Complete Chloroplast Genome Sequence of an Orchid Model Plant Candidate: Erycina pusilla Apply in Tropical Oncidium Breeding

Oncidium is an important ornamental plant but the study of its functional genomics is difficult. Erycina pusilla is a fast-growing Oncidiinae species. Several characteristics including low chromosome number, small genome size, short growth period, and its ability to complete its life cycle in vitro make E. pusilla a good model candidate and parent for hybridization for orchids. Although genetic information remains limited, systematic molecular analysis of its chloroplast genome might provide useful genetic information. By combining bacterial artificial chromosome (BAC) clones and next-generation sequencing (NGS), the chloroplast (cp) genome of E. pusilla was sequenced accurately, efficiently and economically. The cp genome of E. pusilla shares 89 and 84% similarity with Oncidium Gower Ramsey and Phalanopsis aphrodite, respectively. Comparing these 3 cp genomes, 5 regions have been identified as showing diversity. Using PCR analysis of 19 species belonging to the Epidendroideae subfamily, a conserved deletion was found in the rps15-trnN region of the Cymbidieae tribe. Because commercial Oncidium varieties in Taiwan are limited, identification of potential parents using molecular breeding method has become very important. To demonstrate the relationship between taxonomic position and hybrid compatibility of E. pusilla, 4 DNA regions of 36 tropically adapted Oncidiinae varieties have been analyzed. The results indicated that trnF-ndhJ and trnH-psbA were suitable for phylogenetic analysis. E. pusilla proved to be phylogenetically closer to Rodriguezia and Tolumnia than Oncidium, despite its similar floral appearance to Oncidium. These results indicate the hybrid compatibility of E. pusilla, its cp genome providing important information for Oncidium breeding.


Introduction
The Orchidaceae has great diversity in floral morphology and a rich array of species. It comprises the largest family of flowering plants [1]. The extraordinary variety of orchid floral features and appearances ensure a large consumer demand for orchids. Oncidium, a genus in subtribe Oncidiinae, is a popular and important cut flower. It needs about 3 years to reach sexual maturity under natural conditions [2], its flowering being precisely regulated by temperature [3]. Taiwan, located in tropical and subtropical areas with higher temperature, is one of the Oncidium cultivation and hybridization center of world [4]. To develop new commercial species with competitive advantageous traits such as shorter vegetative stage or tropical growth of Oncidium species continues to be a pressing need.
For orchid breeding, it is important for molecular studies of orchid to help in efforts to create unique flower colors and shapes, as well as disease-resistant cultivars that are of high economic value. In Taiwan, genomics approach by transcriptomic database establishment [5][6][7] chloroplast genome [8,9], and transformation technology have been applied in orchid research [10,11]. Few transgenic orchids have been obtained and used in orchid breeding [12]. Sweet pepper ferredoxin-like protein (pflp), a new selection marker developed by You et al., was over-expressing in Oncidium orchid ''Sherry Baby cultivar OM8'' and enanced Erwinia carotovora resistance of transgenic orchid [13]. Transgenic Phalaenopsis expressing coat protein of Cymbidium mosaic virus (CymMV) enhanced protection against CymMV infection through RNAmediated resistance [14]. However, several disadvantages that make orchid breeding by either traditional hybridization or gene engineering difficult are: (1) most of the plants grow slowly; (2) there is a wide range of chromosome numbers e.g. n = 6-30 in Oncidiinae [15]; (3) genome sizes are large and have complex polyploidy caused by spontaneous or man-made hybridization [16][17][18]. Many Oncidium genes have been cloned and studied by ectopic expression in Arabidopsis or Eustoma [19][20][21], however, gain-of-function studies in orchid remains scarce. Therefore, an orchid model plant system is needed for the functional genetic investigation.
Erycina pussila, is a fast-growing epiphytic orchid with a relatively low chromosome number (n = 6; [22]) and small genome size (1.5 pg per 1C nucleus; [23]). Pollination and production of seed capsules rarely occurs in nature [24]. Currently, advances in cultural techniques and precocious flowering have meant that E. pusilla can be grown rapidly, and will produce flowers and fruit in vitro [25,26]. These characteristics make E. pusilla not only an attractive model plant for functional genomic and flowering studies of Oncidium, but also an excellent parent for traditional hybridization methods. To produce attractive traits and breed new commercial orchid species, E. pusilla has been crossed with several important Oncidiinae orchids, and different hybridization compatibility was found with Oncidium, Rodriguezia and Tolumnia [27]. However, systematic molecular investigations and genomic information on E. pusilla remain unclear.
Chloroplast DNA is useful in evolutionary studies because of its simple structure, highly conserved sequence, and maternal inheritance characters [28]. Several plastid regions, such as matK, atpB, psbB, psbC and rpoC1, have been used to identify phylogenetic relationships in orchid [29,30]. Sequencing of complete plastid genomes of different genera has recently provided useful information regarding RNA editing and loss of introns [31,32]. Chloroplast (cp) genomes of 2 orchids, Phalaenopsis aphrodite and Oncidium Gower Ramsey, have been sequenced [8,9]. The availability of chloroplast genome sequences should also help in developing genetic engineering including chloroplast transformation [33,34]. Information on the complete chloroplast genome sequence is not only important for taxonomic classification but also for crop improvement.
To provide information on breeding and molecular aspects of Oncidium, we have sequenced the complete cp genome of E. pusilla using BAC library and next-generation sequencing (NGS). To demonstrate the possibility this orchid as a model, the difference between its cp genome and those of other important orchid species, such as Phalaenopsis and Oncidium, have been compared. Primers were designed based on the various regions of 3 cp genomes and 19 Epidendroideae species were analyzed by PCR. To investigate the cross compatibility of E. pusilla in Oncidiinae species, several cp DNA regions of 36 Oncidium species were obtained and applied for phylogenetic analysis.

