Abstract
Genetic variation evolves during postglacial range expansion of a species and is important for adapting to varied environmental conditions. It is crucial for the future survival of a species. We investigate the nuclear DNA sequence variation to provide evidence of postglacial range expansion of Musa basjoo var. formosana, a wild banana species, and test for adaptive evolution of amplified fragment length polymorphic (AFLP) loci underlying local adaptation in association with environmental variables. Postglacial range expansion was suggested by phylogeographical analyses based on sequence variation of the second intron of copper zinc superoxide dismutase 2 gene. Two glacial refugia were inferred by the average F ST parameter (mean F ST of a population against the remaining populations). Using variation partitioning by redundancy analysis, we found a significant amount of explained AFLP variation attributed to environmental and spatially-structured environmental effects. By combining genome scan methods and multiple univariate logistic regression, four AFLP loci were found to be strongly associated with environmental variables, including temperature, precipitation, soil moisture, wet days, and surface coverage activity representing vegetation greenness. These environmental variables may have played various roles as ecological drivers for adaptive evolution of M. basjoo var. formosana during range expansion after the last glacial maximum.
Similar content being viewed by others
Introduction
In the Quaternary, temperature oscillations are an important historical factor influencing the current distributions of a plant species1. During the last glacial maximum (LGM), most plant species in the Northern hemisphere would have retreated southward toward the tropics or warmer lowland areas, and survived in refugia2. Taiwan is a continental island situated off the coast of the Asian mainland and lies to the south of the Ryukyu Arc and north of the Philippines Archipelago. Although remnants of glaciations at the top of some peaks along the central mountain range (CMR) were found3, the lowlands of Taiwan were not covered with ice but were drier and colder, and the climate changes would have confined species to refugia in low elevations3, 4. In Taiwan, many conifers escaped to the middle elevations during the LGM4. These species were originally distributed at middle and low elevations and would have migrated to the lowlands. A reverse course of events occurred since the LGM1, 2, 5, with species that were previously confined to refugia in the south expanded polewards and lowland forests colonizing at higher elevations1, 2, 6. Environmental gradients can potentially act as selective drivers during postglacial range expansion and result in locally adapted variants2. The current distributions of species are the results of a combinatorial effect by historical events, ecological factors, and stochastic or neutral mechanisms.
Population adaptive divergence is a central issue in evolutionary biology that focuses on understanding the correlations of population genetic variation with environmental heterogeneity7. Studies have shown that natural population divergence is driven by variable environmental conditions and leads to the evolution of locally adapted lineages8,9,10,11. However, gene flow between closely related genetic lineages can be reduced by a combinatorial effect of geography and environment; and geographical isolation may play a larger role than environmental variation in shaping population structure11,12,13. Therefore, investigating the relative roles of geography and environment that influence genetic variation is critical to understand how environmental factors may act as selective drivers and lead to adaptive genetic variation underlying local adaptation in natural populations of a species11, 12. Moreover, identifying environmental factors that play roles in driving adaptive divergence is particularly of interest in biological conservation and ecological restoration14.
In Taiwan, genetic signatures of postglacial expansion into ranges of habitats from refugia were observed in many tree species such as Cyclobalanopsis glauca 15, Cunninghamia konishii 16, Trochodendron aralioides 17, Castanopsis carlesii 18, and Cinnamomum kanehirae 6, 19. Although Taiwan spans a small range of latitude geographically, varied geographical topologies support vegetation from tropical to cool climates20. The dramatic topological differences combined with the influence of tropical and subtropical climates have fostered high habitat diversity and may serve as a driving force for adaptive evolution9,10,11,12. During the postglacial recolonization process, a species would have evolved with adaptive variation invoked by differential selection along environmental gradients occurring in species’ distribution ranges.
Musa basjoo Siebold is a cold hardy banana species and is thought to be originated from southern China21, 22 and is genetically differentiated from other Musa species22, 23. M. basjoo Siebold & Zucc. ex linuma var. formosana (Warb. ex Schum.) Ying is a variety of M. basjoo endemic to Taiwan24. M. basjoo var. formosana is distributed in low elevations in Taiwan25 and has a diploid chromosome number of 2n = 2226. The generation time of bananas under cultivation can be up to 18 months25 and could be longer for natural populations of M. basjoo var. formosana. Fruit from M. basjoo var. formosana is edible but has numerous large and hard seeds. This species is an important germplasm for banana breeding due to its characteristics of cold tolerance and disease resistance. For conservation of this species, it is important to identify the genetic relationships of individuals among populations and to understand the potential for evolutionary adaptation because biodiversity is increasingly threatened by human-induced anthropogenic climate change27.
In this study, the second intron of copper zinc superoxide dismutase 2 gene (Cu/Zn SOD2) was sequenced for 46 individuals from eight populations (Table 1 and Fig. 1) to characterize the postglacial recolonization event. Moreover, amplified fragment length polymorphism (AFLP)28 was also used to survey genetic variation of 112 individuals to examine the genetic relationships of individuals from different populations and test for association of AFLP loci with environmental variables underlying local adaptation driven by environmental gradients. We used variation partitioning based on redundancy analysis (RDA)29 to assess the relative influences of geographical and ecological isolation contributes to AFLP variation. F ST outliers were identified using genome scan methods. In light of understanding local adaptations associated with environmental gradients, multiple univariate logistic regression was used to correlate environmental variables with AFLP loci that have potentially evolved under selection. We hypothesized that frequencies of AFLP outliers may display correlations with environmental gradients underlying local adaptation because the distribution of M. basjoo var. formosana in different geographical regions represents habitat environmental heterogeneities across the CMR. The main goals of this study were to: (i) demonstrate postglacial expansion of M. basjoo var. formosana; and (ii) find evidence for potential adaptive evolution associated with environmental gradients.
Results
Nucleotide diversity, population differentiation, and test for postglacial expansion based on the second intron sequences of Cu/Zn SOD2
We obtained 92 sequences of the second intron of Cu/Zn SOD2 with an aligned length of 1264 bp. No sign of recombination was found examined within populations. However, one non-significant recombination event (P = 0.876) was found between nucleotide sites 15 and 510 when total samples were analyzed. In addition, 2 of 46 samples collected from populations Shouka and Wufeng were heterozygotes and one single-base pair indel was found in the aligned sequences. Sixteen haplotypes were identified in the 92 sequences analyzed (Table 1 and Fig. 2). The number of haplotypes ranged from 1 to 5 for each population. The most common haplotype (haplotype I) was found in all populations examined and the second most common haplotype (haplotype II) was found in four populations (populations Shanmai, Shitou, Shouka, and Wulai) (Table 1 and Fig. 2). Populations Shitou and Wufeng had the greatest number of haplotypes and the population Sandimen had only one haplotype (the most common haplotype I). Although mismatch distribution of the frequency of pairwise differences among haplotypes did not fit tightly with a population expansion model (Kolmogorov-Smirnov test, P < 0.001, Supplementary Fig. S1), the haplotype network displayed “star-like” phylogeny in two separate groups indicating population expansion and subdivision (Fig. 2).
