Introduction

The genus Bidens from Asteraceae family comprises approximately 230 species, which distributed from the tropics to subtropics in Asia, America and other continents [1]. Bidens plants have attracted researchers’ attention from ecologist, pharmacologist and biologist. For example, Sun et al. [2] found Bidens pilosa could tolerate high-cadmium (> 8 mg/kg) soil environment and accumulated to more than 100 mg/kg concentration of cadmium (Cd) in stems, leaves and shoots at the flowering and mature stages, furthermore, Bidens species suggesting it is a Cd-hyperaccumulator which is promising plant in Cd-polluted soil recovery. Phytochemical investigation of Bidens pilosa indicating compounds like polyacetylenes, chalcone glucoside, flavonoid, triterpenoids et al. might be bioactive components which contribute to therapeutic effects in anti-inflammatory, analgesic, antibacterial reported in Bidens plants [3,4,5,6,7,8]. Because of displaying highly ecological and morphological diversity, occupying a wide of habitats including sand dunes, deserts, lava flows, mesic forests, rainforests, scrublands, and wetland bogs, Bidens species especially for Hawaiian Bidens radiation has attracted biologist’ attention for a long time and this genus has been one of the best examples of adaptive radiation in the Hawaiia flora [9]. Based on the study reported by Knope et al. [10], Hawaiian Bidens radiation occurred just within the last 2 million years. Relative short-time divergency and great niche difference among Bidens species caused low levels of genetic variation and high levels of morphological variation, resulted hindered usage of classic DNA markers and morphological markers in resolving the identification and their evolutionary histories. To the best of our knowledge, the classification of Bidens is still controversial, more constructive studies remain to be started to fulfill the gap between current knowledge about evolutionary history and the real situation.

The reason for classical DNA markers such as matK, psbA-trnH, ITS1 and ITS2 are hard to separate closely-related species is that they are usually several hundred of base pairs in length, therefore limiting genetic variation information which is necessary for reaching certain resolution. Based on the records on Flora of China, there are 10 species Bidens species distributed in China [11], what is interesting is that similar morphological and chemical characteristics were found which hindered the studies about both discrimination and phylogenetic history for those species [12, 13]. The first report about molecular identification for Bidens species was conducted by Tsai et al. [12], who applied the noncoding regions from chloroplast genome (trnL intron and trnL-trnF) and nuclear ribosomal DNA (ITS1, 5.8S and ITS2) to Bidens species identification, the results suggested ITS1, 5.8S, ITS2 and trnL intron could only separate B. biternata from B. pilosa var. pilosa. In our preliminary experiments, we retrieved all available ITS1 and ITS2 sequences from NCBI nucleotide database and found that neither ITS1 nor ITS2 could generate reasonable resolution for Bidens species separation (Figs. S1 and S2).

The chloroplast (CP) is an important organelle relating to photosynthesis, and has an independent circular genome with typical quadripartite structure. In angiosperms, CP genomes are highly conserved in gene composition and usually uniparental inheritance, and have low nucleotide substitution rate [14]. Generally, CP genomes are > 100 kb in length, which make CP genomes have significant higher capacity for storing genetic variation than that in widely-used DNA barcodes such as matK, trnH-psbA, ITS1 and ITS2, with only several hundred base pairs in length. In addition, CP genomes are high in copy number in plant cells, for example, about 1000 copies in Arabidopsis thaliana leaf cell was reported [15]. In the past decade, a large number of studies have reported that CP genome is valid tool both in closely-related species identification and in phylogenic-history tracing. For example, Gong et al. [16] unraveled the phylogenetic relationships and hybridization history for Amomum species and developed novel DNA markers for species identification. Song et al. [17] re-constructed Styrax species phylogeny and developed specific DNA barcodes which exhibit higher discriminatory power. Cao et al. [18] reported unique variable sites in CP genes, ndhF, rpl22 and ycf1, which could be applied to distinguish Viola philippica from all other Viola samples including the most closely related species. Based on the previous reports about CP genome and its successful application in species identification, we hypothesized that it might be an effective method to generate reasonable resolution for Bidens pilosa Linn., Bidens bipinnata Linn. and Bidens alba var. radiata and to provide more sound information about phylogeny history of Bidens species.

