate clonally and a perennial life-cycle. All these traits would improve their chances of establishing stable populations in areas where conditions are non-optimal [20,24,25]. The higher frequency of polyploid individuals in non-optimal areas or extreme bioclimatic conditions, if true, could be relevant to the origin of allopolyploid species. Moreover, it is known that hybridization between distinct autopolyploids directly produces allopolyploids [21] and according to this observation, the presence of a high frequency of compatible autopolyploids in an area could trigger the formation of stable populations of the corresponding allopolyploid hybrid.

Ecological Niche Models (ENMs) along with Species Distribution Models (SDMs) [26] are useful tools to visualize the ecological preferences of a species within a specific area and they have been used successfully to examine a wide range of ecological issues. In a number of studies they have also been used to evaluate the effects of climate change on ecosystems and communities, or the effectiveness of conservation plans [27–36]. In terms of species distribution models (SDMs), all the above observations could mean that the frequency of polyploid individuals increases at the periphery of species ranges [37–40], where the suitability for the species is low but not zero. Such Low Suitability Areas (LSAs) would harbor populations with a higher frequency of autopolyploids. If we consider two interbreeding species, we would expect a higher formation of allopolyploids in areas where the periphery of both ranges overlap. If there is a higher frequency of autopolyploids, the probability of producing non-reduced gametes would also be higher [22], increasing the probability of fertile interspecific hybrids arising. Thus, we could consider these areas as more favorable sites to establish stable allopolyploid populations (Fig 1), and such locations could act as potential formation centers or cradles for the hybrid species.

Fig 1 — We show two closely related species with an allopatric distribution. Both of them have a high suitability area around the centroid of their ranges (regions A and B in 1) and a lower suitability area (LSA) around their ranges (areas a and b in 1). According to a number of observations (see text), a higher frequency of autopolyploids (solid squares and circles) could be expected in the peripheral regions of both ranges (the LSAs). The LSAs will overlap if both species have some degree of sympatry and hence, there would be regions with a stronger presence of autopolyploids emanating from both species. This circumstance would favor the formation of fertile tetraploid hybrids (red triangles, species C in 3). We considered such areas as potential cradles or potential formation centers for the hybrid species. Such hybrids could also arise at other locations where the parental taxa could interbreed, although a lower rate of hybrid formation would be expected. This is a possible scenario for the speciation of G. *bermejoi*, where the A and B species would be its parental G. *struthium* subsp. *struthium* and G. *tomentosa*.

Gypsophila bermejoi G.López is an allopolyploid species derived from the parental G. struthium L. subsp. struthium and G. tomentosa L [41]. All these plants are gypsophytes endemic to the Iberian Peninsula, and they are of particular ecological, evolutionary and biochemical interest. Some evidence points to a recent origin of the current populations of G. bermejoi [42,43]. Indeed, based on ENMs and SDMs it has been shown that the parental taxa of G. bermejoi had a high degree of sympatry during the Last Glacial Maximum (LGM), although the climatic conditions may have repressed the formation of this species because of its very narrow bioclimatic niche at that time [42,43]. In the present work gypsum soil locations included in areas where the LSAs of the parental species of G. bermejoi overlap were determined. According to our hypothesis such locations could be potential cradles for G. bermejoi speciation (Fig 1) Thus, we located a set possible centers where the climatic conditions during the (LGM) and the mid-Holocene (MH) period would favor the formation of this species, and then, we assessed the evolution of the bioclimatic conditions at those locations to determine which populations could have survived. Finally, we compared these results with the current locations of G. bermejoi. Our main goal was to find out if this working hypothesis could to some extent predict the current distribution of the species.

2 Material and methods

2.1 Study area and species

All the species included in this work are endemic to the Iberian Peninsula, which is located in the southwestern part of Europe, between latitudes 36°00'08"N—43°47'38"N and longitudes 9°29'00"O—3°19'00"E. With an area of approximately 597,000 km², this is the second largest peninsula on the continent, and it includes most of the territory of Spain and Portugal. In this study, we focused on the gypsum steppes within the Peninsula, containing the gypsum outcrops in this territory and covering a total extension of 32,487 km² (5.4% of the Peninsula) [44]. These habitats are concentrated in the eastern half of the Peninsula, under the influence of a Mediterranean climate.

Our research focuses on the speciation of Gypsophila bermejoi, a natural hybrid derived from G. struthium subsp. struthium (2n = 34) and G. tomentosa (2n = 34) [41]. This genus is included in the Caryophyllaceae family of plants, and it is an allopolyploid species (2n = 68) found on verges and slopes in areas of gypsum soil [44]. While the range of G. bermejoi overlaps that of its parental taxa to some degree [45], the ecological niche of G. bermejoi shows noticeable differences [43]. Gypsum soils are harsh environments for angiosperms due to their extreme xericity, severe nutritional imbalances and the presence of a crusty surface [45,46]. In biochemical terms, gypsophytes can produce a wide range of secondary metabolites and they have a tendency to accumulate certain minerals. The production of saponins and antioxidants, like that of phenols and flavonoids, has been studied widely in Gypsophila [47–49]. As a result of the particular ecological conditions of the gypsisols, the peninsular gypsum outcrops harbor plants that are highly adapted to their environment, and endemic species are common at these sites. This circumstance and the patchy nature of these habitats makes these species a good model to study plant evolution and speciation processes [46].

2.2 Data sources

According to a number of works [50,51] the use of only one algorithm is not the ideal strategy to model ecological niches and potential species distributions. In this research we performed ensambles based on four algorithms: General Linear Models (GLM) [52], Random Forests (RF) [53–55], Mlutivariate Adaptative Regression Splines (MARS) [56,57] and Maximum Entropy (MaxEnt) [58–61]. All these algorithms require two types of data for the study area: species occurrence and environmental data. Here, we used species occurrence data from the Global Biodiversity Information Facility (GBIF) database [62] and a set of WorldClim bioclimatic variables. WorldClim is a database with sets of global climatic data that can be used for ecological modeling and mapping (gridded climate data), with a spatial resolution of about 1 km². The internet portal (www.worldclim.org) offers publically data for 19 bioclimatic variables all them derived from temperature and precipitation data from meteorological stations. [63].

An important aspect of ENM and SDM studies is the uncertainty carried in the data, especially when extrapolating models for future or past conditions, since this data is a result from Global Circulation Models (GCMs). There is considerable variation for some of the bioclimatic variables among the different GCMs. Thus, we used bioclimatic maps for the MH period and LGM generated through three different Global Circulation Model: the Community Climate System Model (CCSM4), the Model for interdisciplinary Research of Climate (MIROC-ESM) and the Max Planck Institute for Meteorology Earth System Model (MPI-ESM) [64–66]. In projections to the LGM, this is particularly important for species highly sensitive to precipitation variables in temperate areas [67, 68]

Modeling species distributions requires correlations between the predictors to be minimized [61]. From a set of available predictors, we selected those with the lowest Pearson correlation coefficient using a clustering algorithm to produce a dendrogram that groups the variables according to their degree of correlation. A jackknife test of the contribution of the variables to the model was also performed. After combining the information obtained through correlation dendrogram and jackknife for variable contribution, and taking into account the abiotic factors determining the vegetation patterns in a semi-arid Mediterranean landscape [69,70], we selected from the less correlated variables those with the highest contribution, and one related with precipitation values during the dry season (Table 1).

Table 1. List of the environmental variables selected to implement both the MaxEnt species distribution models.

Variable	Source	Resolution
bio4—the dispersion of seasonal temperatures	WorldClim	2.5’
bio6—the minimum temperature in the coldest month	WorldClim	2.5’
bio13—the rainfall in the wettest month	WorldClim	2.5’
bio14—the rainfall in the driest month	WorldClim	2.5’

Open in a new tab

2.3 Modeling species distributions

We have used the SDM R package [71,72] to model the potential distribution maps of the studied species. It provides a reproducible and extensible platform for species distribution modeling, suitable for reproducible research. This package includes tools enabling ensembles of models to be performed, model evaluation using different approaches, and SDM projections in space and time [72]. In our case, we used a set of algorithms that included General Linear Models (GLM), Random Forests (RF), Mlutivariate Adaptative Regression Splines (MARS) and Maximum Entropy (MaxEnt), all them currently implemented in R language [71]. For an explication of the methods see [52,61]. These tools essentially assess the probability of the presence of a species in function of the environmental variables selected. The inputs are data files on the presence of the species and a series of maps of the environmental variables considered to be significant. The results can be interpreted as an index of suitability for the particular species to that habitat [59].

The models were fitted and used to generate multiple replicates (30) of each method through partitioning using the bootstrap method. Weighted average ensambles of these models were performed for each species. Then, we obtained species distribution projections for the LGM and the Mid Holocene climatic conditions using three different General Circulation Models: CCSMS, MIROC-ESM and MPI-ESM. Again, an ensemble was performed to get the consensus forecast map.

We dealt with autocorrelation issues by eliminating any redundant presence in each pixel on the scale of the bioclimatic variables used. The original numbers of occurrence records and the numbers left after this step (at a resolution of 2.5 minutes) were: 138 and 60 respectively for G. bermejoi, 256 and 120 for G. struthium subsp. struthium and 349 and 99 for G. tomentosa. The uncertainties were assessed and mapped by studying the Standard Deviation of the ensembles (see suplementary materials 1 and 2). The study focused on the effect of climatic oscillations and as such, we used only bioclimatic variables to generate the models. Given the strict gypsophyllic nature of these plants [44], they are only present on gypsum soils and their suitability on other soils is “0”. Therefore, we applied the models with climate variables to the entire Iberian Peninsula and we then subtracted the gypsum soil territories to evaluate the suitability areas. We validated the models by means of the Area Under the Curve (AUC), and theTrue Skill Statistic (TSS), [73–75].

2.4 An approach to find potential G. bermejoi cradles

As pointed out previously, a number of observations seem to support the hypothesis of a higher frequency of polyploid individuals in the peripheral domains of the ranges for a given species [37–40], where the suitability for the species is weaker. If this statement is true, higher autopolyploid frequencies would be expected at the periphery of the potential areas predicted by the models. Here we used this idea to predict possible centers of G. bermejoi formation. As indicated, our hypothesis is that the gypsum outcrops where the LSAs of both parental taxa overlap (suitable soils with higher frequencies of G. struthium subsp. struthium and G. tomentosa autopolyploids) would be potential cradles for the allopolyploid species G. bermejoi. As such, we assessed whether the results obtained were consistent with the current distribution of G. bermejoi.

