https://orcid.org/0000-0002-1663-6535
Figures
Abstract
Species delimitation is essential to study and conserve biological diversity. It is traditionally based on morphological trait variation observed in one or a few specimens. Nevertheless, such assessments may not sufficiently take intraspecific trait variation into account, misidentifying morphotypes as separate species. The use of high-throughput sequencing data alongside morphological data in taxonomic studies may substantially improve the accuracy of taxonomic assessments. The Musa genus, commonly known for comprising the wild relatives of banana varieties, consists of about seventy described species. However, the taxonomic status of multiple Musa species is uncertain due to typification errors and the lack of high-quality specimens. The species M. splendida and M. viridis from northern Viet Nam only substantially differ from each other in the color of their male flower bracts, which is red to pinkish-red in M. splendida and pink in M. viridis. Consequently, their taxonomic status as separate species has been debated. Here, we studied the genetic relationships between 121 M. splendida and M. viridis plants using high-throughput sequencing data (DArTseq) in which we identified 51,188 single nucleotide polymorphisms. We found that individuals genetically clustered in a principal component analysis (6 clusters), fastStructure analysis (four groups), and ASTRAL-III consensus phylogenetic tree (nine clades) based on their population origin rather than by their taxon identity. In addition, a strong signal for an isolation-by-distance pattern between populations was observed. Plants identified as M. viridis were more closely related to M. splendida plants from the same region than to M. viridis plants from other regions. Hence, we propose to treat M. viridis as a synonym of M. splendida.
Citation: Bawin Y, Mertens A, de Backer S, Vu DT, Le LT, Vu TD, et al. (2025) A splendid banana enigma: Phylogenomic assessment of Vietnamese Musa splendida and Musa viridis populations shows that they are conspecific. PLoS ONE 20(2): e0318252. https://doi.org/10.1371/journal.pone.0318252
Editor: Khalil Kashkush, Ben-Gurion University, ISRAEL
Received: October 15, 2024; Accepted: January 14, 2025; Published: February 11, 2025
Copyright: © 2025 Bawin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All trimmed and quality filtered sequencing data were deposited at the European Nucleotide Archive (ENA) in BioProject PRJEB76413 (https://www.ebi.ac.uk/ena/browser/view/PRJEB76413). All other data used in this study are supplemented as S1 Table.
Funding: This study was funded by Research Foundation Flanders (FWO) (G0D9318N), the Bill and Melinda Gates foundation (BBTV mitigation project OPP1130226). This work is also part of the TETTRIs project funded by the European Union’s HORIZON Innovation Actions under grant agreement No. 101081903. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Species delimitation plays a crucial role in deciphering ecological, biogeographical, and evolutionary patterns across the diversity of life. It is also pivotal for devising effective conservation strategies for endangered species [1]. Traditionally, species have been described based on variation in morphological characters. In plants, some of these morphological characters have a high discriminative power being clearly polymorphic and informative to distinguish among closely related species [2]. However, plants are often characterized by intraspecific morphological variations caused by environmental differences, ontogenetic changes, or genetic mutations [3]. This variability can sometimes lead to misidentification of individuals within the same species as a different species, especially in cases where morphological differences are subtle [2]. Examples of a debatable splitting of species are omnipresent among angiosperm lineages, e.g. Viburnum L. (Caprifoliaceae) [4], Portulaca L. (Portulacaceae) [5], Galium L. (Rubiaceae) [6], Dianthus L. (Caryophyllaceae) [7], and Cyperaceae [8]. Conversely, distinct but closely related species with only very subtle morphological differences (for example in only temporarily available generative characters) are sometimes mistakenly classified under the same taxonomic entity [2]. The integration of molecular data alongside morphological variation significantly enhances the accuracy of species identification, particularly in taxa that are subject to uncertainty in their taxonomic status [9]. Recent advancements in high-throughput sequencing have enabled the cost-effective analysis of numerous genomic loci across large sample sizes, revolutionizing taxonomic studies. Despite their high potential to resolve evolutionary relationships among closely related taxa, these methods are not yet standard in supporting species descriptions [10].
Banana (Musa L.) is a globally significant food crop, with annual production exceeding 135 million tons in 2022 [11]. Most banana cultivars are derived from intraspecific hybridization events of subspecies of the wild species M. acuminata Colla, into which other Musa species like M. schizocarpa N.W.Simmonds, M. balbisiana Colla, and M. textilis Née were either deliberately or accidentally crossed [12, 13]. About 70 species are described in the Musa genus [14]. However, the actual number of Musa species remains enigmatic, as the taxonomy of the genus is believed to be confounded by typification errors and difficulties regarding species identification due to the lack of high-quality herbarium specimens and the occurrence of ephemeral flowers [15–17]. Currently, all Musa species are categorized into two sections: Musa (n = 11) and Callimusa (n = 7/9/10) [14]. The wild relatives of cultivated bananas hold potential for crop improvement, yet many are threatened with extinction due to climate change and habitat degradation and adequate conservation efforts are lacking [18]. Wild Musa species are found throughout Southeast Asia, spanning from eastern India and southern China to Papua New Guinea and northern Australia [19]. Northern Viet Nam, as part of the northern Indo-Burmese ecoregion [19], is renowned for its rich diversity in wild Musa species, with recent surveys identifying six species in the region [20].