Population pairwise F ST values calculated from Cu/Zn SOD2 second intron sequence variation was mostly low (Supplementary Table S1) and averaged 0.053. The southern population Sandimen showed the highest mean pairwise F ST (=0.154) against the remaining populations (Fig. 3), which may have been the southern glacial refugium for M. basjoo var. formosana. Table 1 presents the descriptive statistics of nucleotide diversity of the second intron of Cu/Zn SOD2 include haplotype diversity (Hd) and nucleotide diversity θ S and θ π. Hd ranged between 0 (population Sandimen) and 0.848 (population Shitou), θ π ranged between 0 (population Sandimen) and 0.00173 (population Beishi), and θ S ranged between 0 (population Sandimen) and 0.00168 (population Wufeng). The levels of Hd, θ π, and θ S were not correlated with the number of sequences (sample size) via Pearson’s correlation test (Hd, r = −0.497, P = 0.209; θ π, r = −0.383, P = 0.349; and θ S, r = −0.469, P = 0.240).
Neutrality tests based on Tajima’s D 30 and Fu’s Fs31 revealed non-significant negative values in many populations and even positive values were observed (Table 2). In contrast, for pooled samples, significant negative values in Tajima’s D (=−1.6544, P = 0.015) and Fu’s Fs (=−6.8767, P = 0.008) were found. Moreover, non-significant raggedness index (rg)32 was found for all individual populations and the pooled samples (rg = 0.0697, P = 0.84) (Table 2). Significant values of the sum of square deviations (SSD) statistic were found for several populations (populations Guanhu, Shitou, Shouka, and Wulai) and R 2 index33 was non-significant for all individual populations (Table 2). Nevertheless, non-significant SSD (=01976, P = 0.19) and significant R 2 index (=0.0416, P = 0.04) were found for pooled samples, indicating spatial range expansions34. The mean demographic expansion factor (τ)35, representing the time since the beginning of an expansion, was 1.6167 (0.30225–8.57544, 95% confidence intervals (CIs)). The time at which the expansion event took place was dated with a mean of 21317 (3985–113072, 95% CIs) years before present for M. basjoo var. formosana, which corresponded roughly to the time frame at the end of the LGM, from 25,000–18,000 years before the present2, 5, 36, 37.
AFLP diversity and differentiation
Twelve primer pairs generated a total of 521 AFLP loci in the entire sample with an overall repeatability of 94.68% (Supplementary Table S2). The proportion of AFLP polymorphic loci ranged from 50.7% (population Wulai) to 75.4% (population Shanmai) with an average value of 63.1% (Table 2). The level of Nei’s genetic diversity (H E)38 averaged 0.276 and ranged from 0.231 (population Shouka) to 0.336 (population Wufeng) (Table 2). The mean H E was 0.276.
In AFLP data, the northern population Wulai had the largest average pairwise F ST against the remaining populations (mean F ST = 0.274, Fig. 2 and Supplementary Table S1). In the HICKORY analysis39, the inbreeding coefficient (f ) = 0 model, which estimated θ B (an F ST analogue) with f = 0 best fitted the data by having the lowest deviance information criterion (DIC) and \(\overline{{\rm{D}}}\) (a measure of how well the model fits the data) values, but there was little difference in the DIC values between the f = 0 model and the next best model (full model) (13771.7 vs. 13772.4) (Supplementary Table S3). The full model estimated an f of 0.0789 (0.0091–0.1728, 95% CIs). These results suggested that the degree of inbreeding was low in M. basjoo var. formosana. The M. basjoo var. formosana populations were moderately structured according to HICKORY estimates based on the f = 0 and full model (θ B = 0.1354 and 0.1429, respectively). The moderate level of genetic differentiation between populations estimated with HICKORY was consistent with the values estimated with AFLP-SURV40 (average pairwise F ST = 0.1284, Supplementary Table S1) and hierarchical analysis of molecular variance (AMOVA) (Φ ST = 0.2117, P < 0.001, Supplementary Table S4). When the population Wulai was excluded (because it was highly differentiated from other populations based on DAPC analysis; see result in the following section, Fig. 4), AMOVA revealed a Φ ST value of 0.1308 (P < 0.001). AMOVA Φ ST value based on the three DAPC clusters was 0.2091 (P < 0.001).
In the STRUCTURE analysis41, the maximal ΔK value (change in the log probability) occur at K = 2 (Supplementary Fig. S2). However, the highest mean log likelihood (LnP(D)) was obtained when K = 8. In the LEA analysis42, the minimal cross-entropy was lowest when K = 7 (Supplementary Fig. S3). However, analysis based on discriminant analysis of principal components (DAPC)43 revealed three genetic clusters with the first two linear discriminants described 74.48% of the total AFLP genetic variation (Fig. 3). DAPC results provided prominent phylogenetic breaks among the clusters. In addition, the values of symmetric similarity coefficient44 (SSC) were higher when K = 2, 3, and 4 (0.995, 0.986, and 0.983, respectively) and lower when K ≥ 5 (Supplementary Fig. S4). Therefore, K = 3 could be the most probable genetic clustering scenario according to the results of DAPC, STRUCURE, and LEA analyses (Fig. 3 and Supplementary Fig. S5).
Effect of environmental variables on AFLP variation among populations
Seven environmental variables, including annual mean temperature (BIO1), annual precipitation (BIO12), mean wind speed (WSmean), normalized difference vegetation index (NDVI), soil pH, soil moisture index (TMI)45, and wet days (Supplementary Table S5) were retained as explanatory variables for AFLP variation. Variation partitioning29 revealed a vast amount of unexplained variation (80.94%, fraction [d]), and the proportion of explained variation was 19.06% (fraction [a+b+c]) (Supplementary Table S6). Within the explained variation, 10.81% (P < 0.001) explained by pure environmental variables (fraction [a]) and 8.25% (P < 0.001) by geographically structured environmental variables (fraction [b]). No AFLP variation attributed to pure geographical difference (fraction [c]) was found.