In this study, we reported the CP genomes of Bidens pilosa Linn., Bidens bipinnata Linn. and Bidens alba var. radiata. Then we conducted CP genome-wide comparative analysis among those species with other published available species in Bidens. Our main goals were: (1) to explore the molecular characteristics of those three chloroplast genomes; (2) making comparison CP genome-widely to detect structural variation among Bidens plants. (3) to analyze sequence variations, get highly divergent regions and develop specific DNA markers for Bidens species identification. (4) unraveling the evolutionary history and explore the phylogenetic relationships of Bidens. This study will provide but not limited to useful information to in clarifying species identification and phylogenetic history about Bidens plants.

Materials and methods

Plant materials, DNA extraction and sequencing

Fresh and healthy samples for Bidens pilosa Linn., Bidens bipinnata Linn. and Bidens alba var. radiata were collected at Xiaoguwei County, Guangzhou city, Guangzhou province, China. The leaves were stored in the liquid nitrogen immediately after being removed from the plants and transferred to -80 ℃ refrigerator (Eppendorf, Hamburg, Germany) right back to the laboratory. All samples were identified by Ye YuShi, the engineer in South China Botanical Garden, Chinese Academy of Sciences, and all voucher specimens were deposited in the Second Clinical College of Guangzhou University of Chinese Medicine (voucher ID numbers: BD190529S1 for Bidens pilosa, BD190529S2 for Bidens bipinnata and BD190529S3 for Bidens alba).

Genomic DNA was extracted using a DNA easy Plant Mini Kit (Qiagen Co., Hilden, Germany) following the manufacturers' instructions. The quality and quantity of DNA was determined with 1% gel electrophoresis and Nanodrop2000C (ThermoScientific, Delaware, USA). The DNA was fragmented into 400 ~ 600 bp by Covaris sonication (Covaris M220, Woburn, MA, USA) and thereafter applied to sequencing library construction with the NEBNext® Ultra™ DNA Library Prep Kit Illumina (New England, Biolabs, Ipswich, MA, USA). Libraries were sequenced by Illumina HiSeq4000 platform (Illumina Inc. CA, USA).

Chloroplast genome assembly and annotation

We handled our sequencing reads according to the method published by Zhou et al. [19], the quality of raw reads was evaluated by FastQC (v0.11.9) software [20], low-quality reads were filtered and trimmed by Trimmomatic (v0.39) [21]. Chloroplast-like (CP-like) reads were extracted by blast the clean reads against the collection of CP genomes retrieved from NCBI nucleotide database. After which, CP-like reads were assembled to continuous contigs by SOAPdenovo2 [22] and scaffolded by SSPACE [23], finally, the remaining gaps were filled by Gap Filler package [24].

We predicted and annotated the CP-genomes in two-round steps. Firstly, the assemblies were predicted and annotated by CPGAVAS2 [25] with default parameters, except for taking 2544-genome-model rather than 43-genome-model as reference dataset. Secondly, gene model generated from above were manually curated by Apollo software [26] with supporting information generated by blast searches against Swiss-Prot database. Finally, the original CPGAVAS2’s prediction was updated by the latest GFF3 file. The chloroplast genome were visualized with OGDRAW v1.3.1 [27].

In addition, tRNAs were predicted by tRNAScan-SE software [28]. Relative synonymous codon usage (RSCU) for protein coding genes was calculated by CodonW (http://codonw.sourceforge.net) [29]. Simple sequence repeats (SSRs) analysis was conducted with MISA v2.1 [30], as to long repeats, an online tool REPuter was used for identifying forward (F), palindromic (P), reverse (R), and complement (C) repeats with default parameters[31]. GC content was calculated with GC function planted in sequin package [32].

Analysis for boundary regions of CP-genomes

CP genomes of B. pilosa, B. bipinnata and B. alba and other three species, B. asymmetric, B. campylotheca, and B. cervicate, retrieved from NCBI were thoroughly compared and analyzed at genome-level. In details, contraction and expansion of IR regions among LSC, IRb, SSC and IRa were visualized by IRScope [33]. MAFFT [34] was applied to multiple sequence alignment (MSA) for CP genomes. Nucleotide diversity (Pi) was calculated with DnaSP [35] taking the MSA file as input, the step size and sliding window length were set to 200 bp and 800 bp respectively. In addition, in order to find available molecular markers for species identification, mVISTA [36] was used for CP comparative analysis and visualization, by which highly divergent regions were detected. Candidate molecular markers were verified by in silico PCR with Fast-PCR.

