Search tips
Search criteria 


Logo of eceLink to Publisher's site
Ecol Evol. 2017 September; 7(18): 7311–7333.
Published online 2017 August 9. doi:  10.1002/ece3.3220
PMCID: PMC5606854

Utility of DNA barcoding to identify rare endemic vascular plant species in Trinidad


The islands of the Caribbean are considered to be a “biodiversity hotspot.” Collectively, a high level of endemism for several plant groups has been reported for this region. Biodiversity conservation should, in part, be informed by taxonomy, population status, and distribution of flora. One taxonomic impediment to species inventory and management is correct identification as conventional morphology‐based assessment is subject to several caveats. DNA barcoding can be a useful tool to quickly and accurately identify species and has the potential to prompt the discovery of new species. In this study, the ability of DNA barcoding to confirm the identities of 14 endangered endemic vascular plant species in Trinidad was assessed using three DNA barcodes (matK, rbcL, and rpoC1). Herbarium identifications were previously made for all species under study. matK, rbcL, and rpoC1 markers were successful in amplifying target regions for seven of the 14 species. rpoC1 sequences required extensive editing and were unusable. rbcL primers resulted in cleanest reads, however, matK appeared to be superior to rbcL based on a number of parameters assessed including level of DNA polymorphism in the sequences, genetic distance, reference library coverage based on BLASTN statistics, direct sequence comparisons within “best match” and “best close match” criteria, and finally, degree of clustering with moderate to strong bootstrap support (>60%) in neighbor‐joining tree‐based comparisons. The performance of both markers seemed to be species‐specific based on the parameters examined. Overall, the Trinidad sequences were accurately identified to the genus level for all endemic plant species successfully amplified and sequenced using both matK and rbcL markers. DNA barcoding can contribute to taxonomic and biodiversity research and will complement efforts to select taxa for various molecular ecology and population genetics studies.

Keywords: biodiversity conservation, DNA barcoding, endemics, molecular genetics


Trinidad and Tobago (latitudes 10.0°N and 11.3°N and longitudes 60.3°W and 62°W) are the southernmost islands in the Caribbean and are bordered by the Atlantic Ocean and Caribbean Sea. The islands are located 11.3 and 32 km northeast of the Venezuelan coast of South America (Lesser Antilles; Kenny, Comeau, & Katwaru, 1997; Kenny, 2008). Although both islands are positioned on the South American Continental Shelf, it was proposed that Trinidad separated from the South American continent later (ca. 1,500 years ago) than Tobago (ca. 11,000–13,000 years ago; Kenny, 2008). The flora of Trinidad and Tobago is estimated to include 2, 407 vascular plant species of which approximately 4.5% are endemic (Baksh‐Comeau et al., 2016). This level of endemism is reflective of the close proximity to, and the relatively short geological time frame since separation from the South American mainland, and cannot be compared with oceanic islands in the Greater Antilles, such as Jamaica (12%–50% endemism; MacArthur, 1972; van den Eynden, Oatham, & Johnson, 2008; Baksh‐Comeau et al., 2016).

The Caribbean islands are among the world's most important “biodiversity hotspots” (Mittermeier et al., 2004; Shi, Singh, Kant, Zhu, & Waller, 2005), and, in global terms, can be compared to the Madagascar and Cape Floristic hotspots in terms of the number of endemic genera (Francisco‐Ortega et al., 2007; Maunder et al., 2008; Mittermeier et al., 2004). Based on the International Union for Conservation of Nature (IUCN) Red List categories, and the Global Star rating system, species located in hotspots of high conservation value should be inventoried to assess the distribution and population status of endemics (Baksh‐Comeau et al., 2016). The conservation of natural plant resources in the Caribbean is especially critical for providing essential ecosystem services (Kress & Horvitz, 2005). However, in many biodiversity hotspots, the botanical inventory is usually incomplete, perhaps because taxonomic assignment is frustrated by low discriminatory power of morphological descriptors for very closely related species (Francisco‐Ortega et al., 2007; Zanoni, 1989). DNA barcoding has the potential to support species identification and discovery, vegetation, and floristic species surveys, in addition to studies on ecological forensics, all of which are critical to biodiversity management (Hollingsworth, Li, van der Bank, & Twyford, 2016; Valentini, Pompanon, & Taberlet, 2008; von Crautlein, Korpelainen, Pietiläinen, & Rikkinen, 2011).

Apart from biodiversity conservation, accurate taxonomic assignment is important to the practice of traditional or herbal medicine (Techen, Parveen, Pan, & Khan, 2014). In the Caribbean, herbal remedies are referred to as “bush medicine” (Laguerre, 1987; Mahabir & Gulliford, 1997; Quinlan & Quinlan, 2007). In Trinidad, approximately one‐third of the flora is composed of exotic species which are used as bush medicines according to an ethno‐botanical survey conducted between 2007 and 2008 (Clement, Baksh‐Comeau, & Seaforth, 2015). The danger of collecting plants for use as herbal remedies lies in some medicinal plant species having multiple synonyms, in addition to having a vernacular name, which may be mistakenly used to identify more than one plant species (Bellakhdar, Claisse, Fleurentin, & Younos, 1991). Endangered species may be mistakenly collected to extinction if their identity is confused with more abundant morphologically similar‐looking individuals. Consumption of plant material from misidentified species could also result in serious health risks to end users (Barthelson, Sundareshan, Galbraith, & Woosley, 2006; Bruni et al., 2010; Mahabir & Gulliford, 1997). For example, the Food and Drug Association (FDA) has advised that consumption of products containing aristolochic acid (derived from plants belonging to the Aristolochia genus) has been associated with permanent kidney damage, and development of certain cancers associated with the urinary tract ( Similarly, toxic effects have been reported for fruit and leaf consumption of plant species of the genus Ilex (Weiner & Weiner, 1999) and Maytenus (Da Silva, Serrano, & Silva, 2011). Aristolochia, Ilex, and Maytenus sp. are used in the Caribbean for their proposed medicinal properties (Mahabir & Gulliford, 1997). Preservation of indigenous knowledge concerning medical ethno‐botany is a key aspect of bioprospecting with the proviso of accurate species identification (Harvey & Gericke, 2011; Kumar, Sharma, & Chattopadhyay, 2013; Theodoridis et al., 2012).