AFLP loci potentially evolved under selection and test for association with environmental variables
We performed outlier detection for AFLP loci potentially evolved under selection with two neutrality test methods: DFDST and BAYESCAN46, 47. Multiple univariate logistic regression was used to test for correlations of frequencies of AFLP loci with values of environmental variables using Samβada48. DFDSIT and BAYESCAN identified six (1.15%) and five (0.96%) AFLP loci that potentially evolved under selection (Table 3). Samβada analysis found seven AFLP loci strongly correlated with environmental variables (Table 3). We considered four AFLP loci (P1_17, P3_23, P3_24, and P6_12) as adaptive loci potentially evolved under selection because they were identified by either neutrality test method and correlated strongly with environmental variables. Four AFLP loci that identified as outliers by DFDIST or BAYESCAN and correlated strongly with environmental variables, logistic regression plots were drawn (Fig. 5). In the Samβada analysis, significant positive relationships were found for locus P1_17 with soil moisture and wet days (pseudo-R 2 = 0.251, P < 0.0001; pseudo-R 2 = 0.106, P = 0.0025, respectively) and locus P3_24 with annual mean temperature and NDVI (pseudo-R 2 = 0.094, P = 0.00897; pseudo-R 2 = 0.105, P = 0.00897, respectively). Significant negative relationships were found between locus P3_23 and NDVI (pseudo-R 2 = 0.115, P = 0.0056) and between locus P6_12 and annual mean temperature (pseudo-R 2 = 0.101, P = 0.00368). Samβada analysis reported several types of pseudo-R 2 values, we adopted the Nagelkerke pseudo-R 2 because its calculation is based on log likelihoods rather than on residual deviance and scaled approximately from 0 to 1 equivalent to the unadjusted R 2 in linear regression49.
Discussion
In this study, we surveyed sequence variation in the second intron of Cu/Zn SOD2 and AFLP variation to investigate whether the postglacial range expansion occurred in M. basjoo var. formosana and adaptive evolution in association with the environmental gradients of contemporary populations in this species. Sequence variation in the second intron of Cu/Zn SOD2 suggested range expansion since the LGM. A vast amount of unaccounted variation was observed, which is typical for RDA multivariate analysis50. In addition, a considerable amount of unexplained variation could also be attributed to random processes triggered by ecological drift and dispersal, non-spatially structured biological, and/or unmeasured environmental differences11, 51. In addition, the natural landscapes of Taiwan have been dramatically altered by humans, particularly in the lowlands, which may have played a role in influencing genetic variation surveyed in M. basjoo var. formosana. Nevertheless, a significant amount of AFLP variation can be explained by the environmental variables. Local adaptation in M. basjoo var. formosana populations was suggested by the AFLP loci that were potentially evolved under selection in association with environmental variables, including annual mean temperature, annual precipitation, surface coverage activity (NDVI), soil moisture, and number of days with > 0.1 mm of rain per month (wet days).
Glaciers have had a serious impact on the current distribution of plant species and most subtropical species have retreated toward the tropics or warmer lowland areas2. The discrepancy in detecting population expansion for individual populations could be because of either population reduction, population subdivision, a recent bottleneck, or migration52. Our data revealed that the mismatch distribution for the frequency of pairwise differences among haplotypes did not fit tightly with a population in the expansion model (Kolmogorov-Smirnov test, P < 0.001, Supplementary Fig. S1), which indicates diminishing populations or structured sizes of M. basjoo var. formosana 53, 54. In un-subdivided populations, molecular signature characteristics of sudden expansions might not be observed, even when populations had expanded by several orders of magnitude after the LGM55. Although no consistent evidence for population growth was found for individual populations, our results showed evidence for historical spatial range expansion in pooled samples because of significant negative values of neutrality test statistics (Tajima’s D = −1.6544, P = 0.015; Fu’s Fs = −6.8767, P = 0.008) (Table 2). Moreover, the hypothesis of historical spatial range expansion could not be rejected in pooled samples based on non-significant small rg value (rg = 0.0697, P = 0.84)32, non-significant SSD (SSD = 00096, P = 0.81)56, and significant R 2 (R 2 = 0.0416, P = 0.04)33 (Table 2).
Based on the average F ST value for each population in comparison with the remaining populations, glacial refugia of plant species have been inferred in many plant species in Taiwan. The most divergent populations can be found located in the north-central and south (particularly in the southeastern part) parts of Taiwan for tree species6. Populations served as refugia in the north-central part of Taiwan were found for tree species, including T. aralioides 17, Cu. koinishii 57, Ca. carlesii 18, Machilus thunbergii 58, Ma. Kusanoi 58, and Ci. kanehirae 6 based on chloroplast DNA variation; and Ci. kanehirae 19 and Quercus glauca 59 based on nuclear DNA variation. Species with southern glacial refugia include Cy. glauca 15, T. aralioides 17, Ca. Carlesii 18, Ma. thunbergii 58, Ma. Kusanoi 58, and Ci. kanehirae 6 based on chloroplast DNA variation; and Ci. kanehirae 19 and Q. glauca 59 based on nuclear DNA variation. Our results in M. basjoo var. formosana reflect recolonization events after the latest glacial period that probably originated from the southern population Sandimen due to its highest mean pairwise F ST (=0.154) against the remaining populations based on sequence variation of the second intron of Cu/Zn SOD2 (Fig. 3). However, the population Wulai had the highest average F ST based on AFLP variation (Fig. 3) and clearly distinguished from other populations according to the DAPC genetic clustering analysis (Fig. 4), suggesting that the population Wulai may have been the northern glacial refugium for M. basjoo var. formosana.
In comparison with other angiosperm species that occurred in Taiwan, M. basjoo var. formosana had lower nucleotide diversity in the Cu/Zn SOD2 second intron (Hd = 0.628, θ π = 0.00119, and θ S = 0.00280) compared to introns of chalcone synthase (Chs) and leafy (Lfy) genes in Ci. kanehirae (Chs: Hd = 0.841, θ π = 0.00716, and θ S = 0.00371; Lfy: Hd = 0.895, θ π = 0.00479, and θ S = 0.00805)19 and intron of glyceraldehyde-3-phosphate dehydrogenase gene in Q. glacuca (Hd = 0.840 and θ π = 0.00500)59. However, these comparisons may not be appropriate based on only one gene and the gene sequences compared between species were different. Nevertheless, the level of sequence nucleotide diversity reflects the past demographic events in natural populations. Although populations Sandimen and Wulai may have been the glacial refugia for M. basjoo var. formosana, populations Beishi, Shitou, and Wufeng located in north and central Taiwan across the CMR had higher levels of nucleotide diversity (Table 1). These results suggested that populations Beishi, Shitou, and Wufeng may be the melting pots of diversity5. However, the migration routes may have occurred mostly northward from the southern refugium (population Sandimen) and between adjacent populations except for the population Wulai; and scarcely southward from the northern refugium (population Wulai). This is because the population Wulai had a high average F ST (=0.274) in comparison with all other populations based on AFLP variation (Supplementary Table S1) and individuals of the population Wulai are clearly differentiated from individuals of all other populations (Fig. 4). Compared to other populations, the population Wulai is identified as potential refugium located in higher elevation because the rugged geographic topology might serve as an insular area for the survival of this population during the LGM.