Some wild Musa species in northern Viet Nam, such as M. splendida A.Chev. and M. viridis R.V.Valmayor, Đ.D.Lê & Häkkinen in the Callimusa section, exhibit striking morphological similarities. Both species co-occur in the Yên Bái province and are primarily distinguished by the color of the male flower bracts: pink in M. viridis vs. red to pinkish-red in M. splendida [21]. Another difference between these species is the shape of the leaf lamina basis: rounded in M. splendida vs. cuneate in M. viridis. In contrast to M. splendida, M. viridis was reported as non-rhizomatous [22–24]. Musa paracoccinea A.Z.Liu & D.Z.Li, endemic to the Yunnan province in southern China, also features red male flower bracts and shares geographic proximity with M. splendida and M. viridis [16]. However, M. paracoccinea plants are considered to be taller (4–6 m vs. 3–4 m), to have ovately shaped male flower bracts (vs. lanceolate-ovately shaped bracts in M. splendida), and to produce small, bell-shaped seeds (vs. large, bell- or mushroom-shaped seeds in M. splendida) [23, 24]. It remains unclear whether the morphological differences between these Callimusa species justify their separation into separate species [21, 23]. Previous studies using high-throughput sequencing data of Musa species have shown the high resolution of these data to discriminate between wild Musa species and banana cultivars [12, 25]. Thus, the application of high-throughput sequencing data in taxonomic studies of Musa species can provide insights into the relatedness of closely related species as genetic distinction can go as deep as the specimen level. Nevertheless, high-throughput sequencing data have not routinely been incorporated in taxonomic assessments of the Musa genus.
The Musa genus is known to comprise several taxonomic issues due to the limited availability of high-quality specimens from several parts of these plants [17, 23]. As a consequence, many species are only described based on the morphological characters of one or a few plants, often not seen in the wild but only known from cultivation (e.g. M. haekkinenii N.S.Lý & Haev.) [26]. To evaluate the taxonomic status of Musa species, a thorough study of the diversity within multiple populations of a given (range of) species is necessary in which morphological and molecular characters are jointly assessed. The current study aims to elucidate the evolutionary relationships between two endemic Musa species from northern Viet Nam identified as M. viridis or M. splendida by using genome-wide genetic polymorphisms called through DArTseq analysis. Based on the findings in this study, we will provide recommendations regarding the taxonomic status of both species.
Materials & methods
Taxon sampling and genotyping
In total, 121 plants (divided over nine populations) were selected for this study. Of these, 53 specimens were morphologically identified as M. splendida, while 14 specimens were morphologically identified as M. viridis (Fig 1). The remaining 54 specimens from various populations were only in a vegetative state and thus not identified to species level (further indicated as ‘Musa sp.’ throughout the text; S1 Table). Plants were morphologically assigned to a species based on the expertise of local botanists and on characters that were recorded in the field. Musa plants used in this study were sampled from nine populations in the Lào Cai, Hà Giang, and Yên Bái provinces in northern Viet Nam (Fig 2), located within the described range of both species [21]. All field work was conducted as part of a Bilateral cooperation project between Meise Botanic Garden (Belgium) and the Plant Resources Center (Ha Noi, Viet Nam), which was permitted by the Vietnamese National Foundation for Science and Technology Development (NAFOSTED). Field permits were obtained for the Yên Bái (564/UBND-NV), Hà Giang (255/SNN-CCLN), and Lào Cai provinces (1060/UBND-NC). A modified cetyltrimethylammonium bromide (CTAB) protocol from [27] was used to extract DNA from 15 mg of silica-dried leaf material per individual. Next, 20 μl of DNA was shipped to Diversity Arrays Technology (DArT, University of Canberra, Bruce, Australia) for an enzyme-based genomic complexity reduction approach with the restriction enzymes PstI and MseI (DArTseq) following [28]. Several molecular studies applied DArTseq on Musa plants, showing that it is a reliable method for the identification of genetic variation in these species [29–31]. All DArTseq libraries were subsequently 150 base pairs (bp) single-end sequenced on an Illumina NovaSeq 6000 instrument.
The difference in flower bract color is used as the main morphological discriminant between these species (red in M. splendida and pink in M. viridis).
Labels correspond to the ID of each sampling location and pie charts indicate the proportion of M. splendida (red), M. viridis (pink), and unknown individuals (grey) per population. All populations were located close to the Chay river and Red river (blue lines) in the northern part of the country (blue dots in black circle in inset). Map data from © OpenStreetMap under the Open Database Licence [32].