Phylogenetic analyses and selective pressure evaluation

A phylogenetic tree was reconstructed with the ML-based method with RAxML-ng [37] by providing multiple sequence alignment files generated from whole chloroplast genomes with MAFFT. In detail, there were 15 species were used in the phylogenetic analysis, and H. annuus was regarded as an outgroup. For RAxML-ng, parameters for the model were set to “GTR + R4 + FO”, bs-metric was set to “fbp, tbe”, and 100 starting trees (50 random and 50 parsimony-based) were used to pick the best-scoring topology by setting–tree to “pars, rand” [38]. To evaluate the selection pressure on cp protein-coding genes, we extracted the shared non-redundant genes among species, in which each gene’s CDS-pair of one-by-one species’ combination were extracted and aligned by MAFFT. The rates of synonymous substitutions (Ks) and non-synonymous substitutions (Ka) and Ka/Ks were then calculated by ParaAT2.0 [39], in which KaKs_Calculator [40] is inplanted. The command we applied in this study is “ParaAT.pl –c 11 –h homologs.txt –n CDS –a PEP –p proc –o OUT –k –f axt –m mafft –v”.

Results

Genome characteristics

CP genomes for these Bidens species are in typical quadripartite structures (Fig. 1, Figs. S3 and S4), consisting of LSC, two IRs and a SSC region, where LSC ranged from 151,599 to 154,478 bp, IRs from 24,245 to 26,264 bp, LSC from 83,856 to 84,240 bp and SSC from 17,780 to 18,439 bp, respectively. The sizes of CP genome for those species varied little and are 154,478, 151,611 and 151,599 bp for B. pilosa, B. bipinnata and B. alba (Table 1). GC content are unevenly distributed, i.e., GC contents for LSC, SSC, and IR and CP genome for B. pilosa are 35.64%, 31.34%, 42.42% and 37.52%, B. bipinnata are 35.59%, 30.97%, 43.21% and 37.48%, and B. alba are 35.56%, 30.97%, 43.14% and 37.48%. A total of 129, 129 and 130 genes are found in B. pilosa, B. bipinnata and B. alba, respectively, wherein 86, 85 and 86 are protein-coding genes, 35, 36 and 36 are tRNA genes, in addition the remaining 8 genes for those species are rRNA genes. There are 16, 14, and 15 intron-containing genes in B. pilosa, B. bipinnata, and B. alba, respectively, in which there are 6 intron-containing tRNA genes (trnA-UGC, trnC-ACA, trnE-UUC, trnK-UUU, trnL-UAA, and trnS-CGA) and 12 intron-containing genes (Table S1).

Fig. 1
figure 1

Chloroplast genome map of Bidens pilosa

Full size image
Table 1 Characteristics of the CP genomes
Full size table

As to the protein-coding genes, 5 genes responding for photosystem I (psaA, B, C, I, J), 15 genes for photosystem II (psbA, B, C, D, E, F, H, I, J, K, M, N, T, Z, ycf3), 6 genes for ATP synthase (atpA, B, E, F, H, I), 11 genes for large ribosomal proteins (rpl2, 14, 16, 20, 22, 23, 32, 33, 36), and 12 genes for small ribosomal proteins (rps2, 3, 4, 7, 8, 11, 12, 14, 15, 16, 18, 19) were found for B. pilosa, B. bipinnata and B. alba, Wherein rpl2, 23, rps12 are duplicated genes, clpP is absent from B. bipinnata but kept in B. pilosa and B, alba and psbD only kept in B. pilosa but absent from B. alba and B. bipinnata (Table 2).

Table 2 Common genes annotated in the cp genomes
Full size table

Relative synonymous codon usage

Relative Synonymous Codon Usage (RSCU) for the protein-coding genes of CP genomes for B. pilosa, B. bipinnata and B. alba was calculated by CodonW on the basis of protein-coding genes (Table S2). There are 64 types of codons encoding 20 amino acids were found in this study, in which 6 codons encode Serine (Ser), Leucine (Leu), and Arginine (Arg), 4 codons encode Alanine (Ala), Glycine (Gly), Proline (Pro), Threonine (Thr), 3 codons encode Isoleucine (Ile), 1 codon encode Methionine (Met) and Tryptophan (Trp), respectively, the remaining amino acids were encoded by 2 codons. In addition, we found Leucine was the amino acid in the highest frequency ranging from 10.6% to 10.7% while Cysteine (Cys) was in the lowest ranging from 1.11% to 1.13% in CP genome protein-coding genes. Generally, codon with RSCU > 1 when encoding an amino acid should be regarded as codon usage bias preference [29, 41]. As there is only one codon encoding Methionine and Tryptophan respectively, RSCU for codon of these amino acid was 1. About half of the codons have RSCU > 1 (30/64), 29/30 of which end with base A or T. Moreover, Leu, Arg and Ser are amino acids with the highest RSCU value both for the 3 Bidens plants, exhibiting usage bias for tta, aga, tct, gct with RSCU value of 1.88, 1.87, 1.78 for B. pilosa, 1.88, 1.84, 1.82 for B. bipinnata, 1.87, 1.85, 1.81 for B. alba respectively (Table S2).