AFLP displayed higher level of genetic diversity (average H E = 0.276) in M. basjoo var. formosana compared with other broadleaf tree species occurred in Taiwan, such as Rhododendron oldhamii (average H E = 0.216)11 and species in the genus Salix (average H E = 0.166)12. M. basjoo var. formosana also had a relatively high level of AFLP variation compared to the average H E (=0.230) for 13 plant species summarized in Nybom60. Moreover, the level of AFLP variation of M. basjoo var. formosana was comparable to that of M. balbisiana, another wild banana species, occurred in China (average H E = 0.241)61. Patterns of genetic variation in contemporary populations of a species are influenced by the historical processes that shaped the distribution of a species1, 2, by the landscape ecological properties11, 62, and by life history traits63. High levels of H E in M. basjoo var. formosana may have been related to the low degree of inbreeding revealed by the HICKORY analysis (Supplementary Table S3). Long life span and predominant outcrossing by animal pollinators may account for high AFLP diversity in wild bananas61, 64. The potential build-up of genetic variation during the course of expansion since the LGM, particularly under climate change conditions, is important for species’ survival facing global climate change27. M. basjoo var. formosana harbors a substantial amount of AFLP variation and is an important resource for populations to adapt to changing environmental conditions under natural selection.
In M. basjoo var. formosana, pure environmental and spatially structured environmental factors explained a significant amount of AFLP variation (Supplementary Table S6), which suggests that environments play important roles in influencing genetic variation of this species. Environmental variables such as temperature, precipitation, surface coverage activity, soil moisture, and wet days are the most important ecological drivers influencing genetic variation of M. basjoo var. formosana and correlated strongly with four AFLP loci based on multiple univariate logistic regression analysis (Table 3 and Fig. 5), indicating fitness-related change in AFLP variation. Temperature and precipitation are commonly found to be the ecological drivers strongly correlated with adaptive AFLP variation for many plant species that occur in Taiwan, such as R. oldhamii 11, Keteleeria davidiana var. formosana 9, and Salix species12. Temperature and precipitation appeared to be common ecological drivers for adaptive AFLP variation in natural populations of various plant species8, 65,66,67,68.
Soil properties can explain a non-negligible proportion of the spatial distribution of tree species69 and influence genetic variation among populations of tree species70,71,72. Soil moisture is associated with allozyme genotypes at the glycerate dehydrogenase locus and may play an important role in the adaptation of Pinus edulis 70; and with AFLP loci such as in Fagus sylvatica 71 and Eperua falcate 72. NDVI is a measure of surface coverage activity indicating the level of vegetation greenness and acts as a proxy for a biotic competitive environment. NDVI, estimated with the moderate resolution imaging spectroradiometer, has been shown to be linear with the fraction of absorbed photosynthetically active radiation (fPAR)73, which was not retained as explanatory variable in this study (Supplementary Methods). NDVI and/or fPAR can be influential factors acting on the genetic variation of a species in response to interactions with other species in a local ecological community74; and has been shown to be correlated with population adaptive divergence12 and adaptive divergence between species12, 13.
Understanding environmental variables acting as ecological drivers and playing roles in shaping the contemporary gene pool structure of M. basjoo var. formosana from experienced postglacial range expansion is important. Our results suggest that there are two glacial refugia located in the northern and southern part of the M. basjoo var. formosana distribution range. Postglacial expansion confronting ecological discontinuity may have triggered the evolution of environmentally associated AFLP variation underlying local adaptation. Although a vast amount of AFLP variation was attributed to residual effects, a significant amount of explained variation attributed to environmental effects was found. In conclusion, environmental variables include temperature, precipitation, soil moisture, surface coverage activity, and wet days may have been the most important ecological drivers for adaptive evolution of postglacial expanded M. basjoo var. formosana populations, and have played roles in shaping the current distributions of this species.
Methods
Sampling
A total of 112 individuals from eight populations of M. basjoo var. formosana were collected (Fig. 1; Tables 1 and 2). All samples were subjected to AFLP genotyping and 46 were used to obtain the Cu/Zn SOD2 second intron sequences.
Cloning and sequencing of Cu/Zn SOD2 second intron
The extraction of genomic DNA and total RNA from leaves followed the methods of Dellaporta et al.75 and Clendennen and May76, respectively. First-stranded cDNA was synthesized using total RNA and reverse transcribed (MMLV reverse transcriptase, Promega), and amplified using RACE kit (Invitrogen) with degenerate primers (SOD-F1, 5′-CTCRMKCCDGGNCTCCATGGCTTCC-3′; and SOD-R1, 5′-TTTCCKTCRTCRCCMRCATG-3′). RACE products were then ligated into pGEM-T vector (Promega) and transformed to E. coli. Two full-length Cu/Zn SOD coding sequences were cloned and sequenced (mbCSD1: ABI34606 and mbCSD2: ACY24898).
The BamHI-, BglII-, EcoRI-, or SacI-digested genomic DNA was self-religated and used for the first inverse polymerase chain reaction (IPCR) with CSD-F2 and CSD-R2 primers (5′-GGTGATACCACCAACGGCTGC-3′ and 5′-GCAGCCGTTGGTGGTATCACC-3′). The first IPCR was performed at 94 °C for 6 min, 25 cycles of 20 sec at 94 °C, 15 sec at 60 °C, 2 min at 72 °C, followed by 5 min at 72 °C. The first IPCR amplified fragments were used as template in the second IPCR with CSD-F3 (5′-CACGTGATGAGGAACGACATGC-3′) and CSD-R3 (5′-GCAGCCGTTGGTGGTATCACC-3′) primers in a PCR reaction at 94 °C for 4 min, 25 cycles of 30 sec at 94 °C, 30 sec at 60 °C, 2 min at 72 °C, followed by 5 min at 72 °C. The second IPCR products were subcloned to pUC119 and sequenced (Genbank accession numbers for genomic sequences of Cu/Zn SOD1 and Cu/Zn SOD2 are DQ866814 and GU045759, respectively).