Read data processing
Demultiplexed reads retrieved from DArT were trimmed to remove barcodes and the PstI restriction site remnant at the 5’-end and the MseI restriction site remnant and adapter sequence at the 3’-end with Cutadapt v3.5 [33] in GBprocesS v4.0.0 [34]. Reads with a barcode shorter than the maximum barcode were 3’ trimmed with Cutadapt to compensate for variation in barcode lengths. In addition, reads shorter than 20 bp were deleted. Afterwards, all reads were filtered for maximum five ambiguous nucleotide calls (Ns) and an average base quality score of 25 using the MaxNFilter and AverageQualityFilter in GBprocesS. Reads with internal intact PstI or MseI restriction sites were discarded as well with the RemovePatternFilter in GBprocesS. All trimmed and quality filtered sequencing data were deposited at the European Nucleotide Archive (ENA) in BioProject PRJEB76413.
Trimmed and filtered reads were mapped onto the Musa acuminata subsp. malaccensis ‘DH Pahang’ v4 reference genome sequence [35] using the BWA-MEM algorithm in BWA 0.7.17 [36]. This Musa acuminata reference genome sequence was the most extensively annotated reference genome sequence at the time of our analysis and was demonstrated to be an effective common reference for closely related Musa species [37]. All reads were subsequently indexed, sorted, and filtered for a minimum mapping quality (MAPQ) score of 20 with Samtools v1.17 [38] and tagged with read groups using the Picard AddOrReplaceReadGroups program v3.3.0 [39], resulting in minimum 740,179 and maximum 1,514,724 high-quality mapped reads per sample. Single Nucleotide Polymorphisms (SNPs) in the mapped reads were identified with the Unified Genotyper of the Genome Analysis Toolkit (GATK) v3 [40]. SNP calls were filtered for a minimum minor allele count (MAC) of 4, a minimum SNP quality score of 20, and a minimum genotype quality score of 20 with VCFtools v0.1.17 [41]. Only polymorphic and biallelic SNPs were retained with GATK. Finally, SNP positions with a total read depth below 20, an allele depth below 3, or a data completeness below five percent were removed with a custom python script available on GitLab (https://gitlab.com/ybawin/sequence-data-processing-tetraploids). The data processing work flow and SNP filtering steps were based on approaches applied by previous studies that used similar data [42, 43]. Threshold values for SNP filtering were set after an empirical evaluation of different values on the SNP data of this study.
Read-backed haplotyping
The start and end position of loci was determined using SMAP delineate [44]. The minimum stack depth and cluster depth were set to five and ten, respectively, whereas loci with data in less than five percent of the samples were ignored. Short (< 240 bp) haplotypes were called based on the filtered SNP set and on variation in start and end positions in loci (SMAPs) using SMAP haplotype-sites [44]. Haplotypes were called for loci with minimum 20 reads with a minimum MAPQ score of 20 and if they had a minimum frequency of 5 percent in a locus. Haplotypes that only partially aligned to the locus were included following the best practices for GBS data outlined in the SMAP manual (ngs-smap.readthedocs.io), but haplotypes with insertions or deletions on SNP and SMAP positions were removed. Haplotype frequencies were subsequently converted into discrete dominant calls (0 or 1), applying a frequency bound interval of 20 percent.
Analysis of genetic structure
Based on these haplotype calls, a Jaccard similarity coefficient (J) [45] was calculated for each sample pair and transformed into inversed distances (1–J) using the SMAPapp-Matrix.py script available from the SMAPapps GitLab project (https://gitlab.com/ybawin/smapapps). The resulting genetic distance matrix was used to conduct a Principal Coordinates Analyses (PCoA) with the cmdscale function from the stats package in R v4.3.1 [46]. In addition, genetic stratification in our sampling was evaluated based on SNPs in the filtered VCF file using fastStructure v1.0 [47] with all default settings. The number of subpopulations (K) varied between 1 and 9 and the number of subpopulations with the highest likelihood value was 4. Genetic patterns caused by isolation by distance were inferred by a Mantel test using the mantel.rtest function in the STATS package. The Mantel test was performed on the Jaccard genetic distance matrix and a geographic distance matrix that was created based on longitude and latitude coordinates.
Phylogenetic tree reconstruction
A nucleotide alignment of consensus sequences was created for each locus delineated with SMAP using the SMAPapp-Alignment.py script in the SMAPapps GitLab project. Consensus sequences consisted of both SNPs and invariable sites, and had less than 25 percent missing nucleotide calls. Only alleles supported by minimum 5 reads and with a minimum frequency of 20 percent in the locus were used for the construction of a consensus sequence. Loci with partially overlapping reads were combined into one locus to reduce redundancy between locus alignments. Consensus sequences with only ambiguous nucleotide calls and invariable sites were discarded as well. A maximum likelihood phylogenetic tree was reconstructed for each locus alignment with IQ-TREE 2 [48] using the ModelFinder tool [49] to select the most appropriate substitution model per locus based on the corrected Akaike Information Criterion (AICc). Branch support values (i.e. local posterior probability values) were obtained using non-parametric bootstrapping based on 200 bootstrap replicates. Afterwards, all locus consensus trees were combined into one taxon consensus tree with ASTRAL-III [50].