Our approach to find potential cradles is summarized as:

a Generating species distribution models (SDM)

The ensemble modeling approach was used to produce suitability forecasts based on a series of bioclimatic variables carefully selected (see above) and four different algorithms (GLM, RF, MARS and MaxEnt). These models were then projected onto the geographical space of the Iberian Peninsula in the current climatic conditions, and in two past climatic ages: the LGM and MH period. For the LGM and MH we again performed an ensemble based on three General Circulation Models (CCSMS, MIROC-ESM and MPI-ESM). We generated these models for G. bermejoi and for its parental taxa, G. struthium subsp. struthium and G. tomentosa.

b Obtaining a statistical distribution of suitability values

To achieve our goal it is crucial to define what is “low” suitability. To do so, we carried out an exploratory analysis of the suitability values predicted by the models under the current climatic conditions for both parental taxa. A statistical distribution for these bioclimatic suitability values was obtained, considering only the sites where they are currently present. In other words, we took the readings of the suitability values predicted by the models at the locations where the species is currently found. Our goal here was to define the lowest suitability values predicted by the model at the sites where both taxa occur. (G. struthium subsp. struthium and G. tomentosa). It is important to bear in mind that by doing so, we obtain the lowest suitability values predicted by the SDM for the parental taxa at locations where those plants really grow (Fig 2A).

To establish the intervals of low suitability, we need to know the suitability values predicted by the models at the locations where these taxa are present under current climatic conditions. For each taxon, we considered low suitability values to be all those between the minimum and the first quartile. We tried to assess the suitability values in which a high frequency of autopolyploids is expected for both species. In Fig 2A, the suitability value histograms for both taxa are shown, with the minimum and first quartiles indicated by colored lines. In interval 1 (the minimum between G. *tomentosa* and G. *struthium* subsp. *struthium*), there is low suitability for G. *tomentosa* but no presence of G. *struthium* subsp. *struthium* (making interbreeding impossible). In interval 2 (between the minimum and the first quartile of G. struthium subsp. struthium) there is low suitability for both taxa and according to our hypothesis, a higher frequency of autoplyploids might be expected. Interval number 3 has low suitability for G. *tomentosa* but not for G. *struthium* subsp. *struthium*. If we accept the situations for low suitability, there could be a higher frequency of G. *tomentosa* autopolyploids in this interval but not of G. *struthium* subsp. *struthium*. In these conditions a higher frequency of triploid individuals would be expected, which could also enhance the formation of the G. *bermejoi* tetraploid hybrid. As a consequence, we considered the 2+3 interval to be favorable for the formation of G. *bermejoi*, between the higher minimum (that of G. *struthium* subsp. *struthium*) and the higher first quartile (that of G. *tomentosa*). Fig 2B and 2D show the MaxEnt models during the LGM for G. *struthium* subsp. *struthium* and G. *tomentosa*, respectively, whilst Fig 2C and 2E show the LSAs for both taxa (suitability values between 0.07603 and 0.52207).

c Defining Low suitability intervals

At this point we need to establish a threshold to define what can be considered “low”. To do so, we used the range between the minimum and the first quartile of the statistical distributions for the suitability values of the parental taxa (Table 2). We considered a low suitability interval to be that defined between the minimum and the first quartile of the suitability values predicted by the SDM at the current locations of both taxa (G. tomentosa minimum = 0.07603, first quartile = 0.52207 and struthium struthium minimum = 0.1306, first quartile = 0.4447). It is important to bear in mind that our aim is to assess the suitability interval that could enhance the formation of allopolyploids. In other words, the suitability values favoring a higher frequency of autopolyploids for at least one of the parental taxa. We used the same interval for both taxa and we established the interval between the lowest minimum and the highest first quartile (Fig 2A). We have to consider this step by step, whereby below the lowest minimum no taxon is present and obviously, we must discard that interval. Between both minima, we had low suitability for one taxon and no presence for the other, making interbreeding impossible. The interval defined by the G. struthium subsp. struthium minimum and its first quartile represents a low suitability for both parental taxa (Fig 2A). In these conditions higher frequencies of autopolyploids could be expected for both taxa, as well as a higher rate of allopolyploid formation. However, we also considered the interval between the first quartiles of G.struthium subsp. struthium (0.4447) and G. tomentosa (0.52207). If the working hypothesis were correct, there would be a higher frequency of autopolyploids in this interval for just one taxon. However, according to the triploid bridge mechanism [76,77], this interval could also facilitate the formation of allopolyploids. As a consequence, we consider that the bioclimatic suitability interval between the minimum of G. struthium subsp. struthium (0.1306) and the first quartile of G. tomentosa (0.52207) could favor the formation of G. bermejoi allopolyploid species. Such intervals correspond to a series of pixels in the maps defining the LSAs for each taxon. The next logical step is to locate those regions in the geographical space and to find where the LSAs for the parental taxa overlap.

Table 2. Models evaluation using AUC and TSS.

AUC values next to one and TSS values above 0.6 are considered good and 0.2–0.6 to be fair to moderate.

	MaxEnt	RF	GLM	MARS
G. bermejoi AUC	0.97	0.99	0.94	0.97
G. bermejoi TSS	0.87	0.93	0.86	0.86
G. tomentosa AUC	0.95	0.98	0.87	0.95
G. tomentosa TSS	0.81	0.89	0.7	0.81
G. struthium AUC	0.95	0.98	0.89	0.95
G. struthium TSS	0.78	0.88	0.7	0.77

Open in a new tab

d Low Suitability Areas in the geographic space for the LGM

Once the low suitability interval is defined, a GIS can be used to establish the LSAs for the parental species of G. bermejoi during the LGM, for which we used QGIS [78]. From the potential area predicted for G. struthium subsp. struthium and G. tomentosa during the LGM (Fig 2B and 2D), we selected those areas with suitability values between 0.1306 and 0.52207 (Fig 2C and 2E). According to our working hypothesis, those areas should have more autopolyploid plants and if so, a higher frequency of allopolyploids would be expected in the locations where both these areas overlap (Fig 3B and 3D). A stronger presence of non-reduced gametes in such populations would facilitate the formation of the fertile hybrid G. bermejoi. Thus, we used QGIS to establish where the LSAs for the parental taxa overlap.

Fig 3 — (2011) was used to plot the sites of gypsum soils on the map *(3A)*. During the LGM, the LSAs of the G. *bermejoi* parental taxa overlapped at some locations due to the high degree of sympatry of both taxa (yellow colored areas in 3B and 3C). According to our hypothesis, higher frequencies of autopolyploids and non-reduced gametes would be expected in those areas. The three taxa studied are strict gypsophytes and hence, we filtered all the gypsum outcrops included in the areas where the LSAs for both parental taxa (G. *struthium* subsp. *struthium* and G. *tomentosa*) overlap (3C). At those sites, there would be a higher probability of establishing thriving populations of the fertile G. *bermejoi* hybrid and thus, we consider such places as potential cradles for this species. 3D shows The overlapping LSAs of the parental G. *bermejoi* taxa during the MH period and (3E) the potential cradles for the same climatic period.

e Filtering for locations with suitable soils: Locating the LGM cradles

All the species considered in this work are strict gypsophytes. So far we have located the areas where bioclimatic suitability could favor the formation of G. bermejoi but we have yet to take into account the special edaphic conditions required by this plant. To select the sites with suitable soils, we used a digital version of the gypsum soils maps published by Mota et al. (2011) [44]. Using an algorithm implemented in R programming language [71] and QGIS [78], we filtered the sites with gypsum outcrops from the areas where the low bioclimatic suitability of the parental taxa overlap (Fig 3B and 3C). We consider these locations as potential cradles for G. bermejoi and in such places, the presence of two concomitant factors could favor the establishment of stable populations of this species: suitable soils and according to the working hypothesis, a high frequency of polyploid individuals from the parental G. bermejoi taxa.

f Locating potential cradles for the Mid-Holocene period

The SDM models show dramatic changes in the potential distribution for the species studied during the MH period. New potential bioclimatic areas mean new LSAs and hence, new potential cradles. To locate these potential cradles, we repeated the protocol using the models for the MH period (Fig 3D and 3E): the LSAs for the MH period can be seen in the supplementary materials section (S3).

2.5 Predicting the expected distribution of G. bermejoi under current climatic conditions

Having determined the potential G. bermejoi cradles, it is possible to forecast its expected current distribution by studying the evolution of the bioclimatic suitability of this taxon at such locations or, in other words, by foreseeing the fate of such potential cradles. Our goal here is to find out which of these locations could have maintained stable populations of G. bermejoi despite the dramatic climatic changes that have occurred in the past 25,000 years. By doing so, we could assess which potential cradles might have survived such climatic oscillations and could currently harbor G. bermejoi populations. Hence, we assessed in which of these cradles bioclimatic suitability remained reasonably high for G. bermejoi as follows:

a Defining a survival threshold for G. bermejoi

A “survival” threshold was defined in order to establish which of the possible cradles could have maintained G. bermejoi populations up to the present day. To do so, we generated a model for G. bermejoi under current climatic conditions and we took the suitability values at the points of occurrence of this species. The resulting distribution is shown as a histogram in Fig 4A to 4D (lower histograms) and the minimum suitability value of this distribution was used as a “survival” threshold. We also evaluated the effects of different thresholds in the results (see discussion).

b Evolution of the bioclimatic suitability conditions for G. bermejoi at the cradles

The evolution of bioclimatic suitability of the potential cradles as a whole was studied from the LGM to the present (Fig 4). We treated the potential cradles generated during the LGM (Fig 4A and 4B) separately from those generated in the Mid-Holocene (Fig 4C and 4D). The former were studied at three moments: LGM, MH period and current climatic conditions. Obviously, we only studied suitability during the MH and the current climate for the MH cradles. We then took the suitability values of G. bermejoi produced by the models at the LGM cradles for the MH and current climatic conditions.

Comparative histograms were generated to get an idea of the potential cradles where the suitability values for G. bermejoi remained below the minimum for different climatic periods (Fig 4), sites at which G. bermejoi populations would be doomed. This approach also enabled us to filter the surviving cradles at different moments in the past and under the present climatic conditions.

c Filtering the potential cradles in the geographical space: an expected current distribution prediction

Two series of maps were generated using QGIS [78] to depict the fate of both the LGM and MH potential cradles (Fig 5A to 5D). A color code reflecting the bioclimatic suitability values at those specific locations was used to show how favorable such sites were for G. bermejoi. All the cradles with suitability values below the minimum were removed because G. bermejoi populations could not endure such bioclimatic conditions. Our goal here was to assess which cradles could have harbored G. bermejoi populations up to the present. It is important to note that on this occasion we used the suitability values for G. bermejoi whilst we used those of its parental taxa to locate the potential cradles.

Due to the dramatic climatic changes in the Iberian Peninsula over the last 25,000 years it is possible that the bioclimatic suitability for G. bermejoi may drop abruptly in some potential cradles, areas in which the G. bermejoi populations will be doomed. Using these maps it is possible to find out in which potential cradles the conditions might maintain thriving G. bermejoi populations to date (Fig 5B and 5D). We applied the process for both LGM and MH cradles and in doing so, we can predict the current distribution of G. bermejoi based on an assumption of no dispersal, a prediction that can be compared with the actual distribution of this species (Fig 5D).

