Identification of New Cultivar and Different Provenances of Dendrocalamus brandisii (Poaceae: Bambusoideae) Using Simple Sequence Repeats Developed from the Whole Genome

2.1. Abundance, Frequency, and Characteristics of Microsatellites

Based on bioinformatics analysis, we detected 231,789 microsatellites in the 2813.5 Mb genome of D. brandisii, of which 231,110 (99.71%) microsatellites were located on chromosomes. The number of perfect SSRs and complex SSRs was 203,082 and 28,707, respectively (Figure 1A), with a microsatellite frequency of 82.4/Mb. In 87.62% of perfect SSRs, the proportion of motifs with different numbers of bases varies. The proportion of microsatellites with dinucleotide repeat motifs is the highest at 69.7%; microsatellites with trinucleotide repeat motifs account for 24.56%, followed by which microsatellites with 4–6 nucleotide repeat motifs account for a smaller proportion, being 4.1%, 1.1%, and 0.5%, respectively (Figure 1A).

The 28,707 compound SSRs identified in this study were composed of 2–78 perfect SSRs and intermediate sequence connections of less than 100 bp. Among them, the number of composite SSRs containing two perfect SSRs was the highest (21,789), accounting for 75.9%, followed by the number of composite SSRs containing three perfect SSRs (4069), accounting for 14.2%. The number of composite SSRs containing 4–78 perfect SSRs was relatively low (1–1119), accounting for 0.003% to 3.899% (Table S1). For compound SSRs containing two perfect SSRs, the length of the connecting sequence is mostly 0 nt (9124), accounting for 41.87%, followed by 1 nt (1305) and 2 nt (1258), accounting for 5.99% and 5.77%, respectively. There are fewer connecting sequences with a length of 3 nt to 100 nt (41–357), accounting for 0.19% to 1.64% (Table S2).

2.2. Frequency and Characteristics of Motif Repeat

We identified 280,317 motifs, and the number of motifs with different base numbers decreased with increasing motif length (Figure 1B). The main types of motifs with different base numbers are as follows: the di-nucleotide motifs are mainly AT/AT, AG/CT, and GA/TC; the tri-nucleotide motifs are mainly GCC/GGC, CGC/GCG, and CCG/CGG; the tetra-nucleotide motifs are mainly CATA/TATG, ATAC/GTAT, and TACA/TGTA; the penta-nucleotide motifs are mainly GAGCC/GGCTC, CGAGC/GCTCG, and AAAAG/CTTTT; and the hexa-nucleotide motifs are mainly CATATA/TATATG, ATATAC/GTATAT, and TATACA/TGTATA (Table 1).

Among the 203,082 perfect SSRs identified in this study, the number of repeats of di-nucleotide motifs ranged from 6 to 187, with those with a repeat frequency of 6 to 8 dominant, accounting for a total of 59.00%. The number of repeats of tri-nucleotide motifs ranged from 5 to 127, with those with a repeat frequency of 5 to 6 dominant, accounting for a total of 85.41%. The number of repeats of tetra-nucleotide motifs ranged from 5 to 571, with those with a repeat frequency of 5 to 6 dominant, accounting for a total of 67.15%, The number of repetitions of the penta-nucleotide motif is 5–72, with those with 5–6 repetitions dominating, accounting for a total of 87.80%. The number of repetitions of the hexa-nucleotide motif is 5–42, with those with 5–6 repetitions dominating, accounting for a total of 58.24% (Figure 2; Table S3).

2.3. The Distribution of SSRs in the Genome

Among the 231,789 identified SSRs, 99.71% were evenly distributed on 70 chromosomes of D. brandisii, and 0.29% were not successfully localized on chromosomes (Figure 3A–C). Based on the location of SSRs on the genome and genome annotation files, we divided these SSRs into six regions, including 5’UTR, 3’UTR, exons, introns, intergene, and multiple mapping regions. SSRs located in multiple mapping regions are defined as SSRs that can be mapped to two or more regions simultaneously. As shown in Figure 3D, through the analysis of SSRs, the results show that SSRs are generally mapped to the intergenic region, and multiple mapping SSRs also account for a large proportion, followed by SSRs mapped to introns and exons, while SSRs mapped to 5’UTR and 3’UTR are relatively few in number.