An initial screening found no variation in introns of Cu/Zn SOD1 and Cu/Zn SOD2 except the second intron of Cu/Zn SOD2 in several samples from different populations. Specific CSD2-F1 (5′-CTCCACTGGTAAACCCTCG-3′) and CSD2-R1 (5′-GAGGTCCTGCATGACAACAAG-3′) primers were used to amplify Cu/Zn SOD2 second intron sequences of 46 individuals in PCR reactions with 6 min at 94 °C for 6 min, followed by 30 cycles of 30 sec at 94 °C, 30 sec at 57 °C, 1.5 min at 72 °C, and 5 min holding at 72 °C. The sequential PCR fragments were ligated into pGEM-T vector, transformed into E. coli, and double sequenced using M13 primers and haplotype sequences deposited (GenBank accessions: KX688826~KX688841).
Sequence alignment, nucleotide diversity, haplotype, and demography
Sequences were aligned with Clustal X77 and nucleotide diversity θ π and θ S, haplotype diversity (Hd), R 2 index, and pairwise mismatch distribution estimated using DnaSP v5.078. θ π and θ S were calculated based on the average pairwise number of differences between sequences and the number of segregating sites per sequence, respectively. Population pairwise F ST, raggedness (rg) index of observed mismatch distribution, and neutrality statistics (Tajima’s D and Fu’s Fs) were estimated using Arlequin v3.579. Moreover, goodness-of-fit of the observed mismatch distribution to that expected under the spatial expansion model was tested using the sum of square deviations (SSD) statistic with Arlequin. We generated a haplotype network with the R “pegas” package80, 81.
Neutrality statistics are nearly zero in constant-size populations, whereas a significant negative value and significant positive value reflect processes such as population subdivision or recent bottlenecks. Positive R 2 value and significance of coalescent simulation against neutral model indicate population expansion. The raggedness index is a measure that quantifies the smoothness of the observed mismatch distribution and small rg values represent a population that has experienced sudden expansion32. A significant SSD value indicates departure from spatial expansion34.
To estimate the time since the beginning of an expansion, we used t = τ/2μk, where t is the time elapsed between initial and current population sizes, τ is the estimated number of generations since the expansion, μ is the mutation rate, and k is the sequence length. A mutation rate of 1.5% per 106 year per site was used19. Demographic expansion factor (τ) was estimated using Arlequin. We assumed a mean generation time of 2 years for M. basjoo var. formosana.
AFLP
Genomic DNA (200 ng) was digested with 5 U EcoRI and 5 U MseI (Yeastern Biotech, Taipei, Taiwan) for AFLP genotyping28. Digested products were ligated with EcoRI (0.5 μM) and MseI (5 μM) adaptors in a 10-μl ligation reaction mixture using 30 U T4 DNA ligase (Yeastern Biotech) at 22 °C for 1 hr. Preselective amplification was performed in a total volume of 10-μl reaction buffer, including 1 μl of the digested sample (1:9 dilution), 100 nM of EcoRI (E00: 5′-GACTGCGTACCAATTC-3′) and MseI (M00: 5′-GATGAGTCCTGAGTAA-3′) primers (Supplementary Table S2), 0.25 mM dNTPs, and 1 U Taq (Bernardo Scientific, Taipei, Taiwan). The thermal cycling parameters for preselective amplification were as follows: 2 min at 72 °C and 3 min at 94 °C, followed by 25 cycles of 30 sec at 94 °C, 30 sec at 56 °C, and 5 min holding at 72 °C. Twelve EcoRI (labeled with FAM) and MseI selective primer combinations (E00 and M00 primers with three additional bases; Supplementary Table S2) were used in selective amplification. Selective amplification was performed with 1 μl diluted preselective amplification product (1:9 dilution) in a 10-μl 1 x PCR buffer containing EcoRI and MseI (both 100 nM) selective primers, 0.25 mM dNTPs, 0.75 U Taq with an initial holding at 94 °C for 3 min, followed by 12 cycles of 30 sec at 94 °C, 30 sec at 65 °C with a 0.7 °C touchdown per cycle and 1 min at 72 °C, followed by 24 cycles of 30 sec at 94 °C, 30 sec at 56 °C, 1 min at 72 °C, with a final 5 min holding at 72 °C. Selective amplification were visualized on an ABI PRISM 3100 sequencer (Applied Biosystems, Foster City, CA, USA). AFLP fragments were scored with Peak Scanner v2.0 (Applied Biosystems) for each individual in the range of 150–500 bp with relative fluorescent unit threshold set at 100, and genotyping error rate estimated (Supplementary Table S2). AFLP genotyping data are available in the Supplementary Data.
Environmental variables
Seven environmental variables, including annual mean temperature (BIO1), annual precipitation (BIO12), number of days with > 0.1 mm of rain per month (wet days), mean wind speed (WSmean), normalized difference vegetation index (NDVI), soil pH, and soil moisture index (TMI) were retained as explanatory variables (Supplementary Methods and Supplementary Table S4).
AFLP diversity, structure, and relationships
Nei’s genetic diversity (H E)38 and pairwise F ST based on a Bayesian method with non-uniform prior distribution of allele frequencies82 was calculated using AFLP-SURV40. The numbers of private and fixed private bands were calculated using FAMD83. Hierarchical analysis of molecular variance (AMOVA) was performed with the “poppr” package of R84 and significance tested based on 9999 permutations with the “ade4” package of R85. The Bayesian program, HICKORY v1.139, was used to estimate an F ST analogue (designated θ B) from dominant markers accounting for inbreeding coefficient (f ) using default settings for sampling and chain length parameters (burnin = 5,000; samples = 100,000; thinning = 20). Four models, including a full model, f = 0 model, θ B = 0 model, and f-free model were fitted with the AFLP data (Supplementary Methods).
Initial detection of genetic structure of M. basjoo var. formosana was carried out with two assignment test methods: Bayesian clustering41 and sparse non-negative factorization (sNMF)42 methods. Bayesian clustering method adapted for dominant markers implemented in STRUCTURE41 was used to estimate an individual’s probability of belonging to a homogeneous cluster (K populations). An admixture model was adopted and tested with ten runs, for K ranging from 1 to 9, with 106 iterations and 105 burn-in steps. R package “pophelper”86 was used to summarize mean log likelihood (LnP(D)), change in the log probability (ΔK)87, and symmetric similarity coefficient (SSC)88 to evaluate the fit of different clustering scenarios. Genetic assignment of individuals was also inferred for K values ranging from 1 to 9 based on sNMF algorithm and least-squares optimization with the “LEA” package of R42. The settings for the LEA analysis were: regularization parameter = 100, iterations = 200, and repetitions = 10 with other arguments set to defaults, and the best K evaluated with the means of minimal cross-entropy.