Results
In total, 50,388 loci were identified in 121 samples, covering 4,629,437 bp (1%) of the M. acuminata reference genome sequence. After SNP filtering, 51,188 SNPs were retained in all samples. All SNP and SMAP variation was combined into 174,951 haplotypes that were identified in 36,369 loci. The first three principal coordinates jointly explained more than 40% of the total variation in the haplotypes table and partitioned the Musa plants into six clusters (Fig 3A). Four out of six clusters consisted exclusively of individuals from one population. The two remaining clusters combined individuals from populations 3 and 4 and from populations 7, 8, and 9, respectively. The first principal coordinate predominantly separated populations 1, 2, 3, 4, and 5 from population 6 and from populations 7, 8, and 9. The second principal coordinate mainly isolated plants from population 1 from all other plants (Fig 3A), while the third principal coordinate confirmed the distinct genetic clustering of populations 5 and 6 into separate groups (S1 Fig). Two clusters consisted of both M. splendida and M. viridis plants: the cluster comprising population 6 and the cluster comprising populations 7, 8, and 9 (Fig 3B).
The first and second coordinate explained 25.2% and 7.7% of the total variation in the dataset, respectively. Individuals were colored based on population ID (left) and taxon ID (right).
The number of subpopulations with the highest likelihood value in the fastStructure analysis was four, predominantly corresponding to the following groups: population 1; population 3 and 4; population 5; and population 7, 8, and 9 (Fig 4). Nearly all plants from population 2 were assigned with high membership probabilities to the subpopulation containing population 1 and to the subpopulation containing population 5. These plants also had low membership probabilities to the subpopulation with populations 7, 8, and 9. All individuals from population 6 had substantially high membership probabilities to population 5 and to populations 7, 8, and 9. A few plants from population 1 and 2 were equally assigned to the subpopulation consisting of populations 3 and 4. A positive and significant correlation was found between the genetic distance matrix and the geographic coordinates of the plants based on the results of the Mantel test (observation = 0.79, p-value = 0.0001), providing evidence for a pattern of isolation by distance in the Musa populations.
Numbers below the barplot indicate the population ID of the samples.
All locus and SNP data were combined into 11,859 non-overlapping polymorphic locus alignments. The accessions in the ASTRAL consensus tree were subdivided into nine clades, corresponding to their geographic distribution across the study area (Fig 5). Clustering into nine clades was supported by high local posterior probability values (> 0.9). Plants assigned to M. splendida and M. viridis clustered together in clades 6 and 7 of the phylogenetic tree.
All plants were partitioned into nine clades that were indicated by the curved lines and by their population ID next to the tip labels. Node labels are local posterior probability values. Only values of 0.95 or higher are shown.
Discussion
Recent advances in the acquisition and processing of high-throughput sequencing data enabled the use of genome-wide polymorphic markers to assess genetic relationships in plants. Based on DArTseq data, we studied genetic patterns in plants described as M. splendida or M. viridis in northern Viet Nam. The Musa individuals clustered based on population origin in either the PCoA plot, fastStructure barplot, and phylogenetic tree, showing a strong isolation-by-distance pattern, rather than based on taxon identity. Neither Musa splendida nor M. viridis were grouped into a separate monophyletic group. Individuals from one species were scattered over multiple clades in the tree, clustering together with other individuals from the same population and not with individuals with the same taxon ID. Consequently, the distinct taxonomic delimitation of M. splendida and M. viridis is uncertain. According to their species descriptions, the main difference between these two taxa is the color of the bracts of the male flower bud that ranges from red and (pinkish)-red (M. splendida) to pink (M. viridis) [22–24]. Variation in male bract color of banana plants is often caused by differences in the composition of anthocyanins [51, 52]. Although the proportion of different types of anthocyanins in bracts may vary across individuals of the same species, their resulting coloration patterns were so far always considered to be relatively stable within one species [51, 53]. Therefore, the color of male flower bracts was assumed to be a character with high taxonomic value in Musaceae [51]. Nevertheless, variation in flower bract color has also been observed in M. acuminata subsp. banksii (F.Muell.) N.W.Simmonds, demonstrating that color variation in flower bracts can even occur within a Musa subspecies [54]. Flower color is determined by a variety of abiotic and biotic factors including pollinator diversity, solar radiation, and mean annual rainfall [55]. Different environmental conditions may slightly alter the color of flower bracts, resulting in intraspecific variation. For example, a study in the Epimedium L. genus (Berberidaceae) found large intraspecific variation in flower color, showing that flower color is an unsuitable character for species delimitation in this plant genus [56]. Analogously, the variation in the color of male flower bracts is not a stable character to classify plants assigned to M. splendida and M. viridis into two distinct species. Our field observations of the shape of the leaf lamina basis of Musa plants showed that these characters are also not stable across individuals assigned to the same species. In particular, the shape of the leaf lamina basis varied between inner and other vegetative leaves and was often deformed due to mechanical damage. This genetic study provides evidence that supports the aggregation of all plants in our sampling into one species named M. splendida. This name has priority over M. viridis because it is the oldest available name (ICN Art. 11.3) [57].