2.4. Obtain 34 Pairs of Polymorphic Primers and Perform Gene Mapping on 6 of Them

We utilize the Primer3_core program for Perl language-designed batch primers for all microsatellite flank sequences, resulting in 20,228 pairs of primers. Then, we randomly selected 800 pairs of primers for synthesis, combined PCR and agarose gel electrophoresis to preliminarily screen them, and the results showed that 433 pairs of primers can amplify at least one band (Figure S2). We selected 72 pairs of specific primers that can only amplify one band to conduct TP-M13-SSR PCR amplification on ‘Manxie No.1’ and 11 different geographical provenances of D. brandisii materials. The capillary electrophoresis results showed that the results of the three individuals within each provenance population are consistent, so any one of the samples can be selected as a representative sample for that provenance population. And 34 pairs of primers (Table S4) exhibited polymorphism, with adjacent alleles of each marker differing by multiples of the repeat unit length, indicating that these 34 polymorphic SSR markers are effective and have conservation and transferability across cultivars and provenances.

In addition, we performed gene mapping on 6 SSRs (SSR605, SSR243, SSR711, SSR710, SSR673, and SSR182) located on introns from 34 SSR markers, which, respectively, corresponded to DhB21G011140, DhB31G002880, DhA19G015160, DhA19G013950, DhB10G011540, and DhB31G019250. According to the annotation file, DhB21G011140 belongs to the CSN8/PSMD8/EIF3K family and is related to post-translational modifications, protein turnovers, and chaperones. DhB31G002880 is a gene encoding protein tyrosine and serine/threonine kinase and is associated with signal-transduction mechanisms. DhA19G015160 is an aminotransferase class I and II enzyme involved in the process of amino acid transport and metabolism. DhA19G013950 contains the Xrn1 helical domain, which is involved in the nuclear-transcribed mRNA catabolic process. DhB10G011540 contains the caspase domain, which belongs to the MC family of genes and may play an important role in plant immune responses. DhB31G019250 encodes SOS response-associated peptidase (SRAP). The promoters of these six genes all contain multiple light-responsive elements, hormone-responsive elements, and elements necessary for anaerobic induction, indicating that they may participate in the response process to light, hormones, and anaerobic conditions. In addition, some of these genes also contain elements related to low temperature, drought, defense, and meristem expression (Figure 4).

2.5. Genetic Diversity Analysis of ‘Manxie No.1’ and 11 Different Geographical Provenances of D. brandisii Materials Based on 34 Polymorphic SSRs

Genetic diversity analysis was conducted on the ‘Manxie No.1’ and 11 different geographical provenances of D. brandisii representative materials using 34 polymorphic SSR markers. A total of 34 polymorphic SSR markers amplified 97 alleles, with an average of 2.85 alleles per marker. Among them, SSR201 amplified the largest number of alleles (Na = 10). In contrast, 22 markers, such as SSR431 and 711, amplified the fewest alleles (Na = 2). The range of Ne was 1.0868 to 4.1739, with an average of 1.8280 and an average Ho value of 0.6198, which was higher than the average He value of 0.5072. The average value of I is 0.6403 (range of 0.1732–1.7129). The PIC range of polymorphic information is 0.0767–0.7344, with an average value of 0.3152. There are 17 pairs of primers with PIC < 0.25, indicating low polymorphism. The PIC of seven primer pairs is greater than 0.5, which indicates that they possess high polymorphism (Table 2). The correlation graph between Na and PIC values shows that as the number of alleles increases, the polymorphism of specific loci increases (Figure S3). The genetic parameters calculated based on 34 SSR markers indicate that there is a moderate degree of genetic diversity and variation rate between 11 different geographical provenances of D. brandisii and ‘Manxie No.1’.

2.6. Cluster Analysis of 1 ‘Manxie No.1’ and 11 Different Geographical Provenances of D. brandisii Materials Based on 34 SSR Markers