The relationships between individuals and populations of M. basjoo var. formosana were also visualized based on the discriminant analysis of principal components (DAPC) using the R “adegenet” package43, 89. DAPC first performed a principal component analysis (PCA), and an optimal number of PCs was estimated and retained for further discriminant analysis.
Effect of environmental variables on AFLP variation
Redundancy analysis (RDA) was used to partition AFLP variation and explained by environmental and geographical variables using the R “vegan” package90 and significance tested with 9,999 permutations. Proportions of AFLP genetic variation (adjusted R 2) explained by pure environmental variables, geographically structured environmental variables, pure geographical variables, and residual components were estimated29. The longitude and latitude of sample localities were used as geographical variables.
AFLP outlier detection and association with environmental variables
Two F ST outlier detection methods (DFDIST and BAYESCAN) were used to identify AFLP loci that potentially evolved under selection. DFDIST implemented in Mcheza software46 estimated allele frequencies based on the Bayesian approach82 and the highest and lowest 30% of the initial F ST were removed for calculating the mean F ST. Outliers were identified by observed F ST and H E compared to simulated neutral distributions generated using 105 iterations of coalescent simulations. Loci with unusually high F ST values at the 95% confidence level by applying a false discovery rate of 5% were considered potentially evolved under selection. BAYESCAN uses a Bayesian likelihood approach via a reversible-jump Markov Chain Monte Carlo algorithm in comparing the selection versus neutrality model to identify AFLP loci that are potentially evolved under selection47. Posterior odds (PO), the ratio of posterior probabilities of selection over neutrality, was estimated with the settings of a sample size of 50,000 and thinning interval of 20 among 106 iterations, following 100 pilot runs of 50,000 iterations. When an AFLP locus with log10(PO) > 1.591 was considered to have strong evidence for selection.
Samβada48 was further used to evaluate the associations between frequencies of AFLP loci and values of environmental variables using logistic regression model; and significant fit was identified based on the Wald and likelihood ratio tests with Bonferroni correction at P < 0.01. Given the relationships between values of environmental variables and frequencies of AFLP outlier loci, logistic regression plots were depicted.
References
Hewitt, G. The genetic legacy of the Quaternary ice ages. Nature 405, 907–913 (2000).
Hewitt, G. M. Some genetic consequences of ice ages, and their role in divergence and speciation. Biol. J. Linn. Soc. Lond. 58, 247–276 (1996).
Siame, L. et al. Glacial retreat history of Nanhuta Shan (north-east Taiwan) from preserved glacial features: the cosmic ray exposure perspective. Quat. Sci. Rev. 26, 2185–2200 (2007).
Tsukada, M. Late Pleistocene vegetation and climate in Taiwan (Formosa). Proc. Natl. Acad. Sci. USA 55, 543–548 (1966).
Petit, R. et al. Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300, 1563–1565 (2003).
Kuo, D. C. et al. Two genetic divergence centers revealed by chloroplastic DNA variation in populations of Cinnamomum kanehirae Hay. Conserv. Genet. 11, 803–812 (2010).
Stölting, K. N. Genomic scan for single nucleotide polymorphisms reveals patterns of divergence and gene flow between ecologically divergent species. Mol. Ecol. 22, 842–855 (2013).
Jump, A. S., Hunt, J. M., Martínez-Izquierdo, J. A. & Peñuelas, J. Natural selection and climate change: temperature-linked spatial and temporal trends in gene frequency in Fagus sylvatica. Mol. Ecol. 15, 3469–3480 (2006).
Fang, J. Y. et al. Divergent selection and local adaptation in disjunct populations of an endangered conifer, Keteleeria davidiana var. formosana (Pinaceae). PLoS One 8, e70162 (2013).
Hsieh, Y. C. et al. Historical connectivity, contemporary isolation and local adaptation in a widespread but discontinuously distributed species endemic to Taiwan, Rhododendron oldhamii (Ericaceae). Heredity 111, 147–156 (2013).
Huang, C. L. et al. Influences of environmental and spatial factors on genetic and epigenetic variations in Rhododendron oldhamii (Ericaceae). Tree Genet. Genom. 11, 823 (2015).
Huang, C. L. et al. Genetic relationships and ecological divergence in Salix species and populations in Taiwan. Tree Genet. Genomes 11, 39 (2015).
Nakazato, T., Warren, D. L. & Moyle, L. C. Ecological and geographic modes of species divergence in wild tomatoes. Am. J. Bot. 97, 680–693 (2010).
Escudero, A., Iriondo, J. M., Torres, M. E. (2003) Spatial analysis of genetic diversity as a tool for plant conservation. Biol. Conserv. 113, 351–365 (2003).
Huang, S. F., Hwang, S. Y. & Lin, T. P. Spatial pattern of chloroplast DNA variation of Cyclobalanopsis glauca in Taiwan and East Asia. Mol. Ecol. 11, 2349–2358 (2002).
Hwang, S. Y. et al. Postglacial population growth of Cunninghamia konishii (Cupressaceae) inferred from phylogeographical and mismatch analysis of chloroplast DNA variation. Mol. Ecol. 12, 2689–2695 (2003).
Huang, S. F., Hwang, S. Y., Wang, J. C. & Lin, T. P. Phylogeography of Trochodendron aralioides (Trochodendraceae) in Taiwan and its adjacent areas. J. Biogeogr. 31, 1251–1259 (2004).
Cheng, Y. P., Hwang, S. Y. & Lin, T. P. Potential refugia in Taiwan revealed by the phylogeographical study of Castanopsis carlesii Hayata (Fagaceae). Mol. Ecol. 14, 2075–2085 (2005).
Liao, P. C. et al. Historical spatial range expansion and a very recent bottleneck of Cinnamomum kanehirae Hay. (Lauraceae) in Taiwan inferred from nuclear genes. BMC Evol. Biol. 10, 124 (2010).
Su, H. J. Studies on the climate and vegetation types of the natural forests in Taiwan (III): a scheme of geographic regions. Quar. J. Chin. For. 18, 33–44 (1985).
Makino, T. Makino’s New Illustrated Flora of Japan (The Hokuryukan Co., Ltd. Tokyo, 1979).
Liu, A. Z., Li, D. Z. & Li, X. W. Taxonomic notes on wild bananas (Musa) from China. Bot. Bull. Acad. Sin. 43, 77–81 (2002).