IR contraction and expansion

The length of IR from Bidens species we have sequenced varied in length, is 24,652, 24,245, 26,264 bp for B. alba, B. bipinnata and B. pilosa, respectively. It is more divergent than that of published species (ranging 24,661 to 24,662 bp), indicating IR expansion might occurred in B. pilosa while contraction might occurred in B. bipinnata. In addition, contraction and expansion of IR boundary contributes to genome size variations for CP genomes [13, 16, 42]. 3 CP genomes assembled in this study and 3 other Bidens’ published CP genomes (B. cervicata, B. asymmetrica and B. campylotheca) were included in comparison of IRs regions at four junctions between two IRs (IRa and IRb) and the two single-copy region (LSC and SSC) (Fig. 2). As a result, rps19 genes located at the junctions of IRa/LSC and LSC/IRb, wherein, 166 to 178 bp located at LSC region and 101 to 113 bp located at IRb region. As for rpl2, the whole gene located at IR regions in all Bidens species except for B. bipinnata where 282 to 343 bp located at LSC region. As for ndhF in Bidens plants, located at SSC region, variation of location in SSC was found that ndhF located close to IRb region in the three newly assembled CP genomes (specifically, it crossed IRb and SSC in B. pilosa) while close to IRa region in other 3 species. In addition, it is noteworthy to mention ycf1, which spanned SSC/IRa junction in B. pilosa, B. bipinnata and B. alba while only located at IRa in B. cervicata, B. campylotheca, and B. asymmetrica. Those results suggesting genome structure varied in Bidens species.

Fig. 2
figure 2

Comparison of the boundaries of the LSC, SSC and IR regions. JLB: junction between LSC and IRb; JSB: junction between SSC and IRb; JSA: junction between. SSC and IRa; JLA: junction between LSC and IRa

Full size image

Comparative chloroplast genomic analysis

Genome divergence and sequence identity were calculated using the mVISTA by taking B. pilosa as reference (Fig. 3). The results indicate that the LSC and SSC regions are more divergent than IR regions, this finding is in accordance to previous studies [13, 16, 42]. The results revealed that most of the variation was located in the LSC and SSC regions, and slight variation occurred in the IR regions, whereas the coding regions were more conserved than the non-coding regions. The CP genomes of the three Bidens species in this study more conservative than those of the other three species due to the detection of blank regions. We also found slight differences for rps12 and ycf2, which exist in multiple fragments among six species. Moreover, the highly divergent non-coding regions among the six CP genomes appear in the intergenic spacer regions (IGS), including ndhE-psaC, ndhE-ndhG, trnL_UAG-rpl32, ndhF-rpl32, ndhI-ndhG, and ndhD-ccsA. Among them, the most divergent coding regions are the rpl20, ndhA and matK genes.

Fig. 3
figure 3

Comparison of five chloroplast genomes using mVISTA by taking B. pilosa sequence as a reference. The lower left corner is the color coding of gene function, grey arrows indicate the orientation of genes, red bars represent conserved non-coding sequences, purple bars represent exons, and blue bars represent introns. The y-axis represents the percentage identity (shown: 50–100%)

Full size image

Simple sequence repeat and long repeats analysis

Simple sequence repeats (SSRs) are tandem repeat DNA sequences 1 ~ 6 bp in length that have been widely applied as molecular markers in species authentication [43,44,45,46]. A total of 61, 65, 66 SSRs are detected in B. pilosa, B. alba and B. bipinnata, among which 39, 42 and 42 are mononucleotides, 6, 8, and 8 are dinucleotides, 4, 5, 5 are trinucleotides, 9, 8, 9 are tetranucleotides, 2, 2, 2 are pentanucleotides, interestingly, only 1 hexanucleotide repeat is found in B. pilosa (Fig. 4A-B and Table S3A-B). Noteworthy, mononucleotide repeats (A/T) is the highest in number ranging from 35 (B. pilosa) to 41 (B. bipinnata and B. alba) (Fig. 4B). Moreover, long repeats were characterized by REPuter, 99 repeats were detected from those Bidens plants separately. Wherein there are 45, 36, 37 forward repeat, 11, 19, 20 reverse repeat, 41, 32, 33 palindromic repeat and 2, 12, 9 complement repeat respectively. As for 3 other published species, 99 long repeats were identified for B. asymmetric, B. campylotheca, and B. cervicata in which there are 40, 38, 47 forward repeat, 31, 34, 22 reverse repeat, 22, 22, 27 palindromic repeat and 6, 5, 3 complement repeat respectively (Fig. 5). The above results suggested that both SSR and long repeats varied among species, by which novel molecular markers could be developed in Bidens plants identification.