Based on NTSYS v.2.10e software, the genetic similarity coefficient was calculated using the SM coefficient, and a UPGMA tree was constructed based on the genetic similarity coefficient. Similarly, as with result 2.5, we conducted cluster analysis using a representative sample from each population. The analysis showed that 1 ‘Manxie No.1’ and 11 different geographical provenances of D. brandisii could be clustered into 2 main groups at a similarity coefficient of 0.45. The first group also includes two subgroups; the first subgroup is composed of D. brandisii from Chiang Mai, Thailand; Mu Cang Chai County, Vietnam; and Xishuangbanna Dai Autonomous Prefecture, China (the following text is abbreviated as ‘Xishuangbanna’). The genetic similarity coefficients of any two materials range from 0.8454 to 0.9072. Among them, the genetic similarity coefficient of D. brandisii from Mu Cang Chai County, Vietnam, and Xishuangbanna, China, indicates a closer genetic relationship. The second subgroup includes D. brandisii from Pu’er City, Honghe Hani, and Yi Autonomous Prefecture (the following text is abbreviated as ‘Honghe’); Baoshan City, Yunnan Dehong Dai, and Jingpo Autonomous Prefecture (the following text is abbreviated as ‘Dehong’); Yuxi City and Lincang City in Yunnan Province, China; as well as D. brandisii from Guangzhou City in Guangdong Province, China. The genetic similarity coefficient of any two materials ranges from 0.7938 to 0.9381. The genetic similarity coefficient between D. brandisii in Pu’er and Honghe, China, which are geographically adjacent, is the highest at 0.9381, indicating the closest genetic relationship. The genetic similarity coefficient between D. brandisii in Yuxi and Lincang*, China, which are not geographically adjacent, is the second highest, at 0.9278. The genetic similarity coefficient between D. brandisii from the geographically adjacent Baoshan and Dehong, China, is also relatively high, at 0.9072. The independent formation of the second group by ‘Manxie No.1’ proves that ‘Manxie No.1’ has a relatively distant genetic relationship with D. brandisii, which is consistent with the fact that it is an excellent cultivar artificially selected (Figure 5; Table S5). This result may be due to the limited gene exchange between ‘Manxie No.1’ and D. brandisii. The clustering results indicate that there is a certain correlation between the genetic relationships of different geographical provenances of D. brandisii and their geographical distribution. However, it is not completely related to their geographical distribution. Further verification of the accuracy of the UPGMA tree was carried out through the Cophenetic values subroutine in the Clustering program and the Matrix comparison plot subroutine in the Graphics program. The matrix correlation r value of the correlation test was 0.97856, proving that the reliability of the UPGMA tree is very high.

D. brandisii, Dendrocalamus hamiltonii, and Dendrocalamus asper are all part of the Dendrocalamus genus and are known as the world’s three largest bamboo plants with sweet bamboo shoots. We conducted cluster analysis on 2 samples of D. asper, 1 sample of D. hamiltonii, and the aforementioned 12 materials. The results showed that the genetic relationship between D. hamiltonii and D. brandisii was closer than that between D. asper and D. brandisii (Figure S4), and these three types of sweet dragon bamboo could each cluster into one branch. The clustering results were consistent with existing classifications, confirming the applicability and effectiveness of the microsatellite markers identified in this study.

2.7. SSR Core Primers and DNA Fingerprinting of 1 ‘Manxie No.1’ and 11 Different Geographical Prove-Nances of D. brandisii Materials

In this study, the primer SSR201 can generate eight band patterns, which can distinguish eight different materials, followed by primer SSR84, which can generate nine band patterns and distinguish nine different materials. These two pairs of primers can distinguish all 12 materials. Therefore, it is determined that this primer combination (SSR84 and SSR201) is the core primer for these 12 materials. By using this primer combination, 12 materials can be distinguished (Figure S5). The overall prediction accuracy of 34 pairs of SSR primers for 12 materials was 100%, and the DNA fingerprinting can be found in Figure 6 and Table S6.

3.1. SSR Marker Features Developed Based on the Whole Genome of D. brandisii

This study conducted genome-wide SSR identification and analysis on D. brandisii. A total of 231,789 microsatellites were detected from the genome of D. brandisii at 2813.5 Mb, with a microsatellite frequency of 82.4/Mb. The number of microsatellites was much larger than that of smaller genomes such as Arabidopsis thaliana (119.7 Mb) and Oryza sativa (374.5 Mb). However, the microsatellite frequency was much lower than both (135.5/Mb in A. thaliana and 165.5/Mb in O. sativa). Compared with the smaller genomes of Phyllostachys edulis (2051.7 Mb, 127,593, 62.2/Mb) and Zea mays (2066.4 Mb, 107,742, 52.1/Mb), the number of microsatellites is about twice that of the two, and the frequency of microsatellites is also higher [22]. This is inconsistent with the previous research findings that species with larger gene groups have more SSRs but a lower SSR density [32,33]. The number and density of microsatellites may not be related to the genome size.