3 Results

3.1 Models validation

The AUC and TSS values obtained for the different algorithms and taxa are summarized on Table 2. One must bear in mind that AUC values next to one and TSS values above 0.6 considered good and 0.2–0.6 to be fair to moderate.

3.2 Location and fate of potential LGM and Mid-Holocene cradles for G. bermejoi

We had adopted a two stage strategy to predict the current distribution of G. bermejoi based on our hypothesis. First, we located the potential cradles according to the distributions of both parental taxa (G. struthium subsp.struthium and G. tomentosa) during the LGM and the MH periods, and we then assessed in which of those possible cradles the climatic conditions remained reasonably well suited to G. bermejoi survival. We obtained several possible centers of formation during the LGM, located in the eastern region of the Iberian Peninsula,(Fig 3C). However, the bioclimatic suitability for G. bermejoi is almost 0 at all of these sites during the MH and only one population seems to survive that climatic period (Fig 5A) but no one to the current climatic conditions (Fig 5B).

The dramatic change in the potential distribution provoked by the climatic transition to the MH period favored different potential cradles for G. bermejoi. In this case, the potential cradles were more or less clustered at five locations: the Central, Eastern and South Eastern Region of the Iberian Peninsula; the Ebro Valley; and interestingly, a site in the North West of the peninsula, beyond the Central Mountain System that represents an important geographical barrier. Most of South Eastern areas were ill-fated in terms of G. bermejoi survival and hence, those below the survival threshold were not considered further (Fig 5C). During the MH period, we saw high bioclimatic suitability at several locations in the Ebro Valley and in the Central, Eastern and North western regions. After this period of benign climatic conditions for G. bermejoi at many sites (Fig 5C), its suitability decreased at many of these locations under current climatic conditions. At present, only some of the cradles in the central and north western sites maintain suitability values that would be compatible with the survival of G. bermejoi populations (Fig 5C).

The surviving cradles can be considered (Fig 5D), along with the sites with gypsophyte suitable soils (Fig 3A). Most of these gypsum outcrops remain unoccupied by G. bermejoi under current climatic conditions, although the models show that this was probably not the case in the past (Fig 3C and 3E). The surviving cradles can be interpreted as the expected distribution of G. bermejoi under conditions where there is absolutely no dispersion (Fig 5E). The occurrence of G. bermejoi is shown along with this expected distribution in Fig 5E for comparison.

4 Discussion

The results presented here suggest that the current distribution of G. bermejoi could be explained, at least in some degree, by our hypothesis. Accordingly, we outline some potential key events concerning the recent evolutionary history of this taxon. In general terms, the bioclimatic suitability for G. bermejoi evolved considerably over the last 25,000 years. From the almost null bioclimatic suitability in the LGM, there was an improvement during the MH period that subsequently decreased as the current climatic conditions were established. Our results show how this trend could have affected the possible formation of G. bermejoi, which would not have been present during the LGM. By contrast, several thriving populations of this species could have been established during the benign MH period, thereafter declining as the current climatic conditions set in established.

If we consider the LSAs of the parental taxa during the LGM, we see that they overlap at a number of sites with suitable soils. According to our hypothesis, those areas would harbor higher frequencies of autopolyploids of both taxa and they would therefore be potential cradles for this species. While we cannot rule out the formation of G. bermejoi in places other than these, the formation rate of hybrid individuals would be lower, as would the probability that stable populations could thrive there. It seems reasonable to accept that the climatic transition to the MH period triggered the formation of the first G. bermejoi populations at such sites (Fig 3C). We believe that hybrid populations would have blossomed at such locations as soon as the LGM climate changed.

During the MH period, new climatic conditions generated other potential areas for the parental taxa by producing different LSAs for G. struthium subsp. struthium and G. tomentosa and hence, new potential cradles for G. bermejoi. We think that the further climatic changes filtered those formation centers and our results indicate that only some of them maintained the bioclimatic conditions for G. bermejoi populations to survive. Such surviving populations are located in the Ebro Valley, the central, eastern and south eastern sectors, and interestingly, in the North West beyond the Central mountain range. Most of the observed sites of G. bermejoi are not too distant from the former ones and we think that this distribution is consistent with our hypothesis.

The presence of G. bermejoi populations in North west of Spain, beyond the Central System seems difficult to explain in terms of migration. This mountain range is a geographic barrier extending in ENE-WSW direction, with summits reaching heights of 2500 m [79] and constituting the boundary between two large tertiary basins: the Duero Basin to the North and the Tagus Basin to the South. This mountain range was the outcome of deformation, and it developed through triaxial compression that occurred during the Oligocene and the Lower Miocene period (circa 25 million years ago) [79–80]. Our hypothesis points to an in situ origin of these populations, because this area matches one of the MH cradles that maintained some degree of climatic suitability up to the present. Thus, the presence of G. bermejoi there could be understood in the framework of our hypothesis.

Our predictions include some locations in the Ebro valley, although we feel that their low suitability values (bioclimatic suitability < 0.25) might rule out these sites as potential cradles. We used also other values of climatic suitability as survival thresholds. The first quartile and its mean with the minimum were also considered. These values decrease markedly the number of surviving populations both in the MH and the current climatic conditions. The histograms showing the bioclimatic suitability at the sites of G. bermejoi occurrence according to the models suggest an extremely low frequency of plants below 0.25. Indeed, the species is not currently found in the central Ebro Valley. Moreover, G. bermejoi is currently quite dispersed across the Southern Plateau, a flat and windy region. Significantly, the anatomy and structure of Gypsophila seeds is appropriate for mid-range dispersal, far from the potential cradles, and these seeds can be released from their capsules by strong winds and animals [81]. Indeed, some invasive species of this genus (G. paniculata) apparently use a dispersal mechanism based on tumbleweed [82, 83].

To conclude, we believe that in conjunction with wind dispersal or other mechanisms, our hypothesis could to some extent explain the current distribution of G. bermejoi. However, many questions still remain unanswered regarding the recent evolutionary history of G. bermejoi. For instance, the connection between the gypsum outcrops and the dispersion capacity of G. bermejoi is unclear, and these data will be crucial to disentangle the recent evolutionary history of this taxon. However, our approach seems to shed some light on this issue and we encourage research into the genetic structure of the G. bermejoi populations, as well as cytogenetic studies to assess the rate and frequency of autopolyploid formation in current LSAs of G. struthium subsp. struthium and G. tomentosa.

Conclusions

According to the results obtained here, we conclude that:

The existence of populations of plants similar to those of the current G. bermejoi during the Last Interglacial period cannot be ruled out, although this question is beyond the scope of our study.
The overlap between the LSAs for the parental G. bermejoi taxa during the LGM and the MH may have favored the formation of stable populations of this fertile hybrid soon after the end of this period. The gypsum outcrops in these areas could represent potential cradles or centers of formation for G. bermejoi.
Only some of these potential cradles have retained suitable climatic conditions for this plant to date. These cradles that have survived to date may predict the expected distribution of G. bermejoi, assuming no dispersion. Such predictions seem to match the true distribution of G. bermejoi, particularly if we consider a degree of dispersion by wind or other means. This could explain the presence of this species in the central region of the Iberian Peninsula.
The MH potential cradles could explain the presence of G. bermejoi populations to the North West of the Central mountain range. The presence of these populations is difficult to explain be means of migration, even though it is predicted by our model.
The hypothesis of higher rates of autopolyploid formation in LSAs could to some extent explain the current distribution of G. bermejoi. Although further research will be required to clarify this issue, cytogenetic studies could address this hypothesis, for example assessing the autopolyploid frequencies of G. struthium subsp. struthium and G. tomentosa at their current LSAs.

Supporting information

S1 Fig. Uncertainty maps for the ensembles of the models generated by the four different algorithms (GLM, RF, MaxEnt and MARS) of G. bermejoi, G. tomentosa and G. struthium subsp. struthium.

(TIFF)

Click here for additional data file.^{(174.6KB, tiff)}

S2 Fig. The same for the ensembles performed from models using different climatic models for the projections to the MH and the LGM.

(TIFF)

Click here for additional data file.^{(169.1KB, tiff)}

S3 Fig. Low Suitability Areas (LSAs) for the parental taxa of G. bermejoi during the Mid Holocene.

For this climatic period, the MaxEnt models for G. struthium subsp. struthium and G. tomentosa are shown in A and C respectively. current distribution of this plant according to our hypothesis. B and D show the LSAs for both taxa during the Mid Holocene.

(TIFF)

Click here for additional data file.^{(263.7KB, tiff)}

Data Availability

All relevant data are within the paper and its Supporting Information files.

Funding Statement

The author(s) received no specific funding for this work.