Chloroplast-BAC clone identification
Young E. pusilla leaves (200 g) grown in vitro were collected for isolation of high molecular weight DNA [35]. The DNA was partially digested by HindIII, and the fragments ligated into vector pCC1BAC DH10b (Amplicon, Pullman, WA), which was used for transfection into E. coli. Individual clones were picked up and placed into 384-well plates that contained liquid LB medium with 12.5 mg/L chloramphenicol. The plates were incubated at 37uC overnight and stored at 280uC. Chloroplast specific primers designed by Wu et al. [9] were used to amplify predicted chloroplast regions from a BAC library. The BAC clones containing chloroplast regions of interest were obtained by PCR screening from super pool, plate, row, and spot, as described by Hsu et al. [10]. BAC clones of the chloroplast were identified (clone ID P-5-K16).

Illumina sequncing
BAC plasmids for Illumina sequencing were isolated using the NucleoBond BAC 100 kit (NucleoSpin Blood kit, Macherey-Nagel, Germany). Five micrograms of E. pusilla BAC plasmid were sheared into fragments of 200-600 bp by Bioruptor Next Gen (Diagenode) in 100 ml TE buffer. The purified DNA fragments were treated with T4 DNA polymerase, E. coli DNA polymerase I Klenow fragment and T4 Polynucleotide Kinase. Adapters required for sequencing on the Illumina platform were added to DNA fragments. The ligation products were separated on a 2% agarose gel; those between 270 and 350 bp were excised, eluted from the gel slice, precipitated and resuspended in 15 ml TE, using QIAquick Gel extraction Kit (Qiagen). The adapter-modified DNA fragments were amplified, and the products purified using an Agencourt AMPure XP (Beckman). They were collected in 30 ml of QIAGEN elution buffer (Qiagen). After quantification by Quant-iT dsDNA HS Assay Kit (Invitrogen) and KAPA Library Quantification Kit (KAPABiosystem), the molar concentration was calculated and the quality examined by Expersion DNA 1K Analysis Kit (Bio-Rad). The DNA library was then prepared for sequencing.

Bioinformatics
Sequencing was performed on an Illumina GA IIx platform, using a paired-end strategy at a read-length of 75 bases. Nucleotides with low quality scores (,3) were removed from the sequence reads, and any that had a 100% match to the cloning vector sequence or E. coli sequences were also removed from the subsequent assembly process. De novo assembly was conducted using CLC Genomics Workbench (CLC bio, Cambridge, MA). The gaps between the contigs were filled by PCR.

E. pusilla chloroplast genome annotation
The cp genome was annotated using Dual Organellar GenoMe Annotator (DOGMA) [36]. This program uses a FASTAformatted input file of the complete genomic sequences and identifies putative protein-coding genes by performing BLASTX searches against a custom database of published cp genomes. Both tRNAs and rRNAs were identified by BLASTN searches against the same database of cp genomes. For genes with low sequence identity, manual annotation was performed after identifying the positions of the start and stop codons, as well as the translated amino acid sequence, using the chloroplast/bacterial genetic code. The annotated genome sequences were submitted to NCBI (Accession no: JF_746994).