Generally speaking, there are a large number of di-nucleotide and tri-nucleotide repeats, while the abundance of tetra-nucleotide to hexa-nucleotide repeat sequences is relatively low in the eukaryotic genome [34,35]. Among them, di-nucleotide repeat sequences are considered an indispensable resource due to their highest mutation rate and abundant quantity [36]. In this study, 94.3% of microsatellite repeat motifs were di-nucleotide and tri-nucleotide repeats, of which nearly 70% were di-nucleotide repeats. However, tetra-nucleotide, penta-nucleotide, and hexa-nucleotide repeats only accounted for a very small portion (5.7%). Similar distributions of repeat motifs were also observed in other Poaceous species, such as Ph. edulis, Z. mays, O. sativa, B. distachyon, and S. bicolor, and di-nucleotide and tri-nucleotide repeats still dominate [22]. This is consistent with the usual view that the abundance of microsatellites decreases with increasing repetition length [37,38].

Previous studies have shown that the GC content is higher in animal genomes, while the AT content is higher in plant genomes [39,40]. The data in this study show that the AT content in D. brandisii is also higher than the GC content, which is consistent with other bamboo species such as Guadua chacoensis, Dendrocalamus latiforus, etc. [15,41], and is also consistent with other species such as A. thaliana and S. bicolor [29]. Among many species, AG/CT motifs have the highest proportion, including Poaceae plants such as Ph. edulis, Z. mays, O. sativa, B. distachyon, and S. bicolor and non-Poaceae plants such as Epimedium sagittatum [22,42]. However, in our study, AG/CT motifs (19.44%) were the type with the highest proportion in microsatellites except for AT/AT motifs (35.11%), similar to S. bicolor and A. thaliana [22]. Previous studies have shown that in plants, TC repeat sequences are typically present in the transcriptional region and frequently appear in 5’UTRs. CT microsatellites in 5’UTRs may be involved in antisense transcription and play a role in gene regulation [43]. This suggests that approximately 1/5 of the microsatellite markers developed in this study may be involved in the antisense transcription of D. brandisii, thereby regulating the function of related genes. Monocotyledons have more abundant tri-nucleotide repeat sequences rich in GC than dicotyledonous plants [44]. Monocotyledons such as O. sativa, B. distachyon, and S. bicolor have the highest content of CCG/CGG, while Ph. edulis has the second highest content of CCG/CGG [22]. However, dicotyledonous plants such as A. thaliana and Glycine max rarely exhibit CCG/CGG [44]. In this study, GCC/GGC was the type with the highest proportion of tri-nucleotide repeat motifs in D. brandisii, accounting for 11.17%, followed by CGC/GCG (8.93%) and CCG/CGG (8.01%), which is consistent with this conclusion. Tetra-nucleotide, penta-nucleotide, and hexa-nucleotide repeat sequences, as in most plants, have a lower frequency.

3.2. Genetic Diversity within Bamboo Populations and Determination of Representative Samples for Various Populations of D. brandisii

This study fully considered the typicality and representativeness when selecting samples. We conducted a detailed investigation and analysis of the sample collection site to ensure that the environmental conditions and bamboo growth status of the site can represent the overall situation of the entire study area. At the same time, as bamboo mainly reproduces asexually through rhizomes, in order to avoid duplicate sampling, different samples selected from the same population come from noncontinuous bamboo forests with differences in latitude, longitude, and altitude, ensuring that duplicate samples have relatively independent genetic sources. This precisely proves that the sample we have chosen is more representative. By maintaining a larger sampling distance, the likelihood of duplicate samples being disturbed by the same environment can be reduced, allowing them to reflect the main characteristics of the entire population [45]. The samples we selected showed consistent results under these different conditions, further indicating that they can represent the overall characteristics of geographical provenances to some extent.