The output of the PCoA and fastStructure analysis showed that the Musa populations are aggregated into larger genetic groups. This subdivision predominantly corresponded to their geographic distribution across the study area and the topological variation in the landscape created by the courses of the Chay River and Red River. Subpopulations 1, 2, and 3 in the fastStructure analyses were sampled at a considerable distance from each other in an area east of the Chay River, whereas subpopulation 4 was located east of the Red River. A pronounced population structure was also found in M. balbisiana and M. itinerans Cheesman populations from Vietnam, Laos, and China, which could also be linked to geographic variation across their distribution area caused by river basins and mountain ranges [18, 37, 58]. Moreover, similar patterns of genetic differentiation in Zingiber corallinum Hance populations (Zingiberaceae) from Hainan were attributed to their isolated position on mountain ranges separated by the Changhua river valley [59]. Consequently, river systems and their associated topological variation in northern Viet Nam and neighboring regions seem to form an effective seed and pollen dispersal barrier for several Zingiberales species. However, other environmental factors such as climate variation may additionally determine population genetic structure in Musa species, as shown for Zingiber nudicarpum D.Fang [59]. The Musa populations in our study displayed a strong pattern of isolation by distance. Such patterns were not observed for Vietnamese M. balbisiana populations, which was attributed to human translocations of M. balbisiana plants for nutritional purposes or fiber production [18, 37]. Although M. splendida and M. viridis are used as ornamental species and as a flavor enhancer of rice wine, human translocations may not substantially have changed isolation-by-distance patterns in these taxa. Most populations in our sampling were not admixed, corroborating the limited degree of admixture that was also found in M. balbisiana populations sampled from the same geographic area [18]. A study of Vietnamese ginseng populations (Panax vietnamensis Ha & Grushv.) found evidence for much higher levels of admixture, suggesting that admixture levels vary among different plant groups in Viet Nam [60]. Only populations 2 and 6 displayed considerable evidence for admixture between two genetic groups across all individuals within these populations. The intermediate geographic location of population 6 between subpopulations 3 and 4 may support the possibility of an admixed origin of this population, although these patterns can also be explained by isolation-by-distance [61] Important to note is that we did not find a significant signal of hybridization among the different lineages in our sampling. Hence, these patterns of admixture could not directly be linked to the M. splendida or M. viridis individuals.
Given the high morphological resemblance between M. splendida and other Musa species, a more elaborate study of the genetic relationships between M. splendida and closely related congeneric species may clarify their taxonomic status. As several Musa species were described based on one or a few herbarium specimens, the intraspecific variation in those species has often been ignored. For instance, the species M. exotica R.V.Valmayor is only found in the Cuc Phuong Forest Reservation in the Ninh Binh province in Viet Nam, where it was assumedly planted as an ornamental [21, 23]. As no wild populations of this species are currently known, its conservation status is unclear. It differs morphologically from M. viridis by its orange-red bracts and yellow fruits [23]. Nevertheless, as the color of male flower bracts is seemingly not a stable morphological character of the Musa species analyzed in this study, the value of this character to discriminate between other Musa species is questionable. A thorough assessment of all Callimusa species in southern China and northern Viet Nam based on multiple individuals per species and on a combination of morphological and genetic data is necessary to correctly identify the species relationships within this group of banana species.
Conclusions
Plants identified as M. viridis or M. splendida from northern Viet Nam are genetically more related to plants from the same population than to plants with the same taxon identity. Consequently, we propose to describe all plants assigned to these two taxa in Viet Nam as M. splendida. The results of this study show that the color of male flower bracts may not always be a reliable character for the delimitation of species in the Musa genus, advocating for genetic assessments of other species groups within this genus that mainly differ based on variation within a single character. Taxonomic and molecular research on wild Musa species may benefit from the development of an accessible ex situ living collection of well-documented reference plants that are morphologically and genetically completely characterized. Such efforts may overcome the lack of high-quality herbarium specimens of Musa species. Furthermore, given the evidence for genetic structure and isolation-by-distance in wild Musa populations, future studies may focus on the inference of correlations between genetic patterns and environmental factors to clarify which factors contribute to the genetic structure in Musa species. In addition, research efforts might be more directed towards the study of gene flow within and between Musa species, disentangling the effects of admixture and hybridization from isolation-by-distance patterns. Such insights would be valuable for the in-situ conservation of these banana genetic resources.