Plant materials
Orchids were obtained and collected from a local grower in Taiwan. All orchids were maintained in the greenhouse at National Chung Hsing University, Taichung, Taiwan.

DNA purification and genomic PCR
For chloroplast genomic PCR analysis, total genomic DNA was isolated from leaves using a urea extraction buffer system [37]. The primer designs for Epidendroideae species analysis were based on the various regions of 3 orchid cp genomes. The primer sequences, sequence sizes, and forward primer position in E. pusilla are shown in Table 1. Primers designed by Wu et al. [9] were used for Oncidiinae variety analysis (the sequences are shown in Table 1). Genomic PCR was conducted in a final volume of 50 ml containing 2.5 units of Taq DNA polymerase (Violet gene, Taipei, Taiwan), 1.25 mM of each dNTP, and 10 pmol of each primer. The amplification program used was 30 cycles at 94uC for 30 s, 55uC for 30 s, and 72uC for 90 s. The PCR products were sequenced and assembled using VectorNTI Contig Express software.

Phylogenetic analysis of Oncidiinae species
Four cpDNA regions (trnH-psbA, matK, trnF-ndhJ, and IRb-SSC) of 36 Oncidiinae varieties were obtained by PCR (Accession no: JN598910-JN598996), and from the NCBI database (GQ915119-915130, GU132947-132991, GU136251-136287, GU175342-175358). Alignment of nucleotide sequence was performed using the Clustal X program [38] and adjusted by GeneDoc software. Phylogenetic analysis was conducted using MEGA3.1 [39], and the phylogenetic tree generated using the neighbor-joining method with 1,000 bootstrap trials by means of the neighbor-joining algorithm. Percentages of bootstrap values are indicated on the tree.

Comparison of E. pusilla chloroplast genome with genomes of 2 other orchid genera
The cp genome of E. pusilla is 143,164 bp in size and contains a pair of inverted repeats (IRa and IRb) of 23,439 bp separated by large and small single copy (LSC and SSC) regions of 84,189 and 12,097 bp, respectively ( Figure 1). This genome contains 126 different genes that include 73 protein coding genes, 6 pseudogenes and 19 genes duplicated in the IR region. There are 28 distinct tRNAs and 4 distinct rRNA genes. Fifteen genes contain 1 or 2 introns, and 5 of their introns are within tRNAs. The genome consists of 45.02% protein-coding genes, 46.73% non-coding DNA, which includes the intergenic spacer (IGS) regions, regulatory sequences and introns, 1.94% tRNA and 6.31% rRNA genes. The overall GC and AT content of the cp genome is 36.65% and 63.35%, respectively. The AT content of the LSC and SSC regions is 66.15% and 77.54%, respectively, whereas that of the IR region is 65.21%, including the rRNA gene cluster. The gene order of E. pusilla cp genome is very similar to that in the Oncidium cp genome.
To understand the variation in orchid cp genomes, the E. pusilla cp genome was compared with the previously published Onc. Gower Ramsey and P. aphrodite genomes. The cp genome of E. pusilla has 89 and 84% identity with Onc. Gower Ramsy and P. aphrodite, respectively. Most insertions except NADH dehydrogenase (ndh) genes, deletions or diverse sequence regions occur within intergenic regions ( Figure 2). For example, a 198 bp insertion was found in the trnK intron of P. aphrodite. The rpoC1 introns in both Onc. Gower Ramsey and P. aphrodite were 190 bp longer than in E. pusilla. A 153 bp insertion was found in ycf2 of P. aphrodite, making it 51 amino acids longer than in E. pusilla and Onc. Gower Ramsey. Based on the various regions of 3 orchid cp genomes, the primer for further analysis was designed (Table 1).
Among 3 orchid chloroplast genomes, sequences of ndh genes were most variable. In E. pusilla and O. Gower Ramsey, ndhJ was trunked and ndhK was abscent ( Figure 3A). All ndhF were abscent and ndhD were trunked in three orchid cp genomes ( Figure 3B and Figure 3C). The ndhE was only absent in E. pusilla ( Figure 3D). The ndhA gene sequence was only present in O. Gower Ramsey ( Figure 3E). The ndhB that located in the IR region was 892 bp in E. pusilla, which is 1333 and 1137 bp shorter than in Onc. Gower Ramsey and P. aphrodite, respectively ( Figure 3F). These results indicate that deletion and truncation are common in chloroplastencoded ndh genes of orchid plants.