Previous studies have shown that there is indeed variation within populations of different provenances of D. brandisii, but this variation is very low [46]. This study showed that the Hs value of allelic marker diversity within populations of different provenance of D. brandisii was 0.0382 ± 0.0032; the mean values of the observed number of alleles (Na), the effective number of alleles (Ne), Shannon’s information index (I), and expected heterozygosity (He) among different individuals within various populations are 1.1338 ± 0.2978, 1.0906 ± 0.2168, 0.0742 ± 0.1704, and 0.0507 ± 0.1184. It is generally believed that genetic diversity is low when the Hs value is less than 0.1, the Na value is less than 2, the Ne value is close to or less than 1, the I value is less than 0.5, and the He value is less than 0.3. Given this, it can be seen that these values are relatively low, indicating limited genetic variation and low levels of genetic diversity among different individuals within the same population of D. brandisii. In our study, we selected three independent samples for each provenance population. Only when the results are consistent can we select one sample as a representative sample. If the results are inconsistent, additional samples need to be added for supplementary confirmation. The repeated individuals selected in the study are located at a certain distance and altitude from each other, and the possibility that they are all variant individuals and that the results of the variation reflected in the 34 polymorphic SSR markers are completely consistent is minimal. Therefore, we believe that consistent results from three replicates are sufficient to support the idea of selecting one as a representative sample for preliminary cluster analysis.

In future research, we will continue to improve our sampling methods to further enhance the typicality and representativeness of samples. Nonetheless, we are also aware that increasing the number of samples for each population would be preferable. We plan to expand the sample size, explore more diverse D. brandisii habitats, and conduct more in-depth genetic analysis to capture a more comprehensive picture of its populations and uncover the factors that contribute to limited genetic diversity. This will help us develop more targeted conservation strategies and management plans for D. brandisii resources. Moreover, we will explore innovative research methods and technologies to better understand the genetic characteristics and evolutionary history of D. brandisii. This could include the use of advanced genomic sequencing techniques and computational models to analyze large-scale genetic data. By doing so, we hope to make contributions to the field of D. brandisii research and conservation and promote the sustainable use and protection of these important resources.

3.3. Genetic Diversity and Clustering of 1 ‘Manxie No.1’ and 11 Different Geographical Provenances of D. brandisii

Due to the unique reproductive characteristics of bamboo, the classification problem has been troubling researchers. Molecular markers are genetic markers based on nucleotide sequence differences between individuals or populations, which intuitively and clearly reflect biological diversity at the DNA level. They are suitable and effective for identifying different plant genera, species, and underspecies. Domestic and foreign scholars have used isoenzyme labeling and molecular marker methods such as the RFLP (Restriction Fragment Length Polymer), RAPD (Random Amplified Polymer DNA), AFLP (Amplified Fragment Length Polymer), ITS (Internal Transcribed Spacer), and SSR (Simple Sequence Repeat) to study the genetic diversity and phylogenetic relationships between bamboo species and genera [30,47,48,49,50]. However, due to the characteristics of asexual reproduction of bamboo plants, there is little research on the genetic diversity of bamboo plants at the subspecies level. We used co-dominant SSR molecular marker technology to successfully identify different provenances of D. brandisii and the new cultivar ‘Manxie No.1’. Moreover, we constructed a DNA fingerprinting of 12 materials based on 34 SSR molecular markers to reflect their differences at the DNA level. This provides a molecular basis for the effective identification of different provenance of D. brandisii and the new cultivar ‘Manxie No.1’.

At present, there is relatively scarce information on the genetic diversity and population structure of bamboo germplasm resources. Due to the asexual reproduction of bamboo through rhizomes, it has low intra-species genetic diversity and genetic differences between populations [51]. Therefore, screening for markers with relatively high abundance and polymorphism for bamboo may be very difficult. Abreu developed and screened 7 polymorphic microsatellite markers for Aulonemia aristulata [52], and Vieira developed and screened 16 polymorphic microsatellite markers for the tropical woody bamboo Tribe Bambusae (Poaceae: Bambusoideae) [53]. Weixin developed and screened 20 polymorphic microsatellite markers for Ph. Edulis [32]. In this study, based on the ability of agarose gel to separate DNA fragments, 72 pairs of specific primers that can amplify only 1 band were initially screened from 800 pairs of primers, thus reducing the complexity and uncertainty in subsequent analysis. Then, based on capillary electrophoresis with higher resolution and sensitivity, 34 pairs of polymorphic primers were further selected from 72 pairs of specific primers that can amplify DNA fragments of different lengths in different samples. The differences in fragment lengths reflect the different number of motif repeats in SSR loci, thereby reflecting genetic differences between samples. Using these 34 pairs of polymorphic SSR primers, genetic diversity analysis was conducted on 12 samples (‘Manxie No.1’ and 11 different geographical provenances of D. brandisii). The PIC value takes into account the frequency and number of alleles and can comprehensively reflect the degree of polymorphism of the locus. For example, at an SSR locus, if there are multiple alleles and their frequency distribution in the sample is relatively uniform, the PIC value of the locus will be higher, indicating that the locus has high genetic diversity. In this study, the average PIC value of these 12 samples was 0.3152, ranging from 0.25 to 0.5, indicating that these samples have a moderate level of genetic diversity. This result provides important information for us to understand the genetic background of these samples. The moderate level of genetic diversity means that the samples have both a certain degree of genetic variability and relative stability. This genetic diversity may be due to the samples coming from different geographical provenances, experiencing different environmental choices, and genetic drift. On average, compared with other bamboo species using SSR molecular markers, the genetic diversity level of sweet dragon bamboo is higher than that of Guadua inermis and Ph. edulis [51,54] but lower than the G. angustifolia in the Colombian coffee eco-region and Kuruna debilis [55,56].