Taxonomic treatment
Musa splendida A.Chev., Rev. Bot. Appl. Agric. Trop. 14: 517. 1934. emend. N.S.Lý et al. (Lý & al. 2018: 285)–Lectotype: VIET NAM. Tonkin: Laokay province, Muong-Xen, 4 Dec 1913, A. Chevalier s.n. (P barcode P01767056!).–Epitype (designated by Lý et al. in Phytotaxa 351: 285. 2018):–VIET NAM. Haut-Tonkin: Laokay Province, Phu Lu, 6 Dec 1935, M. Poilane 24969 (P barcodes P00742068!, P00742069! [mounted on two sheets]).
= Musa viridis R.V.Valmayor, Đ.D.Lê & Häkkinen, Philipp. Agric. Sci. 87(1): 115. 2004–Holotype: VIET NAM. Van Chan: Yen Bai province, 29 Nov 1994, L. D. Danh, 1–052 (PHH barcode PHH002), syn. nov.
Emended description
The description is the same as [23], but emended as follows: leaf basis rounded to cuneate on both sides; male bract orange-red to bright red or pink-lilac on both surfaces, fading to yellow towards the base.
Supporting information
S1 Fig. Plots of the second and third principal coordinates (PC) that were constructed based on the haplotype variation across the 121 Musa individuals.
The second and third coordinate explained 7.7% and 7.6% of the total variation in the dataset, respectively. Individuals were colored based on population ID (upper plot) and taxon ID (lower plot).
https://doi.org/10.1371/journal.pone.0318252.s001
(PDF)
S1 Table. Overview of the Musa plants included this study.
Collection ID, taxon ID, country, district, and locality of origin, longitude and latitude coordinates, population ID, and the color of their male flower bracts are listed.
https://doi.org/10.1371/journal.pone.0318252.s002
(DOCX)
Acknowledgments
We want to thank the research team of the Plant Resources Center of Viet Nam for their support in the collection missions related to this work and to the team of Diversity Arrays Technology and to Wim Baert from Meise Botanic Garden for their lab assistance.
References
- 1. Dayrat B. Towards integrative taxonomy. Biol J Linn Soc. 2005;85:407–415.
- 2. Duminil J, Di Michele M. Plant species delimitation: A comparison of morphological and molecular markers. Plant Biosyst. 2009;143:528–42.
- 3. Westerband AC, Funk JL, Barton KE. Intraspecific trait variation in plants: a renewed focus on its role in ecological processes. Ann Bot. 2021;127:397–410. pmid:33507251
- 4. Waswa EN, Mutinda ES, Mkala EM, Katumo DM, Oulo MA, Odago WO, et al. Understanding the Taxonomic Complexes and Species Delimitation within Sambucus L. (Viburnaceae). Divers. 2022;14:906.
- 5. El-Bakatoushi R, Alframawy AM, Samer M, El-Sadek L, Botros W. Evolution of the Portulaca oleracea L. aggregate in Egypt on molecular and phenotypic levels revealed by morphology, inter-simple sequence repeat (ISSR) and 18S rDNA gene sequence markers. Ecol Plants. 2013;208:464–77.
- 6. Li-E Y, Meng Y, Peng DL, Nie ZL, Sun H. Molecular phylogeny of Galium L. of the tribe Rubieae (Rubiaceae)–Emphasis on Chinese species and recognition of a new genus Pseudogalium. Mol Phylogenet Evol. 2018;126:221–32. pmid:29627517
- 7. Franzoni J, Astuti G, Bartolucci F, Bernardo L, Fior S, Casazza G, et al. Different species or altitudinal morphotypes? Testing the taxonomic value of Dianthus brachycalyx (Caryophyllaceae). Taxon. 2024;00:1–14.
- 8. Larridon I, Zuntini AR, Léveillé-Bourret E, Barret RL, Starr JR, Muasya AM, et al. A new classification of Cyperaceae (Poales) supported by phylogenomic data. J Syst Evol. 2021;59:852–95.
- 9. Schlick-Steiner BC, Steiner FE, Seifert B, Stauffer C, Erhard C, Crozier RH. Integrative taxonomy: a multisource approach to exploring biodiversity. Annu Rev Entomol. 2010;55:421–438. pmid:19737081
- 10. Vences M, Mirallas A, Dufresnes C. Next-generation species delimitation and taxonomy: Implications for biogeography. J. Biogeogr. 2024; 51:1709–1722.
- 11.
FAO. FAOSTAT, last update 23 Dec 2023. https://www.fao.org/faostat/en/#home (Accessed 24 Sept 2024).
- 12. Sardos J, Breton C, Perrier X, Van den Houwe I, Carpentier S, Paofa J, et al. Hybridization, missing wild ancestors and the domestication of cultivated diploid bananas. Front. Plant Sci 2022;13:969220. pmid:36275535
- 13. Martin G, Cottin A, Baurens FC, Labadie K, D’Hont A. Resolving the Musa acuminata complex with long-read sequencing. Plant J. 2021;105:136–154.