Conventionally, taxonomic assignment has been the purview of taxonomic experts (Waugh, 2007), however, DNA barcoding may enable rapid and accurate species identification by nonspecialists using nucleotide comparisons of approved gene regions (Coissac, Hollingsworth, Lavergne, & Taberlet, 2016; Hebert, Cywinska, Ball, & deWaard, 2003). A 648‐basepair region of the mitochondrial cytochrome c oxidase 1 gene (“CO1”) is the accepted barcode for almost all animal groups but, it is not a useful barcode in plants because this region (i) has a slow rate of evolution (Chase & Fay, 2009), (ii) is prone to structural rearrangements (Kelly, Ameka, & Chase, 2010; Palmer et al., 2000), and (iii) does not accommodate for the existence of interspecific and intergeneric hybrids in plants (Rieseberg, Wood, & Baack, 2006). Selection of a plant DNA barcode must meet a number of criteria which have already been described elsewhere (Ford et al., 2009; Hollingsworth et al., 2009; Kress & Erickson, 2008; Li et al., 2015). The chloroplast ribulose‐1, 5‐bisphosphate carboxylase/oxygenase large subunit gene (rbcL) and maturase K gene (matK) are the approved barcodes for land plants (CBOL Plant Working Group 2009). However, plant‐plastid barcodes typically have lower resolving power to separate closely related plant species compared to the animal barcode, and in several cases, conspecifics or recently diverged species do not form highly supported, distinct sequence clusters that allow species discrimination (Hollingsworth et al., 2016; van Velzen, Weitschek, Felici, & Bakker, 2012; Zhang et al., 2012). In fact, a uniquely identified species in a given genus is the exception rather than the rule in most plant barcoding studies (Hollingsworth, Graham, & Little, 2011). For these reasons, standard plant barcodes are more appropriately used as “molecular augmentations” to preexisting herbarium identifications as the current plant barcode sequences do not contain sufficient variation to define a species‐level framework for every plant species (Hollingsworth et al., 2016). Further, in using plant barcodes, it is important to understand the limited resolving power of the technique when formulating the objectives of a particular study (Hollingsworth et al., 2016). There are other practical issues to consider which include but are not limited to, the requirement for species‐specific primer combinations which directly determines recovery of matK sequences, DNA extraction methods for recalcitrant species whose genomic DNA may be contaminated with PCR inhibitors, for example, muco‐polysaccharides, proteins, polyphenols, and tannins, the need for and expense involved in automated DNA extraction for high‐throughput processing of large sample sizes, and the difficulty in constructing reference sequence datasets or libraries (Hollingsworth et al., 2011, 2016). The most recent development in plant DNA barcoding is to sequence the complete chloroplast genome which will be used as a “super‐barcode” in order to overcome some of the issues associated with low resolving power of the single or multiple loci barcode approach (Hollingsworth et al., 2016; Li et al., 2015).

In this study, we evaluated the ability of three DNA barcodes (matK, rbcL, and rpoC1) to identify specimens of 14 vascular endemic plant species in Trinidad which are endangered or vulnerable according to The International Union for Conservation of Nature and Natural Resources (IUCN) Red List criteria.


2.1. Plant collection

Fourteen endemic vascular plant species were selected for this study (Table 1; Figure 1). Expeditions were led by Mr. Winston Johnson, retired field plant taxonomy expert of the National Herbarium of Trinidad and Tobago. The main consideration for collection was the fact that the majority of species collected were endangered according to the IUCN Red List criteria (Baksh‐Comeau et al., 2016), and this, therefore, restricted the number of individuals collected. In addition, some mountainous species were difficult to retrieve and accessibility was an issue. Five individuals per species per location were collected in labeled bags and transported on ice to the laboratory. Specimen identification and species assignment were independently confirmed prior to DNA analysis and was based on an assessment of morphological descriptors developed by the National Herbarium of Trinidad and Tobago ( Voucher specimens were deposited at the National Herbarium of Trinidad and Tobago. Information concerning the endemic species used in this study can be accessed through The National Herbarium of Trinidad and Tobago in conjunction with the University of Oxford through the Darwin Initiative online database,

Figure 1

Location map of endemic vascular plant species sampled in this study. The white shaded areas indicate elevation, and it is noted that the majority of endemic species included in this study were located in mountainous regions in North Trinidad

Table 1

Plant species collection data

2.2. DNA extraction, amplification, and sequencing

Total genomic DNA was extracted from freshly collected leaf material according to the modified CTAB protocol of Koboyashi, Horikoshi, Katsuyama, Handa, and Takayanagi (1998). The Kobayashi protocol was selected because it is known to successfully extract amplifiable genomic DNA from a number of woody plant tissue with high amounts of polysaccharides including muco‐polysaccharides, polyphenols, and various secondary metabolites such as alkaloids, flavonoids, and phenols, all of which inhibit PCR amplification (Kobayashi et al. 1998). In some cases, the DNA pellet had to be washed up to three times with Buffer 1 and the chloroform‐isoamyl alcohol extraction step was repeated if the aqueous layer was not clear and/or the color of the pellet was brown and sticky according to the recommendations of Koboyashi et al. (1998). DNA extracts were diluted to 10 ng/μl and this served as the initial working DNA concentration for PCR amplification.

Three markers were assessed for species identification: the recommended two‐locus cpDNA barcode (matK + rbcL; CBOL Plant Working Group 2009), and one cpDNA regions (rpoC1; Chase et al., 2005; Kress, Wurdack, Zimmer, Weigt, & Janzen, 2005; Table 2). PCR amplification reagents and thermal conditions were used according to the CBOL laboratory manual guidelines ( plant_working_group.html). However, optimization was required for each primer pair and for each species which included the following: (i) varying the concentrations of DMSO, Tween 20, and BSA as PCR enhancers, (ii) titrating the concentration of MgSO4, and (iii) assessment of primer annealing temperature through gradient annealing temperature analysis. Optimal amplification was achieved using 1.5 mmol/L MgSO4, 1% Tween‐20, 0.8 mg/ml BSA, and 60°C annealing temperature.

Table 2

Primer data

Amplicons were sequenced in both directions by Amplicon Express (Pullman, WA, USA) using Sanger dideoxy sequencing. Verified base calls were carried out by the sequencing company independently as a first check that the sequence reads were correct. The sequences were also checked against the chromatograms using Sequencher v 5.4.1 (, Gene Codes Corp., Ann Arbor, Michigan, USA). The data from the rpoC1 marker was eliminated from further analysis as the sequences were not clean reads despite repeated sequencing attempts. Sequences obtained from the two barcodes were deposited in GenBank (GenBank Accession Nos. KX228511 to KX228515 and KX212893 to KX212899).

2.3. Data analysis

Data were analyzed using a compendium of supporting methods as there is no one singular approach that best determines barcoding success in species discrimination (Gong, Liu, Chen, Hong, & Kong, 2016; Mao, Zhang, Nakamura, Guan, & Qiu, 2014). As far as possible, the recommendations for data analysis were followed as outlined by Casiraghi, Labra, Ferri, Galimberti, and De Mattia (2010) and Collins and Cruickshank (2013).