Analysis of 5 regions in 19 Epidendroideae species
To analyze the genome variation in orchids, 5 regions were chosen for PCR analysis in 19 species of the Epidendroideae subfamily (Table 2). In the atpH-atpI region, 4 different deletions were found in E. pusilla, Onc. Gower Ramsey, Acampe rigida and Aerangis hyaloids, which varied in size and location ( Figure 4A). The first 400 bp was highly diverse in the petN-psbM region. Three Calanthe species shared the same deletion. A similar but 20 bp longer deletion was found at the same position in Geodorum densiflorum and Phaius mishmensis. Geo. densiflorum contained one more insertion (Table 2 and Figure 4B). Three deletions at different locations of the petN-psbM region were found in Onc. Gower Ramsey, Cymbidium aloifolium and Aer. Hyaloids (Table 2, and Figure 4B). In the accD-psaI region, Onc. Gower Ramsey and Geo. densiflorum shared the same 536 bp deletion. Aer. Hyaloids, which belongs to the Angraecinae subtribe of Vandeae, contained 2 deletions ( Table 2, and Figure 4C). E. pusilla, Bletilla formosana and Primer sequences, annealing position of the forward primer in E. pusilla, and the PCR amplification length are presented. The first 5 sets of primer were used for Epidendroideae analysis. The last 4 sets of primer were used for Oncidiinae phylogenetic analysis [9]. doi:10.1371/journal.pone.0034738.t001 Dendrobium equitans each contained one unique deletion ( Table 2, and Figure 4C). Analysis of the psbE-petL intergene sequences showed 5 different deletions located in a disorderly fashion within E. pusilla, Onc. Gower Ramsey, Cym. aloifolium and Eria corneri ( Table 2, and Figure 4D). In the rps15-trnN region, a deletion found in all 4 Cymbidieae species, including E. pusilla, Onc. Gower Ramsey, Cym. aloifolium and Geo. densiflorum (Table 2, and Figure 4E). With the exception of Aca. rigida, which contained another 295 bp deletion, the other species all shared similar rps15-trnN sequences ( Table 2, and Figure 4E).

Phylogenetic analysis of 36 Oncidiinae species
To investigate the relationship between the molecular study and orchid breeding, 36 important Oncidium species were suitable for phylogenetic analysis (Table 3). Based on the variation in the 3 orchid cp genomes ( Figure 2) and previous studies [9], 4 primers ( Table 1) were chosen for PCR amplification and phylogenetic analysis. Using the matK gene, which is in a highly conserved region, Miltassia (Mtssa) was grouped with Beallara (Bllra), and Zelemnia (Zlm) was grouped with Tolumnia (Tol). Other species could be grouped by genus, but species could not be separated. Erycina was phylogenetically close to Tolumnia and Zelenkocidium, but distant from Oncidium and Odontocidium (data no shown). The primers for amplifying the variable regions, IRb-SSC, could only partially divide the Oncidium genus from the others, which was not suitable for this analysis.
With the exception of members of the Miltassia and Beallara genera, other species could be separated well using trnF-ndhJ region, although the bootstrap scores were low ( Figure 5A). We therefore combined 2 regions, trnF-ndhJ and trnH-psbA, for phylogenetic analysis ( Figure 5B). The combined analysis gave a similar, but more distinguishable, result when compared with that using the trnF-ndhJ region alone. The Miltassia genus was separated from Beallara. Onc. ornithorhynchum and other Oncidium species were separated into 2 different groups. Phylogenetic analysis showed that E. pusilla, Rodriguezia, and Tolumnia were grouped together, distinct from the Oncidium, Odontocidium and Beallara group.