- 14. Häkkinen M. Reappraisal of sectional taxonomy in Musa (Musaceae). Taxon. 2013;62:809–13.
- 15. Argent G. Two interesting wild Musa species (Musaceae) from Sabah, Malaysia. Gard Bull Singapore. 2000;52:203–210.
- 16. Liu AZ, Li DZ, Li XW. Taxonomic notes on wild bananas (Musa) from China. Bot Bull Acad Sin. 2002;43:77–81.
- 17. Häkkinen M, Väre H. Typification and check-list of Musa L. names (Musaceae) with nomenclatural notes. Adansonia. 2008;30:63–112.
- 18. Mertens A, Bawin Y, Vanden Abeele A, Kallow S, Vu DT, Le LT, et al. Genetic diversity and structure of Musa balbisiana populations in Vietnam and its implications for the conservation of banana crop wild relatives. PLoS ONE. 2021;16:e0253255. pmid:34161368
- 19. Janssens SB, Vandelook F, De Langhe E, Verstraete B, Smets E, Vandenhouwe I, 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. 2016;210:1453–1465. pmid:26832306
- 20. Vu TD, Vu DT, Janssens SB, De Langhe E, Le LT, Kallow S, et al. The description, distribution and habitat of wild banana species in northern Viet Nam. Genet Resour Crop Evol. 2023;70:479–504.
- 21.
Sachter-Smith G. The Wild Bananas. A catalogue of wild Musa species and tribute to the work of Markku Häkkinen. 1st ed. Rome: Bioversity International; 2023.
- 22. Chevalier A. Observations sur quelques Bananiers sauvages et cultivés. Rev Bot Appl Agric Colon. 1934;14:506–521.
- 23. Valmajor R, Danh D, Häkkinen M. Rediscovery of Musa splendida A. Chevalier and description of two new species (Musa viridis and Musa lutea). Philipp. Agric. Sci. 2004;87:110–118.
- 24. Lý NS, Lowry PP, Haevermans T. Typification and an emended description of Musa splendida (Musaceae). Phytotaxa. 2018;351:281–288.
- 25. Sardos J, Breton C, Perrier X, Van den Houwe I, Carpentier S, Paofa J, et al. 2022. Hybridization, missing wild ancestors and the domestication of cultivated diploid bananas. Front. Plant Sci. 2022;13:969220.
- 26. Lý NS, Lê CK, Triệu TD, Haevermans A, Lowry PP II, Haevermans T. A distinctive new species of wild banana (Musa, Musaceae) from northern Vietnam. Phytotaxa. 2012;75:33–42.
- 27. Doyle JJ, Doyle JL. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem Bull. 1987;19:11–15.
- 28. Kilian A, Wenzl P, Huttner E, Carling J, Xia L, Blois H, et al. Diversity arrays technology: a generic genome profiling technology on open platforms. Methods Mol Biol. 2012;888:67–89. pmid:22665276
- 29. Martin G, Baurens F-C, Droc G, Rouard M, Cenci A, Kilian A., et al. Improvement of the banana “Musa acuminata” reference sequence using NGS data and semi-automated bioinformatics methods. BMC Genomics. 2016;17:243. pmid:26984673
- 30. Eyland D, Breton C, Sardos J, Kallow S, Panis B, Swennen R, et al. Filling the gaps in gene banks: Collecting, characterizing, and phenotyping wild banana relatives of Papua New Guinea. Crop Sci. 2020;1–20.
- 31. Akech V, Bengtsson T, Ortiz R, Swennen R, Uwimana B, Ferreira C F, et al. Genetic diversity and population structure in banana (Musa spp.) breeding germplasm. Plant Genome. 2024;e20497. pmid:39075664
- 32.
OpenStreetMap contributors. OpenStreetMap. Available under the Open Database Licence from: https://www.openstreetmap.org.
- 33. Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.J. 2011;17:10–12.
- 34.
Schaumont D. 2020. GBprocesS: Genotyping-by-Sequencing Data Processing Toolkit; 2020 [cited 2024 Oct 14] Software [Internet]. https://gitlab.com/dschaumont/GBprocesS
- 35. Belser C, Baurens FC, Noel B, Martin G, Cruaud C, Istace B, et al. Telomere-to-telomere gapless chromosomes of banana using nanopore sequencing. Commun Biol. 2021;4:1047. pmid:34493830
- 36.
Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. arXiv:1303.3997v2 [Preprint]. 2013 [cited 2024 Oct 14]. https://arxiv.org/abs/1303.3997
- 37. Jenny C, Sachter-Smith G, Breton C, Rivallan R, Jacquemoud-Collet J-P, Dubois C, et al. Musa species in mainland Southeast Asia: From wild to domesticate. PLoS ONE. 2024;19:e0307592. pmid:39356650
- 38. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25:2078–2079. pmid:19505943
- 39.