Janssens, S. B. et al. Evolutionary dynamics and biogeography of Musaceae reveal a correlation between the diversification of the banana family and the geological and climatic history of Southeast Asia. New Phytol. 210, 1453–1465 (2016).
Ying, S. S. Musaceae in Flora of Taiwan. (2nd ed.) 5, 704–706 (National Taiwan University Press, Taipei).
Pillay, M., Tenkouano, A. & Hartman, J. Bananas and plantains: future Challenges In Musa breeding in Crop Improvement, Challenges in the Twenty-First Century. Chapter 8, 223–252 (Food Products Press, New York, 2002).
Chiu, H. L., Chen, L. F., Shii, C. T. & Chang, Y. C. Study on ploidy of Musa formosana (Warb.) Hayata in Taiwan. J. Taiwan Agric. Res. 59, 78–85 (2010).
Alberto, F. J. et al. Potential for evolutionary responses to climate change - evidence from tree populations. Glob. Change Biol. 19, 1645–1661 (2013).
Vos, P. et al. AFLP: a new technique for DNA fingerprinting. Nucleic. Acids Res. 23, 4407–4414 (1995).
Borcard, D., Legendre, P. & Drapeau, P. Partialling out the spatial component of ecological variation. Ecology 73, 1045–1055 (1992).
Tajima, F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123, 585–595 (1989).
Fu, Y. X. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147, 915–925 (1997).
Harpending, H. C. Signature of ancient population growth in a low resolution mitochondrial DNA mismatch distribution. Hum. Biol. 66, 591–600 (1994).
Ramos-Onsins, S. E. & Rozas, J. Statistical properties of new neutrality tests against population growth. Mol. Biol. Evol. 19, 2092–2100 (2002).
Schneider, S. & Excoffier, L. Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates very among sites: application to human mitochondrial DNA. Genetics 152, 1079–1089 (1999).
Rogers, A. R. & Harpending, H. Population growth makes waves in the distribution of pairwise genetic differences. Mol. Biol. Evol. 9, 552–569 (1992).
Yu, G. et al. Palaeovegetation of China: a pollen data-based synthesis for the mid-Holocene and last glacial maximum. J. Biogeogr. 27, 635–664 (2000).
Bartlein, P. J. et al. Pollen-based continental climate reconstructions at 6 and 21 Ka: a global synthesis. Clim. Dyn. 37, 75–802 (2011).
Nei, M. Analysis of gene diversity in subdivided populations. Proc. Natl. Acad. Sci. USA 70, 3321–3323 (1973).
Holsinger, K. E. & Lewis, P. O. Hickory: a package for analysis of population genetic data v1.0. http://www.academia.edu/1839794/HICKORY (2003).
Vekemans, X., Beauwens, T., Lemaire, M. & Roldán-Ruiz, I. Data from amplified fragment length polymorphism (AFLP) markers show indication of size homoplasy and of a relationship between degree of homoplasy and fragment size. Mol. Ecol. 11, 139–151 (2002).
Falush, D., Stephens, M. & Pritchard, J. K. Inference of population structure using multilocus genotype data: dominant markers and null alleles. Mol. Ecol. Notes 7, 574–578 (2007).
Frichot, E. & Francois, O. LEA: An R package for landscape and ecological association studies. Methods Ecol. Evol. 6, 925–929 (2015).
Jombart, T., Devillard, S. & Balloux, F. Discriminant analysis of principal components: a new method for the analysis of genetically structured populations. BMC Genet. 11, 94 (2010).
Jakobsson, M. & Rosenberg, N. A. CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23, 1801–1806 (2007).
Thornthwaite, C. W. An approach toward a rational classification of climate. Geogr. Rev. 38, 55–94 (1948).
Antao, T. & Beaumont, M. A. Mcheza: A workbench to detect selection using dominant markers. Bioinformatics 15, 1717–1718 (2011).
Foll, M. & Gaggiotti, O. A genome scan method to identify selected loci appropriate for both dominant and codominant markers: a Bayesian perspective. Genetics 180, 977–993 (2008).
Stucki, S. et al. High performance computation of landscape genomic models integrating local indices of spatial association. Mol. Ecol. Resour. Accepted Author Manuscript.. doi:10.1111/1755-0998.12629 (2016).
Nagelkerke, N. J. D. A note on a general definition of the coefficient of determination. Biometrika 78, 691–692 (1991).
Cottenie, K. Integrating environmental and spatial processes in ecological community dynamics. Ecol. Lett. 8, 175–1182 (2005).
Legendre, P. et al. Partitioning beta diversity in a subtropical broad-leaved forest of China. Ecology 90, 663–674 (2009).
Maruyama, T. & Fuerst, P. A. Population bottlenecks and non-equilibrium models in population genetics. II. Number of alleles in a small population that was formed by a recent bottleneck. Genetics 111, 675–689 (1985).
Excoffier, L., Smouse, P. E. & Quattro, J. M. Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics 131, 479–491 (1992).
Rogers, A. R. Genetic evidence for a Pleistocene population expansion. Evolution 49, 608–615 (1995).
Ray, N., Currat, M. & Excoffier, L. Intra-deme molecular diversity in spatially expanding populations. Mol. Biol. Evol. 20, 76–86 (2003).
Excoffier, L. Patterns of DNA sequence diversity and genetic structure after a range expansion: lessons from the infinite-island model. Mol. Ecol. 13, 853–864 (2004).
Chung, J. D., Lin, T. P., Tan, Y. C., Lin, M. Y. & Hwang, S. Y. Genetic diversity and biogeography of Cunninghamia konishii (Cupressaceae), an island species in Taiwan: a comparison with Cunninghamia lanceolata, a mainland species in China. Mol. Phylogenet. Evol. 33, 791–801 (2004).
Wu, S. H. et al. Contrasting phylogeographic patterns of two closely related species, Machilus thunbergii and Machilus kusanoi (Lauraceae), in Taiwan. J. Biogeogr. 33, 936–947 (2006).
Shih, F. L., Cheng, Y. P., Hwang, S. Y. & Lin, T. P. Partial concordance between nuclear and organelle DNA in revealing the genetic divergence among Quercus glauca (Fagaceae) populations in Taiwan. Int. J. Plant Sci. 167, 863–872 (2006).
Nybom, H. Comparison of different nuclear DNA markers for estimating intraspecific genetic diversity in plants. Mol. Ecol. 13, 1143–1155 (2004).
Wang, X. L., Chiang, T. Y., Roux, N., Hao, G. & Ge, X. J. Genetic diversity of wild banana (Musa balbisiana Colla) in China as revealed by AFLP markers. Genet. Resour. Crop Evol. 54, 1125–1132 (2007).
