**3. Discussion**

*Pyrus hopeiensis* is a valuable wild resource of *Pyrus*, which belongs to the family Rosaceae. Because of its limited distribution and population decline, *P. hopeiensis* is listed among "the wild plants with tiny population" in China. It belongs to one of 13 species of *Pyrus* present in China. In this study, the chloroplast genomes of the two genotypes *P. hopeiensis* HB-1 and *P. hopeiensis* HB-2 and those of three other major pear plants, *P. ussuriensis* Maxin. cv. Jingbaili, *P. communis* L. cv. Early Red Comice, and *P. betulifolia*, were analyzed using high-throughput sequencing for comparative analysis. The chloroplast genome of *Pyrus*, like those of most higher plants, is a typical tetrad consisting of two reverse repeat IR regions and small and large single copy fragments [24]. There was a 46 bp difference in the chloroplast genome size between the two *P. hopeiensis*, which was located in the LSC region. Compared with the other three *Pyrus* species, the total genome length was <225 bp, and the gene number, gene type, gene sequence, LSC, IR, and SSC lengths and GC content were similar. This strongly suggests that chloroplast genomes are highly conserved [25]. The encoded genes of the chloroplast genome are divided into three categories based on their functions. The first is related to chloroplast gene expression, such as tRNAs, rRNAs, and three subunits encoding chloroplast RNA polymerase synthesis. The second is related to photosynthesis, and the third consists of other biosynthesis genes and some genes of unknown function, such as mat and *ycf* [26]. The chloroplast genes of *Pyrus* are similar in composition.

The genomic sequence of the *P. hopeiensis* HB-1 chloroplast was used as a reference sequence to detect single-nucleotide polymorphisms (SNPs) and indels in the other four *Pyrus* species. The results showed a significantly higher variation in the non-coding region than in the coding region and more mutation sites in the intergenic region of the *psbA*-*trnQ\_TTG*, *rpl18*-*rps20*, and *trnT-TGT*\_*trnF\_GAA* genes, which could be used for evolutionary analysis of *Pyrus*. The chloroplast genomes of *Pyrus* show obvious codon preference and similar codon use frequency. Furthermore, the third chloroplast codon has a higher A/T preference. This phenomenon is common in the chloroplast genomes of other higher plants [27]. Although the IR region is highly conserved, the expansion and contraction of the IR region is a common characteristic of the chloroplast genome. The degree of expansion of the IR/SC boundary is similar among the five *Pyrus* species, the two *P. hopeiensis* genotypes contain few genes with different extension positions, and any differences are very small, which is useful in the classification of *Pyrus*, as it can be used as a basis to identify the evolution of the chloroplast genome. The classification and identification of pear species in this study can be utilized for the preservation of pear germplasm resources. The initial identification of species and varieties of *Pyrus* was mainly based on morphological features (leaves, petioles, floral organs, sepals, hairs, fruits, and ventricles) and geographical distribution. For example, based on an investigation of morphological characteristics and natural distribution, Chinese taxonomists believe that *P. hopeiensis*, *P. phaeocarpa*, *P. sinkiangensis,* and *P. serrulata* were all formed by natural crosses [28]. Yu [29] divided *Pyrus* from China into 13 species based on their serrated leaf margins, and these included *P. hopeiensis*, *P. betulifolia*, *P. ussuriensis*, *P. phaeocarpa*, *P. bretschneideri*, *P. pyrifolia*, *P. pashia*, *P. armeniacaefolia*, *P. calleryana*, *P. pseudopashia*, *P. serrulata*, *P. sinkiangensis*, and *P. xerophila*. Anatomical studies in Wang Yingzhong [30] revealed that the anatomical structures of *P. betulifolia* and *P. ussuriensis*, and *P. bretschneideri*, *P. pyrifalia*, and *P. communis* were similar. The results showed that the relationship between *P. ussuriensis* and *P. betulifolia*, *P. bretschneideri,* and *P. pyrifalia* was close. However, it is easy to cross *Pyrus* species and there are no obvious differences in the biological and morphological characteristics among species and varieties, which greatly increases the difficulty of establishing its phylogenetic evolution and classification.