References

1.Adams KL, Wendel JF. Polyploidy and genome evolution in plants. Current opinion in plant biology, 2005, vol. 8, no 2, p. 135–141. 10.1016/j.pbi.2005.01.001 [DOI] [PubMed] [Google Scholar]
2.Otto SP, Whitton J. Polyploid incidence and evolution. Annu Rev Genet 2000;34:401–37. 10.1146/annurev.genet.34.1.401 [DOI] [PubMed] [Google Scholar]
3.Soltis PS, Soltis DE. Ancient WGD events as drivers of key innovations in angiosperms. Current opinion in plant biology, 2016, vol. 30, p. 159–165. 10.1016/j.pbi.2016.03.015 [DOI] [PubMed] [Google Scholar]
4.Tank DC, Eastman JM, Pennell MW, Soltis PS, Soltis DE. Progressive radiations and the pulse of angiosperm diversification. N Phytol 2015. [DOI] [PubMed] [Google Scholar]
5.Jaillon O et al. The grapevine genome sequence suggests ancestral hexaploidization in major angiosperm phyla. Nature, 2007, vol. 449, no 7161, p. 463 10.1038/nature06148 [DOI] [PubMed] [Google Scholar]
6.Van de Peer Y, Fawcett JA, Proost S, Sterck L & Vandepoele K. The flowering world: a tale of duplications. Trends in plant science, 2009, vol. 14, no 12, p. 680–688. 10.1016/j.tplants.2009.09.001 [DOI] [PubMed] [Google Scholar]
7.Soltis PS, Soltis DE. The role of hybridization in plant speciation. Annual review of plant biology, 2009, vol. 60, p. 561–588. 10.1146/annurev.arplant.043008.092039 [DOI] [PubMed] [Google Scholar]
8.Parisod C, Holderegger R, Brochmann C. 2010. Evolutionary consequences of autopolyploidy. New Phytologist 186, 5–17. 10.1111/j.1469-8137.2009.03142.x [DOI] [PubMed] [Google Scholar]
9.Löve A & Löve D. The significance of differences in the distribution of diploids and polyploids. Hereditas, 1943, vol. 29, no 1‐2, p. 145–163. [Google Scholar]
10.Mable BK. Breaking down taxonomic barriers in polyploidy research. Trends in plant Science, 2003, vol. 8, no 12, p. 582–590. 10.1016/j.tplants.2003.10.006 [DOI] [PubMed] [Google Scholar]
11.Grant V. Plant speciation. Columbia University Press, 1981. [Google Scholar]
12.Stebbins GL. Polyploidy, hybridization, and the invasion of new habitats. Annals of the Missouri Botanical Garden, 1985, p. 824–832. [Google Scholar]
13.Brochmann C et al. Polyploidy in arctic plants. Biological journal of the Linnean society, 2004, vol. 82, no 4, p. 521–536. [Google Scholar]
14.Parisod C, Holderegger R & Brochmann C. Evolutionary consequences of autopolyploidy. New phytologist, 2010, vol. 186, no 1, p. 5–17. 10.1111/j.1469-8137.2009.03142.x [DOI] [PubMed] [Google Scholar]
15.Mable BK. Breaking down taxonomic barriers in polyploidy research. Trends in plant Science, 2003, vol. 8, no 12, p. 582–590. 10.1016/j.tplants.2003.10.006 [DOI] [PubMed] [Google Scholar]
16.Petit C and Thompson J.D. Variation in phenotypic response to light availability between diploid and tetraploid populations of the perennial grass Arrhenatherum elatius from open and woodland sites. Journal of Ecology, 1997, p. 657–667. [Google Scholar]
17.Li WL, Berlyn GP & Ashton PMS. Polyploids and their structural and physiological characteristics relative to water deficit in Betula papyrifera (Betulaceae). American Journal of Botany, 1996, vol. 83, no 1, p. 15–20. [Google Scholar]
18.Wullschleger SD, Sanderson MA, McLaughlin SB, Biradar DP & Rayburn AL. Photosynthetic rates and ploidy levels among populations of switchgrass. Crop Science, 1996, vol. 36, no 2, p. 306–312. [Google Scholar]
19.Del Pozo JC, Ramírez-Parra E. Whole genome duplications in plants: an overview from Arabidopsis. Journal of Experimental Botany, 2015, vol. 66, no 22, p. 6991–7003. 10.1093/jxb/erv432 [DOI] [PubMed] [Google Scholar]
20.Fawcett JA, Van de Peer Y. Angiosperm polyploids and their road to evolutionary success. Trends in Evolutionary Biology, 2010, vol. 2, no 1, p. 16–21. [Google Scholar]
21.Ramsey J, Schemske DW. Pathways, mechanisms, and rates of polyploid formation in flowering plants. Ann Rev Ecol Syst 1998; 29:467–501. [Google Scholar]
22.Mason AS, Pires JC. Unreduced gametes: meiotic mishap or evolutionary mechanism?. Trends in Genetics, 2015, vol. 31, no 1, p. 5–10. 10.1016/j.tig.2014.09.011 [DOI] [PubMed] [Google Scholar]
23.Thompson JD, Lumaret R. The evolutionary dynamics of polyploid plants: origins, establishment, and persistence. Trends Eco Evol 1992;7:302–7. [DOI] [PubMed] [Google Scholar]
24.Husband BC. Constraints on polyploid evolution: a test of the minority cytotype exclusion principle. Proceedings of the Royal Society of London. Series B: Biological Sciences 2000;267:217–23. 10.1098/rspb.2000.0990 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Barringer BC. Polyploidy and self-fertiliza-tion in flowering plants. Am J Bot 2007;94:1527–33. 10.3732/ajb.94.9.1527 [DOI] [PubMed] [Google Scholar]
26.Peterson AT, Soberón J. Species distribution modeling and ecological niche modeling: getting the concepts right. Natureza & Conservação, 2012, vol. 10, no 2, p. 102–107. [Google Scholar]
27.Smith A, Alsdurf J, Knapp M, Baer SG, Johnson LC. Phenotypic distribution models corroborate species distribution models: A shift in the role and prevalence of a dominant prairie grass in response to climate change. Global Change Biology, 2017. 10.1111/gcb.1366628211151 [DOI] [PubMed] [Google Scholar]
28.Benito BM., Martínez-Ortega MM, Munoz LM, Lorite J, & Penas J. Assessing extinction-risk of endangered plants using species distribution models: a case study of habitat depletion caused by the spread of greenhouses. Biodiversity and Conservation, 2009, vol. 18, no 9, p. 2509–2520. [Google Scholar]
29.Bueno ML, Pennington RT, Dexter KG, Kamino LHY, Pontara V, Neves DM, et al. Effects of Quaternary climatic fluctuations on the distribution of Neotropical savanna tree species. Ecography, 2017, vol. 40, no 3, p. 403–414. [Google Scholar]
30.Smeraldo S, Di Febbraro M, Ćirović D, Bosso L, Trbojević I, & Russo D. Species distribution models as a tool to predict range expansion after reintroduction: A case study on Eurasian beavers (Castor fiber). Journal for Nature Conservation, 2017. [Google Scholar]
31.Zutta BR, Rundel PW. Modeled Shifts in Polylepis Species Ranges in the Andes from the Last Glacial Maximum to the Present. Forests, 2017, vol. 8, no 7, p. 232 10.3390/f8070232. [DOI] [Google Scholar]
32.Bosso L, Luchi N, Maresi G, Cristinzio G, Smeraldo S, Russo D. Predicting current and future disease outbreaks of Diplodia sapinea shoot blight in Italy: species distribution models as a tool for forest management planning. Forest Ecology and Management, 2017, vol. 400, p. 655–664. [Google Scholar]
33.Benito BM et al. The ecological niche and distribution of Neanderthals during the Last Interglacial. Journal of Biogeography, 2017, vol. 44, no 1, p. 51–61. [Google Scholar]
34.Rovzar C, Gillespie TW, Kawelo K. Landscape to site variations in species distribution models for endangered plants. Forest Ecology and Management, 2016, vol. 369, p. 20–28. [Google Scholar]
35.Benito BM, Cayuela L, Alburquerque FS. The impact of modelling choices in the predictive performance of richness maps derived from species‐distribution models: guidelines to build better diversity models. Methods in Ecology and Evolution, 2013, vol. 4, no 4, p. 327–335. [Google Scholar]
36.Alburquerque FS, Benito B, Beier P, Assunção-Albuquerque MJ & Cayuela L. Supporting underrepresented forests in Mesoamerica. Natureza & Conservação, 2015, vol. 13, no 2, p. 152–158. [Google Scholar]
37.Glennon KL, Ritchie ME & Segraves KA. Evidence for shared broad‐scale climatic niches of diploid and polyploid plants. Ecology letters, 2014, vol. 17, no 5, p. 574–582. 10.1111/ele.12259 [DOI] [PubMed] [Google Scholar]
38.Levin DA. Minority cytotype exclusion in local plant populations. Taxon, 1975, p. 35–43. [Google Scholar]
39.Fowler NL, Levin DA. Ecological constraints on the establishment of a novel polyploid in competition with its diploid progenitor. The American Naturalist, 1984, vol. 124, no 5, p. 703–711. [Google Scholar]
40.Felber F. Establishment of a tetraploid cytotype in a diploid population: effect of relative fitness of the cytotypes. Journal of Evolutionary Biology, 1991, vol. 4, no 2, p. 195–207. [Google Scholar]
41.López G. Gypsophila bermejoi G. López, sp. nov. y algunos comentarios sobre el geénero Gypsophila con relación a Flora Iberica. An Jard Bot Madr.1984; 41(1): 35–38. [Google Scholar]
42.de Luis M, Bartolomé C, García-Cardo Ó & Álvarez-Jiménez J. Gypsophila bermejoi G. López: A possible case of speciation repressed by bioclimatic factors. PloS one, 2018, vol. 13, no 1, p. e0190536 10.1371/journal.pone.0190536 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.de Luis M, Bartolomé C, García-Cardo Ó, Martínez-Labarga JM, & Álvarez-Jiménez J. Sympatric and allopatric niche shift of endemic Gypsophila (Caryophyllaceae) taxa in the Iberian Peninsula. PloS one, 2018, vol. 13, no 11, p. e0206043 10.1371/journal.pone.0206043 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Mota Poveda JF, Sánchez-Gómez P, Guirado JS. Diversidad vegetal de las yeseras ibéricas. El reto de los archipiélagos edáficos para la biología de la conservación. ADIF-Mediterráneo Asesores Consultores, Almería, 2011. [Google Scholar]
45.Mota Poveda JF, Garrido Becerra JA, Pérez-García FJ, Salmerón-Sánchez E, Sánchez-Gómez P, Merlo E. Conceptual baseline for a global checklist of gypsophytes. Lazaroa. 2016; 37: 7–30. 10.5209/LAZA.54044 [DOI] [Google Scholar]
46.Escudero A, et al. Plant life on gypsum: a review of its multiple facets. Biological Reviews, 2015, vol. 90, no 1, p. 1–18. 10.1111/brv.12092 [DOI] [PubMed] [Google Scholar]
47.Acebes B, Bernabé M., Diaz-Lanza AM, Bartolomé C. Two new sulfated saponins from the roots of Gypsophila bermejoi. J Nat Prod. 1998; 61(12): 1557–1559. 10.1021/np9705221 [DOI] [PubMed] [Google Scholar]
48.Acebes B, Daz-Lanza AM, Bernabé M. A saponin from the roots of Gypsophila bermejoi. Phytochemistry. 1998; 49(7): 2077–2079. 10.1016/s0031-9422(98)00404-x [DOI] [PubMed] [Google Scholar]
49.Boscaiu M, Sánchez M, Bautista I, Donat P, Lidón A, Llinares J, et al. Phenolic compounds as stress markers in plants from gypsum habitats. Bull Univ Agric Sci Vet Med Cluj Napoca, Horticulture. 2010; 67(1): 1843–5254. 10.15835/buasvmcn-hort:4903 [DOI] [Google Scholar]
50.Diniz‐Filho JAF, et al. Partitioning and mapping uncertainties in ensembles of forecasts of species turnover under climate change. Ecography, 2009, vol. 32, no 6, p. 897–906. [Google Scholar]
51.Qiao H, Soberón J, & Peterson AT. No silver bullets in correlative ecological niche modelling: insights from testing among many potential algorithms for niche estimation. Methods in Ecology and Evolution, 2015, vol. 6, no 10, p. 1126–1136. [Google Scholar]
52.Guisan A, Edwards TC Jr & Hastie T, Trevor. Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecological modelling, 2002, vol. 157, no 2–3, p. 89–100. [Google Scholar]
53.Breiman L. Random forests. Machine learning, 2001, vol. 45, no 1, p. 5–32. [Google Scholar]
54.Cutler DR, Edwards TC Jr, Beard KH, Cutler A, Hess KT, Gibson J et al. Random forests for classification in ecology. Ecology, 2007, vol. 88, no 11, p. 2783–2792. 10.1890/07-0539.1 [DOI] [PubMed] [Google Scholar]
55.Evans JS, Murphy MA, Holden ZA & Cushman SA, et al. Modeling species distribution and change using random forest. En Predictive species and habitat modeling in landscape ecology Springer, New York, NY, 2011. p. 139–159. [Google Scholar]
56.Friedman JH. Multivariate adaptive regression splines. The annals of statistics, 1991, p. 1–67. [Google Scholar]
57.Friedman J, Hastie T & Tibshirani R. The elements of statistical learning. New York: Springer series in statistics, 2001. [Google Scholar]
58.Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. Ecological modelling, 2006, vol. 190, no 3, p. 231–259. [Google Scholar]
59.Phillips SJ, Dudík M. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography, 2008, vol. 31, no 2, p. 161–175. [Google Scholar]
60.Elith J et al. A statistical explanation of MaxEnt for ecologists. Diversity and distributions, 2011, vol. 17, no 1, p. 43–57. [Google Scholar]
61.Merow C, Smith MJ, Silander JA. A practical guide to MaxEnt for modeling species’ distributions: what it does, and why inputs and settings matter. Ecography, 2013, vol. 36, no 10, p. 1058–1069. [Google Scholar]
62.G.tomentosa: GBIF.org (10th November 2016) GBIF Occurrence Download http://doi.10.15468/dl.my22ws. G. struthium subs. struthium:GBIF.org (11th November 2016) GBIF Occurrence Download http://doi.10.15468/dl.dzjhbc. G. bermejoi:GBIF.org (11th November 2016) GBIF Occurrence Download http://doi.10.15468/dl.engt9m
63.Hijmans RJ et al. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology, 2005, vol. 25, no 15, p. 1965–1978. [Google Scholar]
64.Gent PR. The community climate system model version 4. Journal of Climate, 2011, vol. 24, no 19, p. 4973–4991. [Google Scholar]
65.Watanabe S et al. MIROC-ESM 2010: Model description and basic results of CMIP5-20c3m experiments. Geoscientific Model Development, 2011, vol. 4, no 4, p. 845. [Google Scholar]
66.Gutjahr O, et al. Max Planck Institute Earth System Model (MPI-ESM1. 2) for High-Resolution Model Intercomparison Project (HighResMIP). Geophysical Model Development, 2019, vol. 12, p. 3241–3281. [Google Scholar]
67.Varela S, Lima-Ribeiro MS. & Terribile LC. A short guide to the climatic variables of the last glacial maximum for biogeographers. PloS one, 2015, vol. 10, no 6 10.1371/journal.pone.0129037 [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Goberville E, Beaugrand G, Hautekèete NC, Piquot Y, & Luczak C. Uncertainties in the projection of species distributions related to general circulation models. Ecology and evolution, 2015, vol. 5, no 5, p. 1100–1116. 10.1002/ece3.1411 [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Pueyo Y & Alados CL. Abiotic factors determining vegetation patterns in a semi-arid Mediterranean landscape: different responses on gypsum and non-gypsum substrates. Journal of Arid Environments, 2007, vol. 69, no 3, p. 490–505. [Google Scholar]
70.Dodd MB & Lauenroth WK. The influence of soil texture on the soil water dynamics and vegetation structure of a shortgrass steppe ecosystem. Plant Ecology, 1997, vol. 133, no 1, p. 13–28. [Google Scholar]
71.R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria: URL https://www.R-project.org/. [Google Scholar]
72.Naimi B, Araújo MB. SDM: a reproducible and extensible R platform for species distribution modelling. Ecography, 2016, vol. 39, no 4, p. 368–375. [Google Scholar]
73.Allouche O, Tsoar A, Kadmon R. Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS). Journal of Applied Ecology, 2006, vol. 43, no 6, p. 1223–1232. [Google Scholar]
74.Fielding AH, Bell JF. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation, 1997, vol. 24, no 1, p. 38. [Google Scholar]
75.Leroy B. et al. Without quality presence–absence data, discrimination metrics such as TSS can be misleading measures of model performance. Journal of Biogeography, 2018, vol. 45, no 9, p. 1994–2002. [Google Scholar]
76.Husband B. C. (2004). The role of triploid hybrids in the evolutionary dynamics of mixed-ploidy populations. Biological Journal of the Linnean Society, 82(4), 537–546. [Google Scholar]
77.Felber F & Bever JD. (1997). Effect of triploid fitness on the coexistence of diploids and tetraploids. Biological Journal of the Linnean Society, 60(1), 95–106. [Google Scholar]
78.QGIS Development Team, 2009. QGIS Geographic Information System. Open Source Geospatial Fpoundation; URL http://qgis.osgeo.org. [Google Scholar]
79.De Vicente G, et al. Cenozoic thick-skinned deformation and topography evolution of the Spanish Central System. Global and Planetary Change, 2007, vol. 58, no 1–4, p. 335–381. [Google Scholar]
80.Vergés J, Fernández M. Ranges and basins in the Iberian Peninsula: their contribution to the present topography. Geological Society, London, Memoirs, 2006, vol. 32, no 1, p. 223–234. [Google Scholar]
81.García Casanova M. Germinación de semillas, desarrollo de plántulas y niveles de prolina en dos especies del género Gypsophila. 2011. Tesis Doctoral. [Google Scholar]
82.Leimbach-Maus HB, Parks SR, Partridge CG. Genetic structure of invasive babys breath (Gypsophila paniculata) populations in a freshwater Michigan dune system. bioRxiv, 2018, p. 401950. [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Darwent AL, Coupland R.T. Life history of Gypsophila paniculata. Weeds, 1966, vol. 14, no 4, p. 313–318.] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0232736.r001