Huang, C. L. et al. Disentangling the effects of isolation-by distance and isolation-by-environment on genetic differentiation among Rhododendron lineages in the subgenus Tsutsusi. Tree Genet. Genomes 12, 53 (2016).
Hamrick, J. L. & Godt, M. J. W. Effects of life history traits on genetic diversity in plant species. Philos. Trans. R. Soc. Lond. B 351, 1291–1298 (1996).
Wong, C. et al. Genetic diversity of the wild banana Musa acuminate Colla in Malaysia as evidenced by AFLP. Ann. Bot. 88, 1017–1025 (2001).
Nakazato, T., Bogonovich, M. & Moyle, L. C. Environmental factors predict adaptive phenotypic differentiation within and between two wild Andean tomatoes. Evolution 62, 774–792 (2008).
Manel, S., Poncet, B. N., Legendre, P., Gugerli, F. & Holderegger, R. Common factors drive adaptive genetic variation at different spatial scales in Arabis alpina. Mol. Ecol. 19, 3824–3835 (2010).
Manel, S. et al. Broad-scale adaptive genetic variation in alpine plants is driven by temperature and precipitation. Mol. Ecol 21, 3729–2738 (2012).
Bothwell, H. et al. Identifying genetic signatures of selection in a non-model species, alpine gentian (Gentiana nivalis L.), using a landscape genetic approach. Conserv. Genet. 14, 467–481 (2013).
John, R. et al. Soil nutrients influence spatial distributions of tropical tree species. Proc Natl Acad Sci USA 104, 864–869 (2007).
Mitton, J. B., Grant, M. C. & Yoshino, A. M. Variation in allozymes and stomatal size in pinyon (Pinus edulis, Pinaceae), associated with soil moisture. Am. J. Bot. 85, 1262–1265 (1998).
Pluess, A. R. & Weber, P. Drought-adaptation potential in Fagus sylvatica: linking moisture availability with genetic diversity and dendrochronology. PLoS One 7, e33636 (2012).
Brousseau, L., Foll, M., Scotti-Saintagne, C. & Scotti, I. Neutral and adaptive drivers of microgeographic genetic divergence within continuous populations: the case of the Neotropical tree Eperua falcate (Aubl.). PLoS One 10, e0121394 (2015).
Huemmrich, K. F., Privette, J. L., Mukelabai, M., Myneni, R. B. & Knyazikhin, Y. Time-series validation of MODIS land biophysical products in a Kalahari woodland, Africa. Int. J. Remote Sens. 26, 4381–4398 (2005).
Violle, C. et al. The return of the variance: intraspecific variability in community ecology. Trends Ecol. Evol. 27, 244–252 (2011).
Dellaporta, S. L., Wood, J. & Hicks, J. B. A plant DNA minipreparation: Version II. Plant Mol. Biol. Rep. 1, 19–21 (1983).
Clendennen, S. K. & May, G. D. Differential gene expression in ripening banana fruit. Plant Physiol. 115, 462–469 (1997).
Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. & Higgins, D. G. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25, 4876–4882 (1997).
Librado, P. & Rozas, J. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25, 1451–1452 (2009).
Excoffier, L. & Lischer, H. E. L. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol. Ecol. Res. 10, 64–567 (2010).
Paradis, E. Pegas: an R package for population genetics with an integrated-modular approach. Bioinformatics 26, 419–420 (2010).
R Development Core Team. R: A Language and Environment for Statistical Computing, Version 3.0.0. http://www.R-project.org/ (2013).
Zhivotovsky, L. A. Estimating population structure in diploids with multilocus dominant DNA markers. Mol. Ecol. 8, 907–913 (1999).
Schlüter, P. M. & Harris, S. A. Analysis of multilocus fingerprint data sets containing missing data. Mol. Ecol. Notes 6, 569–572 (2006).
Kamvar, Z. N., Tabima, J. F. & Grünwald, N. J. Poppr: an R package for genetic analysis of populations with clonal, partially clonal, and/or sexual reproduction. Peer J. 2, e281 (2014).
Dray, S. & Dufour, A. B. The ade4 package: implementing the duality diagram for ecologists. J. Stat. Softw. 22, 1–20 (2007).
Francis, R. M. Pophelper: an r package and web app to analyse and visualize population structure. Mol. Ecol. Resour. 17, 27–32 (2017).
Evanno, G., Regnaut, S. & Goudet, J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol. Ecol. 14, 2611–2620 (2005).
Jakobsson, M. & Rosenberg, N. A. CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23, 1801–1806 (2007).
Jombart, T. Adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics 24, 1403–1405 (2008).
Oksanen, J. et al. Vegan: community ecology package. R package version 2.0-1. https://cran.r-project.org/web/packages/vegan/ (2011).
Jeffreys, H. Theory of probability (3rd ed, Oxford University Press, Oxford, 1961).
Menon, S. ArcGIS 10.3: The next generation of GIS is here. Environmental Systems Research Institute, Inc., CA, USA. http://www.esri.com/software/arcgis (2014).
Open Government Data Providing Organization in Taiwan. http://data.gov.tw/node/35430. Open Government Data License, Version 1.0: http://data.gov.tw/license#eng.
Acknowledgements
This work was supported partly by the Taiwan Council of Agriculture under grant numbers 93AS-4.1.1-FB-e2-3 and 94AS-9.1.7-FB-e1-11 to C.W.S. and by the Taiwan Ministry of Science and Technology under grant number of MOST 103-2313-B-003-001-MY3 to S.Y.H. and MOST 104-2311-B-003-002-MY3 to C.W.S. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Author information
Authors and Affiliations
Contributions
C.W.S. and S.Y.H. conceived and designed the experiments. Y.L.L. and P.C.L. collected the samples. Y.L.L. performed the IPCR experiment and characterized the Cu/Zn SOD2 sequences. J.H.C. performed AFLP experiment and haplotype analysis. C.L.H., C.T.C. and S.Y.H. provided analysis tools and performed the analyses for AFLP diversity and its association with environmental variables. J.H.C., C.L.H., S.Y.H. and C.W.S. wrote the paper.
Corresponding authors
Ethics declarations
Competing Interests
The authors declare that they have no competing interests.
Additional information
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Chen, JH., Huang, CL., Lai, YL. et al. Postglacial range expansion and the role of ecological factors in driving adaptive evolution of Musa basjoo var. formosana . Sci Rep 7, 5341 (2017). https://doi.org/10.1038/s41598-017-05256-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-017-05256-6
This article is cited by
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.