Broad Institute. Picard Toolkit; 2019 [cited 2024 Oct 14] GitHub Repository [Internet]. https://broadinstitute.github.io/picard/
- 40. McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The Genome Analysis Toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20:1297–1303. pmid:20644199
- 41. Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, et al. The variant call format and VCFtools. Bioinformatics. 2011;27:2156–2158. pmid:21653522
- 42. Depecker J, Verleysen L, Asimonyio J, Hatangi Y, Kambale J-L, Mwanga Mwanga I, et al. Genetic diversity and structure in wild Robusta coffee (Coffea canephora A. Froehner) populations in Yangambi (DR Congo) and their relation to forest disturbance. Heredity. 2023;130:145–153. pmid:36596880
- 43. Zewdie B, Bawin Y, Tack AJM, Nemomissa S, Tesfaye K, Janssens SB, et al. Genetic composition and diversity of Arabica coffee in the crop’s centre of origin and its impact on four major fungal diseases. Mol. Ecol. 2023;32:2484–2503. pmid:35377502
- 44. Schaumont D, Veeckman E, Van der Jeugt F, Haegeman A, Van Glabeke S, Bawin Y, et al. Stack Mapping Anchor Points (SMAP): a versatile suite of tools for read-backed haplotyping. BioRxiv [Preprint]. 2022;BioRxiv 2022.03.10.483555 [cited 2024 Oct 14]: [9 p.]. https://www.biorxiv.org/content/10.1101/2022.03.10.483555v1
- 45. Jaccard P. The distribution of the flora in the alpine zone. New Phytol. 1912;11:37–50.
- 46.
R Core Team. R: A Language and Environment for Statistical Computing; 2023. R Foundation for Statistical Computing, Vienna, Austria.
- 47. Raj A, Stephens M, Pritchard JK. fastSTRUCTURE: variational inference of population structure in large SNP data sets. Genet. 2014;197:573–589. pmid:24700103
- 48. Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A. et al. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. MBE 2020;37:1530–1534. pmid:32011700
- 49. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods. 2017;14:587–589. pmid:28481363
- 50. Zhang C, Rabiee M, Sayyari E, Mirarab S. ASTRAL-III: polynomial time species tree reconstruction from partially resolved gene trees. BMC Bioinformatics 2018;19:15–30. pmid:29745866
- 51. Simmonds NW. Anthocyanins in Bananas. Ann. Bot. 1954;18:471–482.
- 52. Pazmiño-Durán A, Giusti MM, Wrolstad RE, Glória MBA. Anthocyanins from banana bracts (Musa x paradisiaca) as potential food colorants. Food Chem. 2001;73:327–332.
- 53. Horry JP, Jay M. Distribution of anthocyanins in wild and cultivated banana varieties. Phytochemistry. 1988;27:2667–2672.
- 54. Argent GCG. The wild bananas of Papua New Guinea. Notes Roy. Bot. Gard. Edinburgh. 1976;35:77–114.
- 55. Dalrymple RL, Kemp DJ, Flores-Moreno H, Laffan SW, White TE, Hemmings FA, et al. Macroecological patterns in flower colour are shaped by both biotic and abiotic factors. New Phytol. 2020;228:1972–85. pmid:32533864
- 56. Xu X, Liu L, Liu S, He Y, Li R, Ge F. The taxonomic relevance of flower colour for Epimedium (Berberidaceae), with morphological and nomenclatural notes for five species from China. Phytokeys 2019;118:33–64. pmid:30853835
- 57.
Turland NJ, Wiersema JH, Barrie FR, Greuter W, Hawksworth DL, Herendeen PS, et al. International Code of Nomenclature for algae, fungi, and plants (Shenzhen Code) adopted by the Nineteenth International Botanical Congress Shenzhen, China, July 2017. Regnum Veg. 159. Glashütten: Koeltz Botanical Books; 2018.
- 58. Ge XJ, Liu MH, Wang WK, Schaal BA, Chiang TY. Population structure of wild bananas, Musa balbisiana, in China determined by SSR fingerprinting and cpDNA PCR-RFLP. Mol. Ecol. 2005;14:933–944. pmid:15773926
- 59. Huang R, Chu Q-H, Lu G-H, Wang Y-Q. Comparative studies on population genetic structure of two closely related selfing and outcrossing Zingiber species in Hainan Island. Sci. Rep. 2019;9:17997. pmid:31784623
- 60. Le HTT, Nguyen LN, Pham HLB, Le HTM, Luong TD, Huynh HTT, et al. Target Capture Reveals the Complex Origin of Vietnamese Ginseng. Front Plant Sci. 2022;13:814178. pmid:35909770
- 61. Wiens BJ, Colella JP. That’s Not a Hybrid: How to Distinguish Patterns of Admixture and Isolation By Distance. Mol. Ecol. Res. 2024;0:e14039. pmid:39467042