2.4. BLAST and reference datasets

Verified representative sequences of each taxon were provisionally identified using the BLASTN algorithm available on NCBI ( The similarity indices and query coverage were recorded. A reference sequence library was then constructed for each species which consisted of sequences matching 98%–100% in sequence similarity with 97%–100% query coverage (Larranaga & Hormaza, 2015).

There were significantly different nucleotide lengths sizes for each barcode and the two‐locus combinations prevented acceptable alignment for several species. It was, therefore, difficult to construct multilocus barcodes using the same alignment associated with the corresponding single‐locus barcodes for each taxon in the reference dataset. This difficulty has been reported by others (Hollingsworth et al., 2016). As such, analyses were conducted for separate barcodes.

Sequences were aligned using MAFFT v7 (, Katoh, 2013). The aligned sequences were examined visually, and manual adjustments were made to ensure common start and end lengths. MEGA7 software (Tamura et al. 2011) was used to calculate pairwise distances among the aligned sequences using the Kimura 2‐parameter model (Kimura, 1980) to assess intra‐ and interspecies differences.

2.5. DNA polymorphism analysis

The level of DNA polymorphism of the aligned sequences of each reference sequence dataset was carried out using DnaSP software (; Rozas, 2009; Librado & Rozas, 2009). DNA polymorphism analysis is an approach that can potentially identify important diagnostic differences among sequences that are not detected by distance or tree‐based query assignment methods (DeSalle, Egan, & Siddall, 2005; Pettengill & Neel, 2010). This approach has not been previously applied for analyzing barcode sequences, and this is the first reported use here.

2.6. Direct sequence comparison

The “Species Identifier” suite of tools in the Taxonomy‐aware DNA sequence processing toolkit (TaxonDNA;; Meier, Kwong, Vaidya, & Ng, 2006) was used to explore intra‐ and interspecific genetic distances, matching sequences, and clustering sequences based on pairwise distances. Genetic distance data was used to implement a threshold for determining species identity. This threshold represented the pairwise genetic distance at which 95% of all conspecific individuals were correctly classified. There must be multiple accessions of most species in the reference sequence database, and conspecifics should be present in the database in order to apply the threshold value. BLAST analysis produced similarity hits with sequences belonging to genera outside of the query sequence and as such, demonstrated close genetic affinities to other congeneric species. These sequences were, therefore, also included in the reference sequence dataset.

To evaluate species identity success, the criteria “Best Match” and “Best Close Match” implemented in TaxonDNA were evaluated. “Best Match” is designated if the query sequence is assigned to the genus of the most similar reference sequence. “Best Close Match” is designated if the query sequence is assigned to the genus of the most similar library sequence based on K‐2‐P distance threshold. A query that falls below the determined threshold value will be classified as unidentified (“no match”). The Barcode of Life Data Systems (BOLD), assigns identities using a pairwise genetic distance threshold of 1% for animal species (Ratnasingham & Hebert, 2007). However, the threshold has to be determined for each taxon as it is expected that there is no common threshold value across several different taxonomic groups (Rach, DeSalle, Sarkar, Schierwater, & Hadrys, 2008).

Assignment of each query sequence to a specific taxon was attempted with three possible outcomes for “Best Match” and “Best Close Match” analyses: (i) A “correct” assignment (i.e., the query was assigned to a taxon), (ii) an “ambiguous” assignment (i.e., if there were no barcodes in the library within the set threshold, the assignment was considered to be “ambiguous”), and an “incorrect” assignment (i.e., the query was not assigned to a taxon). A “correct” assignment was then checked with the morphology‐based identification and if there was concordance, the assignment was considered to be TRUE, or FALSE if there was disagreement with the morphology‐based identification (Ross, Murugan, & Li, 2008; Wilson et al., 2011). An “ambiguous” assignment was concluded where the true taxon based on morphological descriptors was not represented in the reference dataset for that barcode which meant that the library was incomplete with inadequate coverage for that specific taxon (Ross et al., 2008; Wilson et al., 2011).

2.7. Tree‐based analysis

Neighbor‐joining (NJ; Saitou & Nei, 1987) analysis was carried out to determine phylogenetic placement of a query sequence in relation to a reference sequence dataset. Bootstrap values >70% at a given branch were considered strong support for the existence of that branch. The model implemented was the K‐2‐P genetic distance model with 1,000 pseudoreplicates (Felsenstein 1985). Using tree‐based criteria, query sequences are assigned to a species when they clustered with barcodes from their correct taxon with high bootstrap support (Elias et al., 2007; Mao et al., 2014). Controversy arises when correct assignment via this method of analysis requires that the taxon be monophyletic and when deep phylogenies cannot be tracked. The NJ algorithm is used here to determine clustering of closely related individuals and not as an absolute confirmation of taxon identification.


3.1. PCR and sequencing

Table 3 summarizes the outcome of PCR amplification and sequencing after optimization. With respect to sequencing, the rpoC1 PCR product was the most difficult to sequence. rbcL gave the cleanest reads compared to rpoC1 and matK. These sequences also had zero InDels for all species except Aristolochia. The rbcL sequences of the reference library mined from GenBank had fewer ambiguous bases (“M,” “S,” “Y,” “K,” “N,” “W”) compared to matK and were easily aligned.

Table 3

PCR amplification and sequencing success

3.2. DNA polymorphism analysis

Tables 4 and 5 summarizes the DNA polymorphism detected in the matK and rbcL sequences. matK sequences had a higher level of polymorphism and had more parsimony informative sites than rbcL sequences regardless of species. matK sequences also enabled a higher percentage of correct identifications compared to rbcL regardless of species. matK sequences yielded a higher number of BLASTN hits to the same genus as the query sequence regardless of species. matK sequences extracted from GenBank had “R,” “Y,” “N,” “S,” “K,” and “M” bases. matK sequences had a higher number of nucleotide differences than rbcL sequences. matK sequences in the reference library dataset had lower CT and C values compared with rbcL sequences especially for Aristolochia, Ilex, and Philodendron species. GenBank accession numbers for all references sequences used in this study are indicated in the NJ trees inferred for each species (Fig. 2a–e for matK sequences and Fig. 3a–f for rbcL sequences).

Figure 2

(a–e) Neighbor‐joining tree for five species based on matK sequences. Clustering of all query sequences of species under study was inferred using the neighbor‐joining method in MEGA6. The condensed tree (50% bootstrap consensus ...

Figure 3

(a–f) Neighbor‐joining tree for six species based on rbcL sequences. Clustering of all query sequences of species under study was inferred using the neighbor‐Joining method in MEGA6. The condensed tree (50% boot strap consensus ...