Advantages of next generation sequencing and BAC libraries for chloroplast genome sequencing
In Taiwan, analysis of orchid genomic sequence have provided valuable information for investigating molecular mechanisms of orchid flowering development, perspectives, and disease resistance pathway [4,[19][20][21]. In the other hand, total cp genomes are useful for evolutionary studies [28]. Total DNA or chloroplast DNA had been used on several occasions as basic materials to obtain cp genomes [8,[40][41][42], but in such studies the possibility of DNA contamination could not be entirely ruled out [43,44]. Other studies show that DNA fragments can be transferred between chloroplasts, mitochondria, and nuclear genomes during evolution [45][46][47]. Sequences of mitochondria and chloroplasts of rice and maize share high percentage sequence homology [48,49]. There are 68 kB cpDNA sequences (42.4% of the cp genome) in the mtDNA of V. vinifera [50]. To reduce the possibility of DNA contamination in our study, we applied BAC library screening by using chloroplast genes as probes for sequencing the complete chloroplast genomes.
For cp genome sequencing, a shotgun library [31] and a PCRbased method were used [9,51]. The PCR-based method relies on the sequence conservation of the chloroplast genome. The products were further validated with Sanger sequencing. Recently, NGS has become a powerful tool for genome sequencing as it is time-saving, low in cost and uses high-throughput technology [40,52]. Various chloroplast genomes, such as 6 woody bamboos belonging to the BEP clade with controversial internal relationships [52] have been successfully sequenced by NGS. However, using NGS sequencing and a BAC library has not previously been used to sequence a complete orchid cp genome.
E. pusilla cp genome BAC clones were identified in our study by using PCR screening. By using chloroplast specific primers [9], the BAC library could be screened using PCR [10], which is easier and faster than traditional hybridization methods [53]. Meanwhile, we also identified BAC plasmids with mitochondria clones which contained chloroplast homologous sequences. Several single nucleotide polymorphisms (SNP), insertion/deletion and homology sequences were found between the mitochondria and chloroplast sequences in E. pusilla (data not shown). Using Illumina sequencing and chloroplast specific BAC plasmids, the possibility of reassembly errors caused by homologous sequences between chloroplast and mitochondria or nucleus DNA could be excluded. Combining a BAC library screened using a PCR approach and Illumina sequencing, we obtained an accurate chloroplast genome sequence efficiently and economically. The chloroplast genome studies will support the identificationy of phylogenitically close relatives of Erycina and will help in their breeding and genetic improvement. Characteristics of E. pusilla chloroplast genome The gene order of the 3 orchid chloroplast genomes was very similar. Unlike other monocot plants, such as maize, rice, and wheat (NC_001320, NC005973, NC_002762, NC_001666), the 3 orchid cp genomes contained the ycf2 gene, which is similar to dicot plants, such as tobacco, Arabidopsis, and Lotus (NC_001879, NC_000932, NC_002694) [8]. The 153 bp longer ycf2 in P. aphrodite (Figure 2) is also found in other dicot plants. The function of the short 153 bp ycf2 in E. pusilla and Onc. Gower Ramsey needs further exploration.
The products of the ndh genes catalyze the transfer of electrons from NADH to plastoquinine, which adjusts the redox level of the photosynthetic electron transporters [54]. Most ndh genes in the chloroplast genome of E. pusilla, Onc. Gower Ramsey, P. aphrodite and the other 14 Oncidiinae species were deleted or truncated [8,9]. The non-functional ndh genes are also found in crassulacean acid metabolism (CAM) and C3 plants, such as Pinus thunbergii, Ketekeeria davidiana, Ephedra equisetina and Welwitshia mirabilis, which belong to autotrophic, heterotrophic, genmospermae, or monocot species, respectively [55][56][57]. Furthermore, in Erodium genus, 11 plastid-encoded ndh genes were intact in Ero. texamum and Ero. carvifolium, but were deleted in Ero. chrysnthum. No morphologic or biological features are associated with ndh gene loss in Erodium [58]. These results indicate that the loss-of-function of chloroplast encodeing the ndh genes might not affect photosynthesis. The ancestral plastid ndh genes of orchids are presumed to have been transferred to the nucleus [8]. The orchid nuclear genome sequences, which are still unavailable, are needed to clarify the horizontal gene transfer questions of ndh genes in orchids.