Fig. 4
figure 4

Analysis of simple sequence repeats (SSRs) in the Bidens chloroplast genomes. A. different SSR types detected in three genomes. B. frequency of identified SSR motifs in different repeat class types

Full size image
Fig. 5
figure 5

Analysis of repeated sequences in Bidens chloroplast genomes. C, F, P and R indicate the repeat types: complement repeat, forward repeat, reverse repeat and palindromic repeat

Full size image

Nucleotide diversity

By analyzing the sliding window of Bidens CP genomes using DnaSP software, a high level of variability was observed (Fig. S5A). The nucleotide diversity (Pi) averaged 0.14923 for the 67 protein-coding genes and 46 IGS regions among six Bidens species, while IR region averaged 0.00758, LSC region averaged 0.05286 and SSC averaged 0.10930 respectively, indicating IR region is more conserved than LSC and SSC region. Taking Pi ≥ 0.05 as threshold [42], 34 divergent regions were found for IGS, such as ndhE-psaC (0.38986), ndhE-ndhG (0.35152), trnL_UAG-rpl32 (0.34343), ndhF-rpl32 (0.33850), ndhI-ndhG (0.32686), and ndhD-ccsA (0.31234) (high Pi values > 0.3) (Fig. S5B), these findings were in accordance with the results of mVISTA. These divergent regions are candidate markers might be used as DNA barcods for further phylogenetic analyses and species identification [18, 42]. As to chloroplast genes, Pi of rpl20, ndhA and matK are 0.02101, 0.01011 and 0.00911 (Fig. S5B) respectively, both higher larger than that of rbcL (0.00650), which has been applied as one of classic DNA barcodes (Table S4).

Specific DNA markers for Bidens

In our preliminary study, we retrieved 5 classic DNA barcodes (ITS1, ITS2, psbA-trnH, matK and rbcL) of Bidens species from NCBI (nucleotide database) for identification analysis, because only sequences for ITS1 and ITS2 were adequate, psbA-trnH, matK and rbcL were remove from further analysis. As a result, we found neither ITS nor ITS2 could separate Bidens species at reasonable resolution (Figs. S1 and S2). Based on divergent regions found above, we designed primers (Table S5) for IGS, including ndhD-ccsA, ndhI-ndhG, ndhF-rpl32, trnL_UAG-rpl32, ndhE-ndhG, ndhE-psaC, matK-rps16, ycf1-trnN_GUU, rps2-atpI, cemA-petA, and petN-psbM are verified by in silico PCR amplification by FactPCR 6.7 [47], in order to develop novel molecular markers for Bidens species. As a result, ndhD-ccsA, ndhI-ndhG, ndhF-rpl32, trnL_UAG-rpl32, ndhE-psaC, matK-rps16, rps2-atpI, cemA-petA, petN-psbM were regarded as candidate markers with high confidence (Table S6).

Selective pressure analysis

We calculated the nonsynonymous (Ka) to synonymous (Ks) substitution ratios for all 68 protein-coding genes of CP genomes from 15 Bidens with KaKs_calculator [40] by ‘Ma’ model and statistically tested by Fisher Exact Test(Fig. 6). Overall, Ka/Ks values were less than 0.5 for the majority genes, suggesting that CP genes of the Bidens are conserved and mainly under negative selection during the evolution process, which is reasonable for necessary roles played by chloroplast genes and is in accordance with previous studies. In addition, Ka/Ks values of 4 genes (ndhE, ndhF, ndhG, and rpl32) were greater than 1, indicating these genes undertaking positive selection pressure during evolution, however, we found only 1 substitution in the multiple sequence alignment (MSA) exist for each gene-pairs, indicating relatively low positive selection occurred to these genes. Further molecular biology studies should be started to evaluate the environmental impact on these genes.