Table 4

DNA polymorphism data for the matK barcode

Table 5

DNA polymorphism data for the rbcL barcode

BLASTN resulted in mixed hits for rbcL sequences for 50% of the species under study. However, Aristolochia and Ilex sequences resulted in 100% genus hits for both matK and rbcL sequences. rbcL sequences extracted from GenBank were clean reads with few sequences having “R,” “Y,” “N,” “S,” “K,” and “M” bases. rbcL sequences of the endemic species also had a higher query coverage (99%–100%) and higher similarity (99%) compared with matK sequences. rbcL sequences also had fewer InDels than matK sequences and contained several regions or blocks of nucleotides with conserved sequences which resulted in high CT (CT = 1) and C values (close to 1).

3.3. Sequence identification

Specimen identification success for the two markers, matK and rbcL, is outlined in Tables 6 and 7. There was a higher proportion of correctly identified species obtained with matK sequences compared with rbcL sequences. rbcL sequences had a higher proportion of ambiguously classified sequences compared to matK sequences. The matK reference dataset also had a higher number of conspecifics compared with the rbcL reference dataset. In terms of genetic distance, Aristolochia sequences shared the highest distance and Metastelma sequences shared the lowest distance, regardless of marker, when compared with other species according to K‐2‐P analysis. All of the endemic species from Trinidad, except Clusia aripoensis, shared a common classification as “ambiguous” regardless of marker. Clusia aripoensis could not be matched according to TaxonDNA's “best match” and “best close match” criteria, and there was no placement of this species into genus‐specific cluster in the NJ tree generated by the K‐2‐P model.

Table 6

Kimura 2‐parameter threshold data and sequence matches in the reference library

Table 7

Species identification success based on best match and best close match analyses

3.4. Reference dataset coverage

Library representation and similarity in the BLAST reference library for each species and marker are summarized in Table 8 and 9. Coverage was representative for only Aristolochia, Ilex, and Philodendron matK sequences. Reference dataset coverage was representative for Aristolochia, Ilex, and Acalypha rbcL sequences. A minimum cutoff value for query coverage was applied at 97% if the subject sequence belonged to the same genus as the query sequence. Similarly, a minimum similarity value was applied at 97% if the subject sequence belonged to the same genus as the query sequence with the exception of Clusia aripoensis which only had 94% maximum similarity regardless of genus. BLASTN hits for Clusia aripoensis also revealed poor reference dataset coverage with the lowest similarity scores (94%) of all the species included in this study. The distances among the conspecifics in the dataset were also very low. This may explain the “no match” designation for Clusia aripoensis.

Table 8

Clustering analysis of matK sequences based on K‐2‐P genetic distances and the neighbor‐joining algorithm

Table 9

Clustering analysis of rbcL sequences based on K‐2‐P genetic distances and the neighbor‐joining algorithm

BLAST searches revealed very poor reference dataset coverage for all species for the rpoC1 marker. Aristolochia had only 11 sequences belonging to this genus (97%–100% query coverage; 97%–98% similarity); Acalypha had no other sequences belonging to this genus; Ilex had four sequences belonging to this genus (96%–100% query coverage; 98% similarity); Maytenus had 15 (98%–100% query coverage; 99% similarity); Metastelma had just five other sequences belonging to this genus (100% query coverage; 97%–98% similarity); Philodendron had no other sequences belonging to this genus. In most cases, the BLAST reference dataset also included other sequences belonging to at least 15 other genera designated as hits with the same query coverage and similarity as those sequences belonging to the same genus as the query sequence. Further analysis was, therefore, not carried out for this marker.

3.5. Clustering of query sequences in NJ trees

matK sequences allowed specific placement of Trinidad species within genus‐specific clusters with moderate to high bootstrap support (>60%–90%; Table 8; Fig. 2a–e), for all but one species, Ilex arimensis, which was not placed into a discernible cluster even though the reference dataset was representative (100%), query coverage and similarity of sequences in this reference dataset was optimal (94%–99%; 99%), but whose sequence variation was low for this marker (0.72%). Polytomies were also evident in the NJ tree generated for Ilex and Philodendron.

rbcL sequences did not allow specific placement of Trinidad species into genus‐specific clusters with moderate to high bootstrap support (>60%) except for Philodendron simmondsii (bs = 73%; Table 9; Fig. 3a–f). Although sequences generally were positioned in genus‐specific clusters, these were not well supported (bs < 50%). There were also polytomies in the NJ trees constructed for rbcL sequences for all species. K‐2‐P genetic distances were lower for rbcL sequences than for matK sequences.


In this study, we examined the suitability of the proposed CBOL Plant Working Group barcoding markers for land plants to confirm the identity of specimens of 14 endemic and rare vascular plant species in Trinidad. Our results indicated that 50% of the species under study were not identified using a barcoding approach due to amplification failure. It was evident that the quality of DNA was an important factor in amplification success and PCR failure may be a result of DNA quality and not necessarily poor primer annealing. The method of DNA extraction and quality of DNA are critical to successful amplification. Others explained amplification failure as a result of poor annealing with standard matK or rbcL primers and highlighted the need to redesign species‐specific primers (Kress & Erickson, 2007; Sass et al. 2007; Fazekas et al., 2008; Lahaye et al., 2008; Casiraghi et al., 2010; Roy et al., 2010). According to Casiraghi et al. (2010), matK sequences were analyzed in different plants but the universality of this barcode ranged from routine success to low recovery. Casiraghi et al. (2010) also acknowledged that even the most conserved rpoB, rpoC1, and rbcL or a portion of matK that demonstrates a rapid rate of evolution, in some plant families, these genes are difficult to amplify. For example, matK and rbcL were able to identify species to the Betus and Salix genus level, but did not allow adequate resolution to distinguish among species belonging to these genera and the rate of amplification was low (only 21% of the Salix samples amplified; Jarvinen et al. 2004; Fazekas et al., 2008; von Crautlein et al., 2011).

DNA barcoding can be suitable for two different purposes: (i) the molecular identification of already described species, and (ii) the discovery of undescribed species (Casiraghi et al., 2010). In a typical DNA barcoding strategy, the sequence of a given species is compared against reference sequences in a library database (sequences of previously identified individuals) for a given barcode. This comparison can result in a query sequence match to another sequence in the library, which leads to species identification (Hajibabaei, Singer, Hebert, & Hickey, 2007). A case where there is no match to any record in the database could also indicate the existence of a new species (Hebert et al. 2004). Trinidad sequences were accurately identified to the genus level for all endemic plant species successfully amplified and sequenced using both matK and rbcL markers. Accurate genus–level identification is important for poorly described (or sampled) groups as well as for the enforcement of quarantine and trafficking regulations as regulators more commonly list genera rather than species (Little, 2011). In this study, our endemics did not match any other species with 100% similarity in the reference libraries created for each Trinidad species. Does this mean new species assignments for Trinidad endemics? Casiraghi et al. (2010) cautions against assigning biological meaning to genetic ranks, unless these sequences are able to clearly and unequivocally link a species to the variability pattern of a single DNA barcoding marker.