Analysis of 5 regions in 19 Epidendroideae species
In the petN-psbM region, 3 species belonging to the Collabiinae shared a deletion of the same size ( Table 2, and at the same position, Figure 4B). However, a 20 bp longer deletion was found in Geo. densiflorum and Pha. mishmensis, which belong to the Eulophiinae and Collabiinae subtribes, respectively. Therefore, the petN-psbM deletion found in the three Collabiinae species might not specific to the Collabiinae subtribe.
Within each detected region, different species that share the same sequences are labeled with white circles; unique deletions are labeled as black circles. The same deletions found in different species are labeled as triangles of the same color. Black stars indicate insertions. '-' indicates that no PCR product was obtained. doi:10.1371/journal.pone.0034738.t002 Four species belongs to Cymbidieae tribe including E. pusilla, Onc. Gower Ramsey, Cym. aloifolium and Geo. Densiflorum shared the same deletion in the rps15-trnN region that is located between the SSC and the IR (Table 3, Figure 3E). .SSC/IR is one of the most variable loci and could be an evolution marker [59]. To examine whether this deletion is commonly exits in Cymbidieae tribe, SSC to IR sequences of Epidendroideae were download from the NCBI database and further analyzed [60,61]. The deletion in rps15-trnN was conserved in all 77 Cymbidieae tribe, including species belonging to Oncidiinae, Crytropodiinae, Eulophiinae, and Maxillarieae subtribes ( Figure S1). Similar to P. aphrodite, other species belonging to the Vandeae tribe, Podochileae tribe, Collabiinae subtribe, or Dendrobieae subtribe contained no such deletion. Together, these findings indicate the deletions in the rps15-trnN were commonly exits in Cymbidieae tribe.

Phylogenetic analysis of Oncidiinae species
Due to the unique morphology, E. pusilla has rendered taxonomic classification and denomination a continual challenge [62]. In 2001, E. pusilla was finally named according to the molecular systematics which were performed based on the ITS, matK, and trnL-F DNA regions [63]. The fast-growing feature of E. pusilla makes it a good parent for breeding. To supply molecular information for Oncidium breeding, 36 Oncidiinae species including tropical-adaption commercial hybrids were further analysis. Though the matK region was a good marker for Oncidiinae phylogeny investigation [63]. In our study, the matK regions within 36 Oncidiinae commercial species were too conserved to separate well. The matK regions might not a good marker to make phylogenetic inferences within commercial hybrids. Another primer for the IRb-SSC region failed to produce a PCR product for many species, which also made phylogenetic analysis using a 4 DNA region combination difficult in this study. However, the phylogenetic analysis using the two regions trnF-ndhJ and trnH-psbA were able to demonstrate a good resolution within 36 Oncidiinae including commercial hybrids. E. pusilla is located close to the Tolumnia and Rodriguezia, while Ionocidium, Oncidesa, and Oncidium belong to another group (Figure 4). The phylogenetic result is able to explicate the hybridization compatibility of E. pusilla [27]. No fruit set and seed production in Ionocidium Popcorn 'Haruri' and Oncidesa Little Dragon crossed with E. pusilla. Fruit can be obtained by crossing E. pusilla with Rod. lanceolata but only few seed would be germinated. However, fruits and progeny could be germinated successfully by crossing E. pusilla with several species belonging to the Tolumnia genus including Tol. Genting Angel. Our phylogenetic analysis using trnF-ndhJ and trnH-psbA thus provides a reference for the hybridization compatibility of E. pusilla. For traditional hybridization breeding, this is important information to select new hybrid parents systematically and create new commercial species efficiently.
Previously, orchid taxonomy has been based on floral traits and morphological features. However, classification is changed frequently because the characteristics of orchid are easily affected by interspecific or intergeneric crossing and changes in environment. E. pusilla and Zelenkocidium Little Angel used to be taxonomically grouped into the Oncidium because of their similar floral appearances. Currently, molecular taxonomy has started to reveal more precise phylogenetic relationship and many species of orchid have been renamed and reclassified. For example, Onc. Little Angel was reclassified as Zelenkocidium Little Angel. Onc. Midas, the hybrid of Zelenkoa onusta and Oncidium flexuosum, was renamed as Zelemnia Midas. According our results, Zelemnia and Zelenkocidium are located close to Tolumnia and distant from Oncidium in the phylogenetic tree in the current study ( Figure 4) thus validating previous results. Beyond the species that we have examined, there might be other species that should be moved to Zelemnia or Zelenkocidium from Oncidium despite having a similar floral appearance to Oncidium. The phylogenetic tree showed Odontoglossum Violetta was phylogenetically distant from Odontoglossum and Odontocidium, and Onc. orithorhynchum was much closer to Erycina and Tolumnia than Oncidium and Odontoglossum. We therefore suggest that the taxonomy of Odm. Violetta and Onc. orithorhynchum should be further checked and compared with the parent. Similar misplacements might occur in many commercial Oncidium species. a possibility that requires further investigation. Accurate Orchid taxonomy is not only desirable for evolutionary studies, but is important for orchid breeding.