Cluster analysis shows that the genetic diversity of D. brandisii and ‘Manxie No.1’ exhibits extensive variation. ‘Manxie No.1’ can be distinguished from different geographical provenances of D. brandisii and D. hamiltonii, and D. asper can also be distinguished from D. brandisii. The SSR84 + SSR201 primer combination can distinguish 15 sweet dragon bamboo materials. However, clustering analysis shows that the genetic relationship between ‘Manxie No.1’ and D. brandisii is farther than that between D. hamiltonii, D. asper and D. brandisii. The possible reason for this result is that the number of microsatellite molecular markers is insufficient, leading to clustering bias. This situation can be improved by further screening polymorphic markers.

3.4. Identification of Candidate Genes Based on Polymorphic SSRs

Based on the development of polymorphic microsatellite markers, we located six of them on introns of six genes, which are non-coding sequences of broken genes and participate in important biological processes such as transcriptional regulation, variable splicing, and nuclear output [57]. The form and function of introns appearing at different positions in the transcript vary [58], and introns can significantly enhance gene expression [59]; this not only affects the transcription, output, and stability of mRNA to increase its content but also improves the efficiency of mRNA translation [60]. Recent studies have shown that introns promote evolution through exon recombination [61] and achieve multiple protein translations of a single gene through variable splicing [62], playing an important role in gene regulation [63]. Through the annotation file, we speculate that these six polymorphic microsatellite markers may have regulatory effects on their respective six genes, thereby regulating processes such as mRNA breakdown metabolism, amino acid transport and metabolism, protein transport and post-translational modification, signal transduction, plant immunity, and SOS response in D. brandisii. Among them, the homologous gene OsLOL3 of DhB10G011540 corresponding to SSR673 in rice can respond to cold stress, heat stress, and pathogen infection (such as Madurella grisea and Rhizoctonia solani). Yeast two-hybrid experiments have shown that OsMC3 can interact with OsLOL1 and may play a role in programmed cell death (PCD), disease resistance, and growth in rice [64], with PCD being the most typical defense stress response in plants; it plays a crucial role in resisting pathogen infection [65], which is consistent with our prediction that DhB10G011540 may play an important role in plant immune response. The promoters of these six genes contain elements related to light response, hormone response, anaerobic induction, low temperature, drought, defense, and meristem expression, suggesting that they may be involved in related response processes, which provides a reference for future in-depth research.

It should be noted that our discussion here is purely speculative. We do not claim any functional correlation between the SSRs and traits at present. Further research is needed to explore any potential functional correlations between these SSRs and specific traits in D. brandisii. For example, future studies could investigate how the presence of these SSRs might affect the phenotypic characteristics of the plant under different environmental conditions.

4.1. Plant Materials, Genomic DNA Isolation, and Detection

A new cultivar, ‘Manxie No.1’, and 11 different geographical provenances of D. brandisii materials were obtained from the main distribution areas of D. brandisii in China, Thailand, and Vietnam (Figure 7). There is a certain distance between the duplicate individuals of each provenance population, and they are distributed in discontinuous bamboo forests with different latitudes, longitudes, or altitudes to ensure that the duplicate samples have relatively independent genetic sources and can reflect the main characteristics of the entire germplasm population [45]. Finally, these fresh leaves were collected in a sampling bag containing a discolored silicone desiccant and were stored in a −80 °C refrigerator for future use.