There is no single optimal method to determine the resolving power of DNA barcodes for all taxa (Austerlitz et al., 2009; Casiraghi et al., 2010; Collins & Cruickshank, 2013; Meyer & Paulay, 2005; Moritz & Cicero, 2004; Ross et al., 2008). Different approaches exist for matching an unknown query sequence with sequences in a reference database or library and tend to be based on ad hoc criteria which may include the frequency of the highest hits, percentage sequence similarity, bootstrapping, BLAST scores or tree‐based clustering assessment (Kress et al., 2009; Wilson et al. 2011). Although there is no consensus on the “best approach” and in reality, the most appropriate approach may be dependent on a number of variables, it is recommended that, as far as possible, the taxonomic origin and assignments be independently confirmed (i.e., using morphological characters) to improve the accuracy of taxonomic assignment through barcoding (Hollingsworth et al., 2016; Wilson et al. 2011).

The main challenge to using distance‐based methods to species identification is that no single genetic distance threshold distinguishes all species (Ferguson, 2002; DeSalle et al., 2005; Little and Stevenson 2007; Wilson et al. 2011). A threshold value calculated from genetic distances may be more appropriate than using a single arbitrary 1% or 3% threshold (Meier et al., 2006; Fazekas et al. 2009; Collins & Cruickshank, 2013). In this study, there was little change in the proportion of “correct,” “ambiguous,” and “incorrect” assignments when threshold values of 1%, 3% and a separate calculated threshold for each reference sequence library dataset were used. Despite using threshold values calculated from K‐2‐P genetic distances for each taxon, all of the endemic species were still classified as “ambiguous” but, they were all assigned to the correct genus for matK and rbcL barcodes. Clusia aripoensis was the only species with a “no match” status based on rbcL sequence comparisons. Two reasons for this result may be explained as: (i) there was poor library sequence database coverage, and (ii) genetic distances were higher than the calculated threshold for this taxon. In this study, DNA barcoding was useful in flagging atypical specimens or in identifying cryptic species for further taxonomic investigation (Hajibabaei et al., 2007).

The low rate of “correct” classification for both methods that provide “ambiguous” and “no match” classifications are important because they reveal several gaps in the approach to analysis including (i) reference sequence library coverage, (ii) low genetic variation among barcode sequences, and (iii) whether markers are targeting regions of the genomes whose genetic distances can vary from species to species (Hollingsworth et al., 2016). Therefore, the need for further research into understanding the cause of the “ambiguous” or “no match” status in identity is highlighted. One approach to ensure good reference library coverage would be to barcode congenerics for each species selected for study sharing the same geography. Even if this were feasible, in terms of availability of specimens, there is no guarantee that these congeneric barcodes would be sufficient to discriminate among all species as was found to be the case with Dendrobium species (Singh, Parveen, Raghuvanshi, & Babbar, 2012).

Tree‐based methods involve assignment of a query sequence to a certain taxon if it is found in a clade consisting of reference sequences with moderate to high bootstrap support. These methods require appropriate alignment of all sequences which may be difficult for highly divergent sequences (Mao et al., 2014; Wilson et al. 2011). While barcode libraries are somewhat similar to molecular phylogenetic data (i.e., they are both built from sequence information from different species), DNA barcodes do not usually have sufficient phylogenetic signal to infer evolutionary relationships (Hajibabaei et al., 2006, 2007). In this study, NJ trees were used to establish clustering of query sequences into correct genus‐specific groups with strong bootstrap support and were not used to infer phylogeny. Poor resolution in tree topologies with low bootstrap scores and polytomies obtained for rbcL sequences were obtained which may be due to inadequate low genetic distances for most species (Hebert et al., 2003; Hollingsworth et al., 2016; Kress et al., 2009; Ross et al., 2008; Wilson et al. 2011). Others have reported low resolution in rbcL because it is known to have insufficient nucleotide sequence variability to distinguish among closely related species (Kress & Erickson, 2007; Newmaster, Grguric, Shanmughanandhan, Ramalingam, & Ragupathy, 2013).

In this study, it was difficult to concatenate relevant sequences mined from GenBank for the matK and rbcL markers for each species. As such, we analyzed separate markers. Hollingsworth et al. (2016) also reported on the difficulty in concatenating sequences available in reference libraries. Others found no improvement in species identification using a combined multilocus approach and loci rarely discriminated among samples that were not already correctly classified using the better performing of the two loci separately (Lahaye et al., 2008). In fact, it seems counterintuitive to combine a high‐performing marker with a low performing marker in an effort to improve the proportion of correct assignments. In this study, the matK marker had a higher percentage of correct identifications compared to the rbcL marker.


DNA barcoding has the potential to distinguish among species that are closely related and among those which are evolutionarily divergent using single barcodes. We have found that barcoding success is dependent on having taxonomically appropriate representation in the reference sequence database, the genetic distance among the sequences in this database, the species under study which affects both technical and species discrimination success, the accuracy of identity of species in the reference sequence database, the barcodes used and whether there is a high level of monophyly among species of a given genus. In other words, the performance of the matK and rbcL approved barcodes appeared to be species‐specific or genus‐specific, which is what has been cautioned by others (Casiraghi et al., 2010). The “best close match” tool implemented in the TaxonDNA suite was useful because of its ability to discriminate among “correct,” “ambiguous,” “incorrect,” and “no match” classifications for each species in the reference sequence database in addition to query sequences. The tree‐based method generally reflected the genetic distances among the sequences in the reference sequence database, and in most cases, our endemic species were positioned in clusters that were genus‐specific based on the matK barcode. This was not the case for the rbcL barcode as the tree topology was poorly resolved due to very low variation among the sequences of the reference sequence database. Others have used different barcodes such as ITS2 in similar ethno‐pharmacology‐based identifications with success (Chen et al., 2010; Gao et al., 2010). DNA barcoding also involves massive sample sets with often industrial‐scale laboratory practices and bioinformatics pipelines (Hollingsworth et al., 2016). These challenges are especially important to developing countries with high levels of biodiversity but with limited resources to conduct DNA barcoding work.


The authors declare that they have no conflict of interest.


This work was funded by The University of the West Indies Campus Research and Publications Grant No. CRP.3.MAR13.11. The authors wish to thank Mr. Stephen Narine and Ms. Keisha Manaure for field assistance.