Decision Letter 0

Daniel de Paiva Silva

10 Sep 2019

PONE-D-19-14999

Using Species Distribution Models to locate the potential cradles of the allopolyploid Gypsophila bermejoi G. López (Caryophyllaceae)

PLOS ONE

Dear Mrs. Bartolomé,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by 9th/Nov/2019. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Daniel de Paiva Silva, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments:

Dear Dr. Bartolomé,

after a careful evaluation by two independent reviewers, one positive (major review) and one negative (rejection) that also raised interesting improvements that need to be taken care of, I believe you manuscript deserves a major review. Please focus on answering/changing the analyses suggested by both reviewers in the next review round.

Best regards,

Daniel de Paiva Silva

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following financial disclosure:

"NO"

Please provide an amended Funding Statement that declares *all* the funding or sources of support received during this specific study (whether external or internal to your organization) as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now

Please state what role the funders took in the study. If any authors received a salary from any of your funders, please state which authors and which funder. If the funders had no role, please state: "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

c. Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

3. Please include a copy of Table 2which you refer to in your text.

4. We note you have included a table to which you do not refer in the text of your manuscript. Please ensure that you refer to Table 1 in your text; if accepted, production will need this reference to link the reader to the Table.

5. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Overall I really enjoyed this manuscript and believe it is a significant contribution towards novel applications of ecological niche models. The authors start from a deep theoretical knowledge and develop their models with the intention of testing a clear hypothesis based on phylogenetic/taxonomic knowledge. This may seem trivial, but it makes this manuscript stand out from other studies that use ENMs to identify past refugia/cradles for species or populations. I believe the strategy employed by the authors to identify cradles is very solid and creative, with a strong basis. However, there are a series of improvements to make it suitable for publication. I will describe the major ones here:

1) Authors need to account for ENM uncertainty in both algorithm and climate scenarios. There are a variety of algorithms and past climate models, in a way that using only one of each is a risky and undesirable strategy. Much of variation in model's results come from those different sources. (see Diniz-Filho et al, 2009 - Partitioning and mapping uncertainties in ensembles of forecasts of species turnover under climate change; Qiao et al 2015 - No silver bullets in correlative ecological niche modelling: insights from testing among many potential algorithms for niche estimation; Varella et al. 2015 - A Short Guide to the Climatic Variables of the Last Glacial Maximum for Biogeographers; Goberville et. al. 2015 - Uncertainties in the projection of species distributions related to general circulation models for discussions about those issues).

There are a series of R packages and scripts that may help you fit ENM using other algorithms (BIOMOD, sdm, ZOON are some examples).

2) While I am satisfied with the strategy of using occurrence data of another species as a proxy of soil in the current period, using those same locations to identify cradles in the past is not straightforward, as the distribution is a interplay of several factors and could be different in the past, which means authors may be missing areas with suitable soil. I recommend the authors to directly include soil variables within the models (see Velazco et al. 2018 - Using worldwide edaphic data to model plant species niches: An assessment at a continental extent; Hengl et al. 2014 - SoilGrids1km — Global Soil Information Based on Automated Mapping) or to also create past models for O.tridentata. There is also the need of a better explanation about the procedures at this step.

3)The authors should explore a little further the effects of their survival threshold. While I understand why the authors chose the LPT (Least Presence Threshold), I believe they should consider using a higher threshold as using the LPT based on data from online

databases is risky due to the data being more prone to errors. The suitability frequencies histogram is a great way to visualize your data, and it seems there is a single point in a value of

very low suitability, the LPT is highly affected by those situations and there is no way to be sure that this point represents a viable population.

Therefore, I recommend exploring other thresholds (a common threshold for ENMs is the value that maximizes specificity and sensitivity, max TPR+TNR) and at least evaluate the effects

of the different threshold in the results.

4)This is a suggestion, and I do not think is an option authors must do for this manuscript. There is a nice R package for plant dynamic plant modelling called MigClim which I believe would be an extremely nice addition to the discussion of the cradle dynamics throughout time.

I made other minor comments and comments related to the text directly in the draft and attached the document.

Again congratulations for your study, it was very pleasant and exciting to read.

Reviewer #2: General comments

The authors did a good job attempting to determine the potential Gypsophila bermejoi cradles, but some assumptions remain unclear. For example, the gypsum soils at Iberian Peninsula are ancient soils? In other words, is it ok to use Ononis tridentata current distribution as a proxy to locate gypsum soils and, ultimately, G. bermejoi LGM cradles? There is scientific evidence to confirm that the studied species don’t occur in another soils unless the gypsum ones?

Considering Methods section, the authors must add information to allow work replication. For example, I could not find the occurrence data. Another important point here is that the authors chose to use only one SDM algorithm (MaxEnt) even though its TSS scores are low, and in the Discussion section they try to explain G. bermejoi historical geographic distribution based only on this result. Actually, the Discussion section seems inconsistent and speculative at some points. I hope the authors improve both sections, Methods and Discussion, once the main paper subject has scientific relevance.

Specific Comments

L 65-68: Funny English.

L 72-74: I can’t see your point here.

L 110: Why these species are of ecological, evolutionary and biochemical interest?

L 111: Which taxon? All of them?

L 115 – 116 and 138 – 146: This is Methods.

Figure 1: The legend do not address step 2. Figures resolution must be improved.

L 144: The word “species” must be excluded.

L 160 – 162: Funny English.

L 175: Occurrence data from GBIF must be cleaned up. Please, make sure that you have avoid bias. A useful option could be the package CoordinateCleaner (see Boria et al. 2014 and Zizka et al. 2019).