Genomic DNA was extracted from leaf samples of 3 individuals in each population using a slightly improved cetyltrimethylammonium bromide (CTAB) method [66]. NanoDrop one (Thermo Fisher Scientific, Waltham, MA, USA) was used to determine the DNA concentration, and 1.0% agarose gel electrophoresis was used to detect the DNA quality.

4.2. Microsatellite Identification and Primer Design

All microsatellites in the entire genome sequence of D. brandisii (the genome assembled in reference [7] (https://doi.org/10.1111/jipb.13592 (accessed on 10 December 2023))) were identified using the Perl language platform-based microsatellite identification tool (MISA) (http://pgrc.ipk-gatersleben.de/misa/ (accessed on 6 January 2024)) [67]. The parameter settings are as follows: the di-nucleotide motif should be repeated at least 6 times, and the tri-nucleotide, tetra-nucleotide, penta-nucleotide, and hexa-nucleotide motifs should be repeated at least 5 times. When the sequence length between two simple sequence repeats (SSRs) is less than 100 bp, it is considered a composite SSR [68]. MISA is still used to analyze the number of motifs, repetitions, the number, length, and starting and ending positions of microsatellites with different repeat types on the genome.

The Primer3_core program (https://github.com/primer3-org/primer3 (accessed on 20 January 2024)) suitable for Perl language is used for batch primer design of microsatellite flanking sequences. The main parameter settings are as follows: (i) the primer length is 18–25 bp, (ii) Tm value is 57 °C–62 °C, (iii) the size of the target PCR product is 120–400 bp, and (iv) the GC-content is 40–60%.

4.3. Screening for Specificity and Polymorphism of Primers

The length of SSR is the main factor affecting its polymorphism, and the sequence length is directly proportional to the polymorphism. When the SSR length is ≥ 20 bp, there is a high degree of polymorphism [37,69]. Therefore, the motifs with a length of 2–6 bp are sorted in descending order according to the number of repetitions. We selected 800 pairs of SSR primers that ranked high and a distribution covering 70 chromosomes for synthesis and used ‘Manxie No.1’ genomic DNA as a template for PCR amplification. When the size of the agarose gel electrophoresis band of the PCR amplification product meets the expectation, it can be identified as a specific primer.

Subsequently, TP-M13-SSR PCR was performed using genomic DNA from different geographical provenances of D. brandisii and the new cultivar ‘Manxie No.1’ as templates, and the obtained specific primers were further screened for polymorphism. All PCR and electrophoresis were evaluated for consistency and reproducibility of results through multiple repeated experiments. This method consists of three primers: a forward primer with an 18 bp M13 tail sequence (5’–TGTAAAAACGACGGCCAGT–3’) added at the 5’ end, an M13 universal forward primer with a fluorescent group (6-carboxy-X-rhodamine (ROX, red), hexachloro-6-carboxy-flurrescine (HEX, yellow), tetrachloro-6-carboxy-flurrescine (TET, green) or 6-carboxy-flurrescine (FAM, blue)), and a normal specific reverse primer. The PCR system and reaction program adopt a two-step method, as detailed in Table S7. Fluorescence capillary electrophoresis SSR analyzed PCR end-products on the ABI 3730xl DNA analyzer (Thermo Fisher Scientific, USA), and the results were analyzed using Genemarker v2.2.0 software (https://softgenetics.com/products/genemarker/ (accessed on 22 May 2024)) [70]. Here, due to the relatively limited genetic variation and low level of genetic diversity among different individuals within the same provenance population of D. brandisii [46], we analyzed and read the results of three individuals in each provenance population. If the three results are completely consistent, we selected any one of them as the representative sample of the provenance population. Otherwise, repeat individuals will be added for further validation to determine representative samples for genetic diversity analysis and cluster analysis.

4.4. Data Analyses and Fingerprint Construction

POPGEN32 v1.32 software [71] was used to calculate Nei’s diversity index (Nei’s), Shannon’s information index (I), average observed heterozygosity (HO), and expected heterozygosity (He). PowerMarker v. 3.25 software [72] was used to calculate polymorphism information content (PIC), and NTSYS v. 2.10e software [73] was used to calculate genetic distance and plot the UPGMA cluster tree. Other data analyses and fingerprint construction were completed in Excel 2019.