Hosein FN, Austin N, Maharaj S, et al. Utility of DNA barcoding to identify rare endemic vascular plant species in Trinidad. Ecol Evol. 2017;7:7311–7333.


  • Austerlitz F., David O., Schaeffer B., Bleakley K., Olteanu M., Leblois R., … Laredo C. (2009). DNA barcode analysis: A comparison of phylogenetic and statistical classification methods. BMC Bioinformatics, 10(14), S10. [PMC free article] [PubMed]
  • Baksh‐Comeau Y. S., Maharaj S. S., Adams C. D., Harris S. A., Filer D. L., & Hawthorne W. D. (2016). An annotated checklist of the vascular plants of Trinidad and Tobago with analysis of vegetation types and botanical ‘hotspots’. Phytotaxa, 265, 36–293.
  • Barthelson R. A., Sundareshan P., Galbraith D. W., & Woosley R. L. (2006). Development of a comprehensive detection method for medicinal and toxic plant species. American Journal of Botany, 93, 566–574. [PubMed]
  • Bellakhdar J., Claisse R., Fleurentin J., & Younos C. (1991). Repertory of standard herbal drugs in the Moroccan pharmacopoea. Journal of Ethnopharmacology, 35, 123–143. [PubMed]
  • Bruni I., De Mattia F., Galimberti A., Galasso G., Banfi E., Casiraghi M., et al. (2010). Identification of poisonous plants by DNA barcoding approach. International Journal of Legal Medicine, 124, 595–603. [PubMed]
  • Casiraghi M., Labra M., Ferri E., Galimberti A., & De Mattia F. (2010). DNA barcoding: A six‐question tour to improve users’ awareness about the method. Briefings in Bioinformatics, 11, 440–453. [PubMed]
  • CBOL Plant Working Group (2009). A DNA barcode for land plants. Proceedings of the National Academy of Sciences of the United States of America, 106, 12794–12797. [PubMed]
  • Chase M. W., & Fay M. F. (2009). Barcoding of plants and fungi. Science, 325, 682–683. [PubMed]
  • Chase M. W., Salamin N., Wilkinson M., Dunwell J. M., Kesanakurthi R. P., Haidar N., et al. (2005). Land plants and DNA barcodes: Short‐term and long‐term goals. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 29, 1889–1895. [PMC free article] [PubMed]
  • Chen S., Yao H., Han J., Liu C., Song J., Shi L., et al. (2010). Validation of the ITS2 region as a novel DNA barcode for identifying medicinal plant species. PLoS ONE, 5(1), e8613. [PubMed]
  • Clement Y. N., Baksh‐Comeau Y. S., & Seaforth C. E. (2015). An ethnobotanical survey of medicinal plants in Trinidad. Journal of Ethnobiology and Ethnomedicine, 11, 67 [PubMed]
  • Coissac E., Hollingsworth P. M., Lavergne S., & Taberlet P. (2016). From barcodes to genomes: Extending the concept of DNA barcoding. Molecular Ecology, 25, 1423–1428. [PubMed]
  • Collins R. A., & Cruickshank R. H. (2013). The seven deadly sins of DNA barcoding. Molecular Ecology Resources, 13(6), 969–975. [PubMed]
  • Da Silva G., Serrano R., & Silva O. (2011). Maytenus heterophylla and Maytenus senegalensis, two traditional herbal medicines. Journal of Natural Science, Biology and Medicine, 2, 59–65. [PMC free article] [PubMed]
  • DeSalle R., Egan M. G., & Siddall M. (2005). The unholy trinity: Taxonomy, species delimitation and DNA barcoding. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 360, 1905–1916. [PubMed]
  • Elias M., Hill R. I., Willmott K. R., Dasmahapatra K. K., Brower A. V. Z., et al. (2007). Limited performance of DNA barcoding in diverse community of tropical butterflies. Proceedings of the Royal Society of London B: Biological Sciences, 274, 2881–2889. [PMC free article] [PubMed]
  • Fazekas A. J., Burgess K. S., Kesanakurti P. R., Graham S. W., Newmaster S. G., Husband B. C., Percy D. M., et al. (2008). Multiple multilocus DNA barcodes from the plastid genome discriminate plant species equally well. PLoS ONE, 3, e2802. [PubMed]
  • Fazekas A. J., Kesanakurti P. R., Burgess K. S., Percy D. M., Graham S. W., Barrett S. C., … Husband B. C. (2009). Are plant species inherently harder to discriminate than animal species using DNA barcoding markers? Molecular Ecology Resources, 9(s1), 130–139. [PubMed]
  • Felsenstein J. (1985). Confidence limits on phylogenies: An approach using the bootstrap. Evolution, 39(4), 783–791. [PubMed]
  • Ferguson J. W. H. (2002). On the use of genetic divergence for identifying species. Biological Journal of the Linnean Society, 75, 509–516.
  • Ford C. S., Ayres K. L., Toomey N., et al. (2009). Selection of candidate coding DNA barcoding regions for use on land plants. Botanical Journal of the Linnean Society, 159, 1–11.
  • Francisco‐Ortega J., Santiago‐Valentin E., Acevedo‐Rodriguez P., Lewis C., Pipoly J. III, Meerow A. W., et al. (2007). Seed plant genera endemic to the Caribbean Island biodiversity hotspot: A review and a molecular phylogenetic perspective. Botanical Review, 73, 183–234.
  • Gao T., Yao H., Song J., Liu C., Zhu Y., Ma Y., et al. (2010). Identification of medicinal plants in the family Fabaceae using a potential DNA barcode ITS2. Journal of Ethnopharmacology, 130, 116–121. [PubMed]
  • Gong W., Liu Y., Chen J., Hong Y., & Kong H.‐H. (2016). DNA barcodes identify Chinese medicinal plants and detect geographical patterns of Sinosenecio (Asteraceae). Journal of Systematics and Evolution, 54, 83–91.
  • Hajibabaei M., Singer G. A. C., Hebert P. D. N., & Hickey D. A. (2007). DNA barcoding: How it complements taxonomy, molecular phylogenetics and population genetics. Trends in Genetics, 23, 167–172. [PubMed]
  • Hajibabaei M., et al. (2006). Benchmarking DNA barcodes: An assessment using available primate sequences. Genome, 49, 851–854. [PubMed]
  • Harvey A. L., & Gericke N. (2011). Bioprospecting: Creating a value for biodiversity. Chapter 15 In: Pavlinov I., editor. (ed.), Research in biodiversity – Models and applications (pp. 323–338). InTech, Retrieved from
  • Hebert P. D. N., Cywinska A., Ball S. L., & deWaard J. R. (2003). Biological identifications through DNA barcodes. Proceedings of the Royal Society of London B: Biological Sciences, 270, 313–321. [PMC free article] [PubMed]