L 178: Please, make sure that you have used WorldClim 2 (Fick and Hijmans 2017), avoiding some anomalies of the older version pointed out by Escobar et al. 2014.

L 180: Besides CCSM4, what climate models have you used for past and future projections?

L 184 – 192: What are your predictor variables at first, all the 19 from Bioclim? Which clustering algorithm have you used? Have you performed Jackknife with all the predictor variables or after the correlation dendrogram? Did you include soil traits, since you have mention references 51 and 52? Please, write Methods making sure that anyone can replicate your work.

L 187 and 189: “jack-knife” or “jackknife”?

Table 1: After reading the previous paragraph I have this sensation that you have choose predictor variables according to your personal experience. Please, explain in details how do you get to these four variables.

L 202: The hyperlink is outdated.

L 204: What did you consider “redundant presence”? Which protocol did you use to deal with autocorrelation?

L 209: Couldn’t understand the sentence “The study focused on the effect of climatic oscillations and as such, we used only bioclimatic variables to generate the models.”

L 211: The reference 58 is out of context. If you know for sure that these plants are strictly gypsophyllic and “they are only present on gypsum soils and their suitability on other soils is “0”, please insert a reference that confirms that.

L 214 - 216 : So you adjust the functions to avoid interpreting difficult curves? I don’t understand.

L 217: About model evaluation metrics, I strongly recommend the follow reading to the authors: “Leroy et. al. 2018 - Without quality presence – absence data, discrimination metrics such as TSS can be misleading measures of model performance”.

General comment: There are several algorithms that could be used to perform Species Distribution Modelling / Ecological Niche Modelling, so I want to know why did you choose to use only maxent? Is this the better choice to your data? Why?

L 223: “frequencies would be”

L 271: Where is the Table 2?

L 304: In my understanding you used the current location of Ononis tridentata to identify gypsum soils, assuming that its location is the same since LGM, is that correct? So, could you insert some references which ensures this information?

L 309: Does the algorithm have a reference? Again, make sure that anyone can replicate your work.

Figure 3: Please, make sure that all colors, shadows and dots in your map are explained by the figure’s legend and review all details. The figure is self-explanatory, but with the legend becomes confusing.

L 336: Considering that you are using Species Distribution Modeling and it is a standard way to identify species current distribution (and you can include soils traits as well, see Velazco et al. 2017), why did you use all of these procedures? It seems that you are using current occurrence points to point out where the species.

L 471: You wrote more than three paragraphs without cite anyone besides your previous papers about G. bermejoi. Besides, most of your arguments are speculative, like “According to our hypothesis”, “It seems reasonable to accept”, “We believed”, “We think that”, “Seems difficult to explain”, “We feel that”.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: André Felipe Alves de Andrade

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

Attachment

Submitted filename: PONE-D-19-14999_reviewer.pdf

Click here for additional data file.^{(1.9MB, pdf)}

PLoS One. 2020 May 19;15(5):e0232736. doi: 10.1371/journal.pone.0232736.r002

Author response to Decision Letter 0

14 Feb 2020

Dear all, editors and reviewers

The following letter gives an answer to all the changes and explanations requested.

Response to the requirements of the journal and reviewers:

1. We have done again all the models, calculations and maps from scratch. We performed ensembles of models generated using four different algorithms (General Linear Models, MaxEnt, Random Forest and MARS) with 30 bootstrap replications each.

2. Ensembles using Three General Circulation Models were performed for the projections to the Mid Holocene and Last Glacial Maximum for each species.

3. The Uncertainties were assessed and mapped at the former steps.

4. Gypsum soils maps were used in spite of using the indicator species Ononis tridentata.

5. Several survival tresholds for Gypsophila bermejoi were explored to assess their effects on the results.

6. We have revised all the references suggested and taken them into account to improve our research.

7. All the small comments have been addressed.

to Reviewers 1 and 2

We are grateful for the comments made on our study and we appreciate sincerely the encouraging comments. We hope you find all the topics complete and clear with all the information added in this revised manuscript.

We appreciate your feedback and hope to have responded effectively.

Kind regards

The authors

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(9.5KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0232736.r003

Decision Letter 1

Daniel de Paiva Silva

20 Mar 2020

PONE-D-19-14999R1

Using Species Distribution Models to locate the potential cradles of the allopolyploid Gypsophila bermejoi G. López (Caryophyllaceae)

PLOS ONE

Dear Mrs. Bartolomé,

We would appreciate receiving your revised manuscript by May 04 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Daniel de Paiva Silva, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Dear authors, congratulations! The manuscript is nearly there. Reviewer #1, which was one of the original reviewers believe it is done. Still, the new Reviewer#3, who is also a SDM specialist, still believes very minor corrections that need to be done before the text is ready. As soon as you perform those, please resubmit your manuscript up to June 19th 2020. As soon as I receive it, I will forward it only to reviewer #3, so s/he may evaluate if the minor changes were done.

All the best and beware with the COVID-19 virus.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #3: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Reviewer #1: First of all, congratulations for the time taken to perform the modifications. I know it may not have been simple to run the experiment all over, but it was very well done and I believe the additions really improved the overall quality of the manuscript.

I believe the manuscript is ready to be published and the findings are very interesting!

This is a very nice piece of work and presents a very good application for ecological niche models!

I do believe there are some minor issues the authors must address, but it does not change my recommendation to accept the manuscript.

I found three sentences that I believe are related to the previous version of the manuscript, which state that only Maxent was used, which is not true in this new version. The sentences are at lines 216, 233 and 359. This is only a small issue, but can confuse readers, please address those.

Reviewer #3: Comments to the authors

The study evaluates possible cradles for Gypsophila bermejoi species, an allopolyploid species, based on its parent species G. struthium subsp . struthium and G. tomentosa, both diploids species. For these, they projected the parent species environmental suitability to Last Glacial Maximum (LGM), Mid Holocene (MH), and current time periods, and established as potential cradles regions where both species have low suitability values. Then they compared these cradles areas with the G. bermejoi environmental suitability spatial distributions from LGM, passing through MH, until current time period.

The study is well designed and do have a nice idea and hypothesis. My only concern is about the ensemble approach on SDM analyses. I see that the authors added the ensemble framework after their first revision. Nonetheless, they did not reanalyzed their data according to the ensemble results, maintaining only the MaxEnt results in the construction of cradles hypothesis. I know that is a lot of work to do in reanalyzing all data according to the ensemble results. However, they have to point out that the ensemble SDM analyses was only done to evaluate the uncertainties bringing from different “algorithms” and “AOGCMs”. Moreover, they have to show, in some way, that the results come from MaxEnt do not differ, as much as, the results of the other “algorithms”, somehow analytical. Finally, it was missing if the Maxent results, used to build their hypothesis, was come from the ensemble of the three AOGCMs for LGM, and MH.

Specific comments:

Line 22: “Allopolyploid” is miswritten.

Line 81: “Plans” is miswritten.

Lines 216-219: Why only MaxEnt model? I do not understand this part. And what about the ensemble results?

Line 225: Something is misplaced in this sentence. Please rewrite.

Line 233: I do not understand “The MaxEnt approach Ensembles”. You have to enlighten it in the previous sessions, as I asked in the third comment.

Line 359: Why, again only Maxent? And what about the ensemble results?

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #3: No

PLoS One. 2020 May 19;15(5):e0232736. doi: 10.1371/journal.pone.0232736.r004

Author response to Decision Letter 1

2 Apr 2020

Dear all, editors and reviewers

The following letter gives an answer to all the changes and explanations requested.

Response to the requirements of the journal and reviewers:

1. We have corrected all the miswritten words.

2. Sentences at lines 216, 233 and 359 have been corrected. Those lines mention MaxEnt as the only algorithm used and are remnants of the first version of the manuscript. For the current version all the calculations were repeated from scratch using four algorithms, and three General Circulation Models for the projections to the LGM and Mid Holocene.

to Reviewer 1

Again, we are grateful for your encouraging comments. All the small comments have been addressed, including the corrections of the miswritten words and the sentences at lines 216, 233 and 359.

We appreciate your feedback and hope to have responded effectively.

to Reviewer 3

We appreciate sincerely the comments made on our study.

All the mistakes have been corrected, including sentences at lines 216, 233 and 359 which state that MaxEnt was the only algorithm used, which is not true in this new version. Those lines were remnants of the previous version of the manuscript.

In the current version we state clearly that four algorithms were used to perform the ensembles and, later, three General Circulation models were used for the projections to the LGM and Mid Holocene.

We hope you find all the topics complete and clear with all the information added in this revised manuscript.

Kind regards

The authors

Attachment

Submitted filename: Response to Reviewers2.docx

Click here for additional data file.^{(9.5KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0232736.r005

Decision Letter 2

Daniel de Paiva Silva

22 Apr 2020

Using Species Distribution Models to locate the potential cradles of the allopolyploid Gypsophila bermejoi G. López (Caryophyllaceae)

PONE-D-19-14999R2

Dear Dr. Bartolomé,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Daniel de Paiva Silva, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Congratulations! Well done!

Stay safe in the pandemic!

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #3: (No Response)

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #3: (No Response)

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #3: (No Response)

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #3: (No Response)

**********

6. Review Comments to the Author

Reviewer #3: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #3: No

PLoS One. doi: 10.1371/journal.pone.0232736.r006

Acceptance letter

Daniel de Paiva Silva

27 Apr 2020

PONE-D-19-14999R2

Using Species Distribution Models to locate the potential cradles of the allopolyploid Gypsophila bermejoi G. López (Caryophyllaceae)

Dear Dr. Bartolomé:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Daniel de Paiva Silva

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Fig. Uncertainty maps for the ensembles of the models generated by the four different algorithms (GLM, RF, MaxEnt and MARS) of G. bermejoi, G. tomentosa and G. struthium subsp. struthium.

(TIFF)

Click here for additional data file.^{(174.6KB, tiff)}

S2 Fig. The same for the ensembles performed from models using different climatic models for the projections to the MH and the LGM.

(TIFF)

Click here for additional data file.^{(169.1KB, tiff)}

S3 Fig. Low Suitability Areas (LSAs) for the parental taxa of G. bermejoi during the Mid Holocene.

(TIFF)

Click here for additional data file.^{(263.7KB, tiff)}

Attachment

Submitted filename: PONE-D-19-14999_reviewer.pdf

Click here for additional data file.^{(1.9MB, pdf)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(9.5KB, docx)}

Attachment

Submitted filename: Response to Reviewers2.docx

Click here for additional data file.^{(9.5KB, docx)}

Data Availability Statement

All relevant data are within the paper and its Supporting Information files.