Fig. 6
figure 6

Pairwise Ka/Ks ratios in different genes. The scale factors associated with each value are displayed on the right side of the graph. The color close to red indicates that the gene has a high Ka/Ks ratio. Red indicates positive selection that is significantly enriched, while blue indicates negative selection that is significantly reduced. See Table S7 for abbreviations

Full size image

Phylogenetic reconstruction of Bidens

Phylogenetic relationships among 14 Bidens species were re-constructed by RaxML-ng software. The phylogenetic tree indicating that the three assemblies of Bidens in China are monophyletic and separated from other Bidens species (Fig. 7). B. bipinnata and B. alba were clustered together with supporting rate of 1.0, to which B. pilosa was sister group, indicating they are closer related species than others included in this study. Furthermore, all Bidens constructed monophyletic relationships indicated they might diverge from the same ancestor and evolved independently around the world. These findings made us confident in proposing a hypothesis that Bidens species were indeed highly adaptive to the changing environment world-widely, and were suffering a higher level of radiation adaptive selection compared with that in Bidens species occurred at Hawwii.

Fig. 7
figure 7

The phylogenetic tree reconstruction base ML

Full size image

Discussions

In this study, we collected three Bidens species distributed in China and assembled their complete chloroplast genome. We found CP genomes for Bidens species varied little in size, but also conserved in genome structure, gene composition and gene order, these findings were in accordance with previous reports that the structure of plastome geome in most angiosperm is generally maternal-inherited and highly conserved [48]. GC content is unevenly distributed at LSC, IR and SSC region, wherein GC content of IR is the highest while the lowest in SSC, higher GC content found in IR region might be resulted by the presence of rRNA (rrn4.5, rrn5, rrn16, and rrn23) that has been reported previously in Asteraceae cp genomes [49,50,51]. A CP gene is rarely lost arbitrarily, it is either trasferred to the nuclear genome or functionally replaced by a nuclear gene [52]. Interesting in our study, we found clpP was lost from B. bipinnata and psbD lost from B. alba and B. bipinnata. Noteworthy, deletion of one intron from clpP has been reported in Vicia sepium [53].

SSRs have high polymorphism and have been widely used in the study of phylogeny and biogeographic development [54, 55]. Similar number of SSRs and long repeats were found in Bidens, while it varied among species in type, which were vastly distributed in the IGS region of LSC, wherein the majority of the SSRs are mononucleotide repeats (A/T) and the majority of long repeats are forward and palindromic repeats, respectively. These information might be useful in developing novel molecular markers. IR expansion and contraction is one of reason causing size variation for CP genome which has been reported in Polystachya dendrollifora [42], Amomum villosum [16], Aristolochia debilis [19], and Styrax species [17, 56]. The length of IR varied among Bidens species indicated IR expansion and contraction might have been occurred in B. pilosa and B. bipinnata respectively. Detailed comparison of four IR/SC junctions of Bidens species showed that the border structures were highly similar to each another (Fig. 2). Although the boundary regions of the cp genome were relatively stable, we found rpl22, ycf1 located at boundray regions might have been shiffted by IR expansion and contraction.

Because Pi for protein-coding genes and Intergenic regions (IGS) averaged the highest among IR, LSC and SSC, we hypothesized IGS like ndhD-ccsA, ndhI-ndhG, ndhF-rpl32, trnL_UAG-rpl32, ndhE-ndhG, ndhE-psaC, matK-rps16, ycf1-trnN_GUU, rps2-atpI, cemA-petA and petN-psbM were candidate markers for Bidens species' separation. In addition, we tested them by in silico PCR and found nhD-ccsA, ndhI-ndhG, ndhF-rpl32, trnL_UAG-rpl32, ndhE-psaC, matK-rps16, rps2-atpI, cemA-petA, and petN-psbM were highly confident candidates. However, further study based on wet-experiment should be conducted to get the exact conclusion.

It seems that ndhE, ndhF, ndhG, and rpl32 in Bidens were under positive selection while the majority of genes were under purifying selection, which is logically sound as CP genes are functional important for photosynthesis. Additionally, we found Bidens in China clustered together but formed monophyletic clade with other Bidens species from phylogenetic analysis, indicating Bidens species might are highly adaptive species to the changing environment world-widely.

Conclusions

CP genomes are helpful in revealing intra-species relationships, but also in identification of closely-related inter-species identification. In this study, we assembled CP genomes for 3 Bidens species endemic to China and conducted CP-genome-widely comparison thoroughly. The genome structure variation, IR expansion and contraction, phylogenetic relationships, mutation hotspot and candidate molecular markers found in this study would provide useful information for genetic diversity analysis and molecular identification for Bidens species.