  • Hebert P. D. N., Stoeckle M. Y., Zemlak Y. S., & Francis C. M. (2004). Identification of birds through DNA Barcodes. PLoS Biology, [PMC free article] [PubMed]
  • Hollingsworth M. L., Clark A., Forrest L. L., Richardson J., Pennington R. T., Long D. G., et al. (2009). Selecting barcoding loci for plants: Evaluation of seven candidate loci with species‐level sampling in three divergent groups of land plants. Molecular Ecology Resources, 9, 439–457. [PubMed]
  • Hollingsworth P. M., Graham S. W., & Little D. P. (2011). Choosing and using a plant DNA barcode. PLoS ONE, 6, e19254. 668 [PubMed]
  • Hollingsworth P. M., Li D.‐Z., van der Bank M., & Twyford A. D. (2016). Telling plant species apart with DNA: From barcodes to genomes. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 371, 20150338 [PubMed]
  • Jarvinen P., Palmé A., Morales L. O., Lännenpää M., Keinänen M., Sopanen T., & Lascoux M. (2004). Phylogenetic relationships of Betula species (Betulaceae) based on nuclear ADH and chloroplastmatK sequences. American Journal of Botany, 91, 1834–1845. [PubMed]
  • Katoh S. (2013). MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Molecular Biology and Evolution, 30, 772–780. [PubMed]
  • Kelly L. J., Ameka G. K., & Chase M. W. (2010). DNA barcoding of African Podostemaceae (river‐weeds): A test of proposed barcode regions. Taxon, 59, 251–260.
  • Kenny J. S. (2008). The biological diversity of Trinidad and Tobago: A naturalist's notes (pp. 10–12). Maraval, Port of Spain: Prospect Press.
  • Kenny J., Comeau P., & Katwaru L. (1997). A survey of biological diversity, Trinidad and Tobago, Port of Spain: United Nations Development Programme.
  • Kimura M. (1980). A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution, 16, 111–120. [PubMed]
  • Koboyashi N., Horikoshi T., Katsuyama H., Handa T., & Takayanagi K. (1998). A simple and efficient DNA extraction method for plants, especially woody plants. Plant Tissue Culture and Biotechnology, 4, 76–80.
  • Kress W. J., & Erickson D. L. (2007). A two‐locus global DNA barcode for land plants: The coding rbcL gene complements the non‐coding trnH‐psbA spacer region. PLoS ONE, 2(6), e508. [PubMed]
  • Kress J. K., & Erickson D. L. (2008). DNA barcodes: Genes, genomics, and bioinformatics. Proceedings of the National Academy of Sciences of the United States of America, 105, 2761–2762. [PubMed]
  • Kress W. J., Erickson D. L., Jones F. A., Swenson N. G., Perez R., Sanjur O., et al. (2009). Plant DNA barcodes and a community phylogeny of a tropical forest dynamics plot in Panama. Proceedings of the National Academy of Sciences of the United States of America, 106, 18621–18626. [PubMed]
  • Kress W. J., & Horvitz C. (2005). Habitat alteration in the Caribbean: Natural and human‐induced In Krupnick G. A., editor; , & Janzen D. H., editor. (Eds.), Plant conservation: A natural history approach (pp. 147–151). Chicago, IL: The University of Chicago Press.
  • Kress W. J., Wurdack K. J., Zimmer E. A., Weigt L. A., & Janzen D. H. (2005). Use of DNA barcodes to identify flowering plants. Proceedings of the National Academy of Sciences of the United States of America, 102, 8369–8374. [PubMed]
  • Kumar S., Sharma S., & Chattopadhyay S. K. (2013). The potential health benefit of polyisoprenylated benzophenones from Garcinia and related genera: Ethnobotanical and therapeutic importance. Fitoterapia, 89, 86–125. [PubMed]
  • Laguerre M. S. (1987). Afro‐Caribbean folk medicine. South Hadley, MA: Bergin and Garvey Publishers.
  • Lahaye R., van der Bank M., Bogarin D., Warner J., Pupulin F., Gigor G., et al. (2008). DNA barcoding the floras of biodiversity hotspots. Proceedings of the National Academy of Sciences of the United States of America, 105, 2923–2928. [PubMed]
  • Larranaga N., & Hormaza J. I. (2015). DNA barcoding of perennial fruit tree species of agronomic interest in the genus Annona (Annonaceae). Frontiers in Plant Science, 6, 589 [PubMed]
  • Little D. P., & Stevenson D. Wm. (2007). A comparison of algorithms for the identification of specimens using DNA barcodes: Examples from gymnosperms. Cladistics, 23, 1–21.
  • Li X., Yang Y., Henry R. J., Rossetto M., Wang Y., & Chen S. (2015). Plant DNA barcoding: From gene to genome. Biological Reviews, 90, 157–166. [PubMed]
  • Librado P., & Rozas J. (2009). DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics, 25, 1451–1452. [PubMed]
  • Little D. P. (2011). DNA barcode sequence identification incorporating taxonomic hierarchy and within taxon variability. PLoS ONE, 6(8), e20552. [PubMed]
  • MacArthur R. H. (1972). Geographical ecology: Patterns in the distribution of species. Princeton, NJ: Princeton University Press.
  • Mahabir D., & Gulliford M. C. (1997). Use of medicinal plants for diabetes in Trinidad and Tobago. Revista Panamericana de Salud Pública, 1, 174–179. [PubMed]
  • Mao Y.‐R., Zhang Y.‐H., Nakamura K., Guan B.‐C., & Qiu Y.‐X. (2014). Developing DNA barcodes for species identification in Podophylloideae (Berberidaceae). Journal of Systematics and Evolution, 52, 487–499.
  • Maunder M., Leiva A., Santiago‐Valentín E., Stevenson D. W., Acevedo‐Rodríguez P., Meerow A. W., et al. (2008). Plant conservation in the Caribbean island biodiversity hotspot. The Botanical Review, 74, 197–207.
  • Meier R., Kwong S., Vaidya G., & Ng P. K. L. (2006). DNA barcoding and taxonomy in diptera: A tale of high intraspecific variability and low identification success. Systematic Biology, 55, 715–728. [PubMed]
  • Meyer C. P., & Paulay G. (2005). DNA barcoding: Error rates based on comprehensive sampling. PLoS Biology, 3, e422. [PubMed]
  • Mittermeier R. A., editor; , Gil R. R., editor; , Hoffman M., editor; , Pilgrim J., editor; , Brooks T., editor; , Mittermeier C. G., editor; , Lamoreux J., editor; , & da Fonseca G. A. B., editor. (Eds.) (2004). Hotspots revisited: Earth's biologically richest and most threatened terrestrial ecoregions. Mexico DF: CEMEX.
  • Moritz C., & Cicero C. (2004). DNA barcoding: Promises and pitfalls. PLoS Biology, 2, e354. [PubMed]
  • Newmaster S. G., Grguric M., Shanmughanandhan D., Ramalingam S., & Ragupathy S. (2013). DNA barcoding detects contamination and substitution in North American herbal products. BMC Medicine., 11, 222 [PubMed]
  • Palmer J. D., Adams K. L., Cho Y., Parkinson C. L., Qiu Y. L., & Song K. (2000). Dynamic evolution of plant mitochondrial genomes: Mobile genes and introns and highly variable mutation rates. Proceedings of the National Academy of Sciences of the United States of America, 97, 6960–6966. [PubMed]
  • Pettengill J. B., & Neel M. C. (2010). An evaluation of candidate plant DNA barcodes and assignment methods in diagnosing 29 species in the genus Agalinis (Orobanchaceae). American Journal of Botany, 97, 1391–1406. [PubMed]
  • Quinlan M. B., & Quinlan R. J. (2007). Modernization and medicinal plant knowledge in a Caribbean horticultural village. Medical Anthropology Quarterly, 21, 169–192. [PubMed]
  • Rach J., DeSalle R., Sarkar I. N., Schierwater B., & Hadrys H. (2008). Character‐based DNA barcoding allows discrimination of genera, species and populations in Odonata. Proceedings of the Royal Society of London B: Biological Sciences, 275, 237–247. [PMC free article] [PubMed]
  • Ratnasingham S., & Hebert P. D. N. (2007). BOLD: The Barcode of Life Data System ( Molecular Ecology Notes, 7, 355–364. [PubMed]
  • Rieseberg L. H., Wood T. E., & Baack E. (2006). The nature of plant species. Nature, 440, 524–527. [PubMed]
  • Ross H. A., Murugan S., & Li W. L. S. (2008). Testing the reliability of genetic methods of species identification via simulation. Systematic Biology, 57, 216–230. [PubMed]
  • Roy S., Tyagi A., Shukla V., Kumar A., Singh U. M., Chaudhary L. B., et al. (2010). Universal plant DNA barcode loci may not work in complex groups: A case study with Indian berberis species. PLoS ONE, 5(10), e13674. [PubMed]
  • Rozas J. (2009). DNA Sequence polymorphism analysis using DnaSP In Posada D., editor. (Ed.), Bioinformatics for DNA sequence analysis; Methods in molecular biology series Vol. 537 (pp. 337–350). NJ, USA: Humana Press, NJ, USA.
  • Sass C., Little D. P., Stevenson D. W., & Specht C. D. (2007). DNA barcoding in the cycadales: Testing the potential of proposed barcoding markers for species identification of cycads. PLoS ONE, 2, e1154. [PubMed]
  • Saitou N., & Nei M. (1987). The neighbor‐joining method‐A new method for reconstructing phylogenetic trees. Molecular Biology and Evolution, 4, 406–425. [PubMed]
  • Shi H., Singh A., Kant S., Zhu Z., & Waller E. (2005). Integrating habitat status, human population pressure, and protection status into biodiversity conservation priority setting. Conservation Biology, 19, 1273–1285.
  • Singh H. K., Parveen I., Raghuvanshi S., & Babbar S. B. (2012). The loci recommended as universal barcodes for plants on the basis of floristic studies may not work with congeneric species as exemplified by DNA barcoding of Dendrobium species. BMC Research Notes, 7, e42. [PMC free article] [PubMed]
  • Tamura K., Peterson D., Peterson N., Stecher G., Nei M., & Kumar S. (2011). MEGA5: Molecular evolutionary genetic analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution, 28, 2731–2739. [PubMed]
  • Techen N., Parveen I., Pan Z., & Khan I. (2014). A DNA barcoding of medicinal plant material for identification. Current Opinion in Biotechnology, 25, 103–110. [PubMed]
  • Theodoridis S., Stefanaki A., Tezcan M., Aki C., Kokkini S., & Vlachonasios K. E. (2012). DNA barcoding in native plants of the Labiatae (Lamiaceae) family from Chios Island (Greece) and the adjacent Cesme‐Karaburun Peninsula (Turkey). Molecular Ecology Resources, 12, 620–633. [PubMed]
  • Valentini A., Pompanon F., & Taberlet P. (2008). DNA barcoding for ecologist. Trends in Ecology & Evolution, 24, 110–117. [PubMed]
  • van den Eynden V., Oatham M. P., & Johnson W. (2008). How free access internet resources benefit biodiversity and conservation research: Trinidad and Tobago's endemic plants and their conservation status. Oryx, 42, 400–407.
  • van Velzen R., Weitschek E., Felici G., & Bakker F. T. (2012). DNA barcoding of recently diverged species: Relative performance of matching methods. PLoS ONE, 7(1), e30490. [PubMed]
  • von Crautlein M., Korpelainen H., Pietiläinen M., & Rikkinen J. (2011). DNA barcoding: A tool for improved taxon identification and detection of species diversity. Biodiversity and Conservation, 20, 373–389.
  • Waugh J. (2007). DNA barcoding in animal species: Progress, potential and pitfalls. BioEssays, 29, 188–197. [PubMed]
  • Weiner M. A., & Weiner J. A. (1999). Herbs that heal: Prescription for herbal healing. Mill Valley, CA: Quantum Books.
  • Wilson J. J., Rougerie R., Schonfeld J. J., Janzen D. H., Hallwachs W., Hajibabaei M., et al. (2011). When species matches are unavailable are DNA barcodes correctly assigned to higher taxa? An assessment using sphingid moths. BMC Ecology, 11, 18. [PubMed]
  • Wilson J. J., Rougerie R., Schonfeld J., et al. (2011). When species matches are unavailable are DNA barcodes correctly assigned to higher taxa? An assessment using sphingid moths. BMC Ecology, 11, 18. [PubMed]
  • Zanoni T. (1989). Hispaniola In Campbell D. G., editor; , & Hammond H. D., editor. (Eds.), Floristic inventory of tropical countries: The status of plant systematics, collections, and vegetation, plus recommendations for the future (pp. 336–340). New York: The New York Botanical Garden.
  • Zhang C. Y., Wang F. Y., Yan H. F., Hao G., Hu C. M., & Ge X. J. (2012). Testing DNA barcoding in closely related groups of Lysimachia L. (Myrsinaceae). Molecular Ecology Resources, 12, 98–108. [PubMed]

Articles from Ecology and Evolution are provided here courtesy of Wiley-Blackwell