[pone.0232736.ref001] 1.Adams KL, Wendel JF. Polyploidy and genome evolution in plants. Current opinion in plant biology, 2005, vol. 8, no 2, p. 135–141. 10.1016/j.pbi.2005.01.001 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref002] 2.Otto SP, Whitton J. Polyploid incidence and evolution. Annu Rev Genet 2000;34:401–37. 10.1146/annurev.genet.34.1.401 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref003] 3.Soltis PS, Soltis DE. Ancient WGD events as drivers of key innovations in angiosperms. Current opinion in plant biology, 2016, vol. 30, p. 159–165. 10.1016/j.pbi.2016.03.015 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref004] 4.Tank DC, Eastman JM, Pennell MW, Soltis PS, Soltis DE. Progressive radiations and the pulse of angiosperm diversification. N Phytol 2015. [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref005] 5.Jaillon O et al. The grapevine genome sequence suggests ancestral hexaploidization in major angiosperm phyla. Nature, 2007, vol. 449, no 7161, p. 463 10.1038/nature06148 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref006] 6.Van de Peer Y, Fawcett JA, Proost S, Sterck L & Vandepoele K. The flowering world: a tale of duplications. Trends in plant science, 2009, vol. 14, no 12, p. 680–688. 10.1016/j.tplants.2009.09.001 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref007] 7.Soltis PS, Soltis DE. The role of hybridization in plant speciation. Annual review of plant biology, 2009, vol. 60, p. 561–588. 10.1146/annurev.arplant.043008.092039 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref008] 8.Parisod C, Holderegger R, Brochmann C. 2010. Evolutionary consequences of autopolyploidy. New Phytologist 186, 5–17. 10.1111/j.1469-8137.2009.03142.x [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref009] 9.Löve A & Löve D. The significance of differences in the distribution of diploids and polyploids. Hereditas, 1943, vol. 29, no 1‐2, p. 145–163. [Google Scholar]

[pone.0232736.ref010] 10.Mable BK. Breaking down taxonomic barriers in polyploidy research. Trends in plant Science, 2003, vol. 8, no 12, p. 582–590. 10.1016/j.tplants.2003.10.006 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref011] 11.Grant V. Plant speciation. Columbia University Press, 1981. [Google Scholar]

[pone.0232736.ref012] 12.Stebbins GL. Polyploidy, hybridization, and the invasion of new habitats. Annals of the Missouri Botanical Garden, 1985, p. 824–832. [Google Scholar]

[pone.0232736.ref013] 13.Brochmann C et al. Polyploidy in arctic plants. Biological journal of the Linnean society, 2004, vol. 82, no 4, p. 521–536. [Google Scholar]

[pone.0232736.ref014] 14.Parisod C, Holderegger R & Brochmann C. Evolutionary consequences of autopolyploidy. New phytologist, 2010, vol. 186, no 1, p. 5–17. 10.1111/j.1469-8137.2009.03142.x [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref015] 15.Mable BK. Breaking down taxonomic barriers in polyploidy research. Trends in plant Science, 2003, vol. 8, no 12, p. 582–590. 10.1016/j.tplants.2003.10.006 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref016] 16.Petit C and Thompson J.D. Variation in phenotypic response to light availability between diploid and tetraploid populations of the perennial grass Arrhenatherum elatius from open and woodland sites. Journal of Ecology, 1997, p. 657–667. [Google Scholar]

[pone.0232736.ref017] 17.Li WL, Berlyn GP & Ashton PMS. Polyploids and their structural and physiological characteristics relative to water deficit in Betula papyrifera (Betulaceae). American Journal of Botany, 1996, vol. 83, no 1, p. 15–20. [Google Scholar]

[pone.0232736.ref018] 18.Wullschleger SD, Sanderson MA, McLaughlin SB, Biradar DP & Rayburn AL. Photosynthetic rates and ploidy levels among populations of switchgrass. Crop Science, 1996, vol. 36, no 2, p. 306–312. [Google Scholar]

[pone.0232736.ref019] 19.Del Pozo JC, Ramírez-Parra E. Whole genome duplications in plants: an overview from Arabidopsis. Journal of Experimental Botany, 2015, vol. 66, no 22, p. 6991–7003. 10.1093/jxb/erv432 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref020] 20.Fawcett JA, Van de Peer Y. Angiosperm polyploids and their road to evolutionary success. Trends in Evolutionary Biology, 2010, vol. 2, no 1, p. 16–21. [Google Scholar]

[pone.0232736.ref021] 21.Ramsey J, Schemske DW. Pathways, mechanisms, and rates of polyploid formation in flowering plants. Ann Rev Ecol Syst 1998; 29:467–501. [Google Scholar]

[pone.0232736.ref022] 22.Mason AS, Pires JC. Unreduced gametes: meiotic mishap or evolutionary mechanism?. Trends in Genetics, 2015, vol. 31, no 1, p. 5–10. 10.1016/j.tig.2014.09.011 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref023] 23.Thompson JD, Lumaret R. The evolutionary dynamics of polyploid plants: origins, establishment, and persistence. Trends Eco Evol 1992;7:302–7. [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref024] 24.Husband BC. Constraints on polyploid evolution: a test of the minority cytotype exclusion principle. Proceedings of the Royal Society of London. Series B: Biological Sciences 2000;267:217–23. 10.1098/rspb.2000.0990 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0232736.ref025] 25.Barringer BC. Polyploidy and self-fertiliza-tion in flowering plants. Am J Bot 2007;94:1527–33. 10.3732/ajb.94.9.1527 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref026] 26.Peterson AT, Soberón J. Species distribution modeling and ecological niche modeling: getting the concepts right. Natureza & Conservação, 2012, vol. 10, no 2, p. 102–107. [Google Scholar]

[pone.0232736.ref027] 27.Smith A, Alsdurf J, Knapp M, Baer SG, Johnson LC. Phenotypic distribution models corroborate species distribution models: A shift in the role and prevalence of a dominant prairie grass in response to climate change. Global Change Biology, 2017. 10.1111/gcb.1366628211151 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref028] 28.Benito BM., Martínez-Ortega MM, Munoz LM, Lorite J, & Penas J. Assessing extinction-risk of endangered plants using species distribution models: a case study of habitat depletion caused by the spread of greenhouses. Biodiversity and Conservation, 2009, vol. 18, no 9, p. 2509–2520. [Google Scholar]

[pone.0232736.ref029] 29.Bueno ML, Pennington RT, Dexter KG, Kamino LHY, Pontara V, Neves DM, et al. Effects of Quaternary climatic fluctuations on the distribution of Neotropical savanna tree species. Ecography, 2017, vol. 40, no 3, p. 403–414. [Google Scholar]

[pone.0232736.ref030] 30.Smeraldo S, Di Febbraro M, Ćirović D, Bosso L, Trbojević I, & Russo D. Species distribution models as a tool to predict range expansion after reintroduction: A case study on Eurasian beavers (Castor fiber). Journal for Nature Conservation, 2017. [Google Scholar]

[pone.0232736.ref031] 31.Zutta BR, Rundel PW. Modeled Shifts in Polylepis Species Ranges in the Andes from the Last Glacial Maximum to the Present. Forests, 2017, vol. 8, no 7, p. 232 10.3390/f8070232. [DOI] [Google Scholar]

[pone.0232736.ref032] 32.Bosso L, Luchi N, Maresi G, Cristinzio G, Smeraldo S, Russo D. Predicting current and future disease outbreaks of Diplodia sapinea shoot blight in Italy: species distribution models as a tool for forest management planning. Forest Ecology and Management, 2017, vol. 400, p. 655–664. [Google Scholar]

[pone.0232736.ref033] 33.Benito BM et al. The ecological niche and distribution of Neanderthals during the Last Interglacial. Journal of Biogeography, 2017, vol. 44, no 1, p. 51–61. [Google Scholar]

[pone.0232736.ref034] 34.Rovzar C, Gillespie TW, Kawelo K. Landscape to site variations in species distribution models for endangered plants. Forest Ecology and Management, 2016, vol. 369, p. 20–28. [Google Scholar]

[pone.0232736.ref035] 35.Benito BM, Cayuela L, Alburquerque FS. The impact of modelling choices in the predictive performance of richness maps derived from species‐distribution models: guidelines to build better diversity models. Methods in Ecology and Evolution, 2013, vol. 4, no 4, p. 327–335. [Google Scholar]

[pone.0232736.ref036] 36.Alburquerque FS, Benito B, Beier P, Assunção-Albuquerque MJ & Cayuela L. Supporting underrepresented forests in Mesoamerica. Natureza & Conservação, 2015, vol. 13, no 2, p. 152–158. [Google Scholar]

[pone.0232736.ref037] 37.Glennon KL, Ritchie ME & Segraves KA. Evidence for shared broad‐scale climatic niches of diploid and polyploid plants. Ecology letters, 2014, vol. 17, no 5, p. 574–582. 10.1111/ele.12259 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref038] 38.Levin DA. Minority cytotype exclusion in local plant populations. Taxon, 1975, p. 35–43. [Google Scholar]

[pone.0232736.ref039] 39.Fowler NL, Levin DA. Ecological constraints on the establishment of a novel polyploid in competition with its diploid progenitor. The American Naturalist, 1984, vol. 124, no 5, p. 703–711. [Google Scholar]

[pone.0232736.ref040] 40.Felber F. Establishment of a tetraploid cytotype in a diploid population: effect of relative fitness of the cytotypes. Journal of Evolutionary Biology, 1991, vol. 4, no 2, p. 195–207. [Google Scholar]

[pone.0232736.ref041] 41.López G. Gypsophila bermejoi G. López, sp. nov. y algunos comentarios sobre el geénero Gypsophila con relación a Flora Iberica. An Jard Bot Madr.1984; 41(1): 35–38. [Google Scholar]

[pone.0232736.ref042] 42.de Luis M, Bartolomé C, García-Cardo Ó & Álvarez-Jiménez J. Gypsophila bermejoi G. López: A possible case of speciation repressed by bioclimatic factors. PloS one, 2018, vol. 13, no 1, p. e0190536 10.1371/journal.pone.0190536 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0232736.ref043] 43.de Luis M, Bartolomé C, García-Cardo Ó, Martínez-Labarga JM, & Álvarez-Jiménez J. Sympatric and allopatric niche shift of endemic Gypsophila (Caryophyllaceae) taxa in the Iberian Peninsula. PloS one, 2018, vol. 13, no 11, p. e0206043 10.1371/journal.pone.0206043 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0232736.ref044] 44.Mota Poveda JF, Sánchez-Gómez P, Guirado JS. Diversidad vegetal de las yeseras ibéricas. El reto de los archipiélagos edáficos para la biología de la conservación. ADIF-Mediterráneo Asesores Consultores, Almería, 2011. [Google Scholar]

[pone.0232736.ref045] 45.Mota Poveda JF, Garrido Becerra JA, Pérez-García FJ, Salmerón-Sánchez E, Sánchez-Gómez P, Merlo E. Conceptual baseline for a global checklist of gypsophytes. Lazaroa. 2016; 37: 7–30. 10.5209/LAZA.54044 [DOI] [Google Scholar]