Pollen morphological identification, cytological markers, isozymes, and other methods have also been studied with a view to classifying *Pyrus*. The pollen morphology of *P. sinkiangensis* is similar to that of the Western pear, indicating a close relationship [31]. The pollen morphology of the Western pear is obviously different from that of the Oriental pear. The pollen morphology of *P. calleryana* has many primitive characteristics, and it is a primitive species of *Pyrus* in China. The pollen morphology of *P. bretschneideri*, also present in China, has the characteristics of both *P. pyrifalia* and *P. ussuriensis*, and may be a natural hybrid of *P. pyrifalia* and *P. ussuriensis*. Cytological markers enabled the analysis of the number, banding, karyotype, and meiosis behavior of the chromosomes. *P. phaeocarpa* has a similar karyotype to that of *P. betulifolia*, and those of *P. sinkiangensis*, *P. hopeiensis*, and *P. serrulata* were also similar [32,33]. In 1983, Lin Bonian and Shen Dexu [34] proved, through the use of the peroxidase isozyme, that *P. bretschneideri* and *P. pyrifolia* were closely related. However, these methods have few characteristic sites, poor polymorphism, and low accuracy, and provide a limited amount of information. To date, the relationships among *Pyrus* species, their origin, evolution of cultivation systems, and the origin of some suspicious species and hybrids remain unclear.

In recent years, molecular markers based on DNA, such as restriction fragment length polymorphisms (RFLPs) and simple sequence repeats (SSRs) have been used to investigate the genetic relationships, genetic diversity, and germplasm of *Pyrus*. However, there remain some deficiencies in the study of the interspecific relationships and origins of hybrids. Results based on random amplification of polymorphic DNA (RAPD) showed that the origin of *P. sinkiangensis* involved the crossing of many Eastern and Western pear species and that the genetic relationship between *P. bretschneideri* and *P. pyrifolia* is very close [35]. RAPD, inter sequence simple repeats (ISSR), and other DNA markers showed that *P. hopeiensis*, *P. betulifolia* and *P. phaeocarpa* were closely related to each other. In the same way, *P. phaeocarpa* is considered to be a hybrid of *P. betulifolia* and *P. ussuriensis*, whereas *P. hopeiensis* is a hybrid of *P. phaeocarpa* and *P. ussuriensis*. In a study using RAPD, *P. hopeiensis* and *P. phaeocarpa* shared some spectral bands with *P. betulifolia* and *P. ussuriensis* [36–38]. Zheng et al. identified a close relationship between *P. ussuriensis* and *P. hopeiensis* using internal transcribed spacer (ITS) sequences, which is consistent with the results in our study [39].

Because the chloroplast genome is the second-largest genome after the nuclear genome, it is maternally inherited in most angiosperms; thus it reflects the maternal evolutionary history, and this helps us to understand the maternal ancestors of suspected hybrids. The coding and non-coding regions of the chloroplast genome evolve at different rates, making them suitable for systematic research at different levels. The coding region is highly conserved and is only suitable for phylogenetic studies of families, orders, and higher taxonomic levels, whereas the non-coding regions are less constrained by function and the rapid evolutionary rate is suitable for plant phylogenetic studies at interspecific and subspecies levels. At present, the successful design of a set of universal primers for the chloroplast gene non-coding regions (such as *trnS*-*psbcc*, *trnL*-*trnF* and *accD*-*pasI*) [40] has made the study of chloroplast non-coding regions a hot topic in studies of the systematic relationship of *Pyrus*. Phylogenetic trees based on combinations of the sequences of *trnL*-*trnF* and *accD*-*psaI* in the chloroplast non-coding regions have further confirmed the theory of an independent evolution of the Oriental pear and the Western pear from the background of matrilineal evolution, and have shown the close relationship between *P. bretschneideri* and *P. pyrifolia* [41]. A study of the *trnL*-*trnF* region of cpDNA showed that *P. sinkiangensis* is closely related to the Western pear and the Oriental pear; the relationship between *P. betulifolia* and *P. ussuriensis* is close; *P. bretschneideri* is a hybrid of *P. ussuriensist*, *P. phaeocarpa*, and *P. pyrifolia*; and that the Western pear and Oriental pear are related to each other [42]. The sequences of these regions are highly conserved, with only limited sites available to provide information to unravel the phylogeny of *Pyrus*. However, no comprehensive and systematic cpDNA sequence analysis of *Pyrus* exists in China. To further our understanding of the inter-species relationships of *Pyrus* and to reveal the origin of hybrids and explore the evolutionary model of Eastern and Western pears, a wider range of representative species and varieties of Eastern and Western pears must be selected, and the nuclear gene fragments inherited by their parents should be combined, especially the low copy nuclear gene introns.