[pone.0232736.ref046] 46.Escudero A, et al. Plant life on gypsum: a review of its multiple facets. Biological Reviews, 2015, vol. 90, no 1, p. 1–18. 10.1111/brv.12092 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref047] 47.Acebes B, Bernabé M., Diaz-Lanza AM, Bartolomé C. Two new sulfated saponins from the roots of Gypsophila bermejoi. J Nat Prod. 1998; 61(12): 1557–1559. 10.1021/np9705221 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref048] 48.Acebes B, Daz-Lanza AM, Bernabé M. A saponin from the roots of Gypsophila bermejoi. Phytochemistry. 1998; 49(7): 2077–2079. 10.1016/s0031-9422(98)00404-x [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref049] 49.Boscaiu M, Sánchez M, Bautista I, Donat P, Lidón A, Llinares J, et al. Phenolic compounds as stress markers in plants from gypsum habitats. Bull Univ Agric Sci Vet Med Cluj Napoca, Horticulture. 2010; 67(1): 1843–5254. 10.15835/buasvmcn-hort:4903 [DOI] [Google Scholar]

[pone.0232736.ref050] 50.Diniz‐Filho JAF, et al. Partitioning and mapping uncertainties in ensembles of forecasts of species turnover under climate change. Ecography, 2009, vol. 32, no 6, p. 897–906. [Google Scholar]

[pone.0232736.ref051] 51.Qiao H, Soberón J, & Peterson AT. No silver bullets in correlative ecological niche modelling: insights from testing among many potential algorithms for niche estimation. Methods in Ecology and Evolution, 2015, vol. 6, no 10, p. 1126–1136. [Google Scholar]

[pone.0232736.ref052] 52.Guisan A, Edwards TC Jr & Hastie T, Trevor. Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecological modelling, 2002, vol. 157, no 2–3, p. 89–100. [Google Scholar]

[pone.0232736.ref053] 53.Breiman L. Random forests. Machine learning, 2001, vol. 45, no 1, p. 5–32. [Google Scholar]

[pone.0232736.ref054] 54.Cutler DR, Edwards TC Jr, Beard KH, Cutler A, Hess KT, Gibson J et al. Random forests for classification in ecology. Ecology, 2007, vol. 88, no 11, p. 2783–2792. 10.1890/07-0539.1 [DOI] [PubMed] [Google Scholar]

[pone.0232736.ref055] 55.Evans JS, Murphy MA, Holden ZA & Cushman SA, et al. Modeling species distribution and change using random forest. En Predictive species and habitat modeling in landscape ecology Springer, New York, NY, 2011. p. 139–159. [Google Scholar]

[pone.0232736.ref056] 56.Friedman JH. Multivariate adaptive regression splines. The annals of statistics, 1991, p. 1–67. [Google Scholar]

[pone.0232736.ref057] 57.Friedman J, Hastie T & Tibshirani R. The elements of statistical learning. New York: Springer series in statistics, 2001. [Google Scholar]

[pone.0232736.ref058] 58.Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. Ecological modelling, 2006, vol. 190, no 3, p. 231–259. [Google Scholar]

[pone.0232736.ref059] 59.Phillips SJ, Dudík M. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography, 2008, vol. 31, no 2, p. 161–175. [Google Scholar]

[pone.0232736.ref060] 60.Elith J et al. A statistical explanation of MaxEnt for ecologists. Diversity and distributions, 2011, vol. 17, no 1, p. 43–57. [Google Scholar]

[pone.0232736.ref061] 61.Merow C, Smith MJ, Silander JA. A practical guide to MaxEnt for modeling species’ distributions: what it does, and why inputs and settings matter. Ecography, 2013, vol. 36, no 10, p. 1058–1069. [Google Scholar]

[pone.0232736.ref062] 62.G.tomentosa: GBIF.org (10th November 2016) GBIF Occurrence Download http://doi.10.15468/dl.my22ws. G. struthium subs. struthium:GBIF.org (11th November 2016) GBIF Occurrence Download http://doi.10.15468/dl.dzjhbc. G. bermejoi:GBIF.org (11th November 2016) GBIF Occurrence Download http://doi.10.15468/dl.engt9m

[pone.0232736.ref063] 63.Hijmans RJ et al. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology, 2005, vol. 25, no 15, p. 1965–1978. [Google Scholar]

[pone.0232736.ref064] 64.Gent PR. The community climate system model version 4. Journal of Climate, 2011, vol. 24, no 19, p. 4973–4991. [Google Scholar]

[pone.0232736.ref065] 65.Watanabe S et al. MIROC-ESM 2010: Model description and basic results of CMIP5-20c3m experiments. Geoscientific Model Development, 2011, vol. 4, no 4, p. 845. [Google Scholar]

[pone.0232736.ref066] 66.Gutjahr O, et al. Max Planck Institute Earth System Model (MPI-ESM1. 2) for High-Resolution Model Intercomparison Project (HighResMIP). Geophysical Model Development, 2019, vol. 12, p. 3241–3281. [Google Scholar]

[pone.0232736.ref067] 67.Varela S, Lima-Ribeiro MS. & Terribile LC. A short guide to the climatic variables of the last glacial maximum for biogeographers. PloS one, 2015, vol. 10, no 6 10.1371/journal.pone.0129037 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0232736.ref068] 68.Goberville E, Beaugrand G, Hautekèete NC, Piquot Y, & Luczak C. Uncertainties in the projection of species distributions related to general circulation models. Ecology and evolution, 2015, vol. 5, no 5, p. 1100–1116. 10.1002/ece3.1411 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0232736.ref069] 69.Pueyo Y & Alados CL. Abiotic factors determining vegetation patterns in a semi-arid Mediterranean landscape: different responses on gypsum and non-gypsum substrates. Journal of Arid Environments, 2007, vol. 69, no 3, p. 490–505. [Google Scholar]

[pone.0232736.ref070] 70.Dodd MB & Lauenroth WK. The influence of soil texture on the soil water dynamics and vegetation structure of a shortgrass steppe ecosystem. Plant Ecology, 1997, vol. 133, no 1, p. 13–28. [Google Scholar]

[pone.0232736.ref071] 71.R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria: URL https://www.R-project.org/. [Google Scholar]

[pone.0232736.ref072] 72.Naimi B, Araújo MB. SDM: a reproducible and extensible R platform for species distribution modelling. Ecography, 2016, vol. 39, no 4, p. 368–375. [Google Scholar]

[pone.0232736.ref073] 73.Allouche O, Tsoar A, Kadmon R. Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS). Journal of Applied Ecology, 2006, vol. 43, no 6, p. 1223–1232. [Google Scholar]

[pone.0232736.ref074] 74.Fielding AH, Bell JF. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation, 1997, vol. 24, no 1, p. 38. [Google Scholar]

[pone.0232736.ref075] 75.Leroy B. et al. Without quality presence–absence data, discrimination metrics such as TSS can be misleading measures of model performance. Journal of Biogeography, 2018, vol. 45, no 9, p. 1994–2002. [Google Scholar]

[pone.0232736.ref076] 76.Husband B. C. (2004). The role of triploid hybrids in the evolutionary dynamics of mixed-ploidy populations. Biological Journal of the Linnean Society, 82(4), 537–546. [Google Scholar]

[pone.0232736.ref077] 77.Felber F & Bever JD. (1997). Effect of triploid fitness on the coexistence of diploids and tetraploids. Biological Journal of the Linnean Society, 60(1), 95–106. [Google Scholar]

[pone.0232736.ref078] 78.QGIS Development Team, 2009. QGIS Geographic Information System. Open Source Geospatial Fpoundation; URL http://qgis.osgeo.org. [Google Scholar]

[pone.0232736.ref079] 79.De Vicente G, et al. Cenozoic thick-skinned deformation and topography evolution of the Spanish Central System. Global and Planetary Change, 2007, vol. 58, no 1–4, p. 335–381. [Google Scholar]

[pone.0232736.ref080] 80.Vergés J, Fernández M. Ranges and basins in the Iberian Peninsula: their contribution to the present topography. Geological Society, London, Memoirs, 2006, vol. 32, no 1, p. 223–234. [Google Scholar]

[pone.0232736.ref081] 81.García Casanova M. Germinación de semillas, desarrollo de plántulas y niveles de prolina en dos especies del género Gypsophila. 2011. Tesis Doctoral. [Google Scholar]

[pone.0232736.ref082] 82.Leimbach-Maus HB, Parks SR, Partridge CG. Genetic structure of invasive babys breath (Gypsophila paniculata) populations in a freshwater Michigan dune system. bioRxiv, 2018, p. 401950. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0232736.ref083] 83.Darwent AL, Coupland R.T. Life history of Gypsophila paniculata. Weeds, 1966, vol. 14, no 4, p. 313–318.] [Google Scholar]

PERMALINK

Using species distribution models to locate the potential cradles of the allopolyploid Gypsophila bermejoi G. López (Caryophyllaceae)

Miguel de Luis

Julio Álvarez-Jiménez

Francisco Javier Rejos

Carmen Bartolomé

Roles

Abstract

1 Introduction

Fig 1. A possible mechanism of cradle formation.

2 Material and methods

2.1 Study area and species

2.2 Data sources

Table 1. List of the environmental variables selected to implement both the MaxEnt species distribution models.

2.3 Modeling species distributions

2.4 An approach to find potential G. bermejoi cradles

a Generating species distribution models (SDM)

b Obtaining a statistical distribution of suitability values

Fig 2. Low Suitability Areas (LSA) for G. struthium subsp. struthium and G. tomentosa.

c Defining Low suitability intervals

Table 2. Models evaluation using AUC and TSS.

d Low Suitability Areas in the geographic space for the LGM

Fig 3. A digital version of the gypsum soils maps published by Mota et al.

e Filtering for locations with suitable soils: Locating the LGM cradles

f Locating potential cradles for the Mid-Holocene period

2.5 Predicting the expected distribution of G. bermejoi under current climatic conditions

a Defining a survival threshold for G. bermejoi

Fig 4. The bioclimatic suitability of G. bermejoi cradle locations in the LGM and Mid-Holocene periods were evaluated in the Mid-Holocene and current climate conditions.

b Evolution of the bioclimatic suitability conditions for G. bermejoi at the cradles

c Filtering the potential cradles in the geographical space: an expected current distribution prediction

Fig 5. Fate of the potential G. bermejoi cradles and expected current distribution of this plant according to our hypothesis.

3 Results

3.1 Models validation

3.2 Location and fate of potential LGM and Mid-Holocene cradles for G. bermejoi

4 Discussion

Conclusions

Supporting information

Data Availability

Funding Statement

References

Decision Letter 0

Daniel de Paiva Silva

Roles

Author response to Decision Letter 0

Decision Letter 1

Daniel de Paiva Silva

Roles

Author response to Decision Letter 1

Decision Letter 2

Daniel de Paiva Silva

Roles

Acceptance letter

Daniel de Paiva Silva

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases