2.1. Biological Materials

Twenty-two batches of biological materials were collected from the Germplasm Source and Seedling Breeding Center of “Guang chenpi” (Table 1). 22 Citrus reticulata “Chachi” samples were different in variety, plant propagation, rootstock, and tree age.

2.2. DNA Extraction

Genomic DNA was extracted using the plant DNA extraction kit (TSP101-200) of Tsingke. The quality of the extracted genomic DNA was checked by 1% agarose gel electrophoresis with DL2000 DNA marker, and the concentration of them was carried out through the NanoDrop 1000 (Thermo Fisher Scientific, Waltham Massachusetts, US).

2.3. PCR and Sequencing of DNA Barcodings

Genomic DNA was diluted to 15 ng·μl⁻¹ and then was amplified by performing polymerase chain reaction (PCR) using 4 pair of universal primers of DNA barcodings listed in Table 2 [18]. PCR was performed under the following conditions: initial denaturation at 98°C for 2 min, followed by 30 cycles with 98°C denaturation for 10 s, annealing at the melting temperatures (TM) listed in Table 2 for 10 s, and extension at 72°C for 10 s. The final extension step was performed for 5 min at 72°C. Next, an aliquot of the amplification product was resolved on 1% agarose gel electrophoresis documented with a gel documentation system and further analyzed by sequencing.

2.4. DNA Library Construction and Illumina Sequencing

Genomic DNA will be randomly interrupted, the end will be repaired, “A” will be added, and the unique connector of DNBSEQ-T7 sequencer will be added. Then, DNA libraries will be constructed by PCR enrichment. Finally, the DNA library was denatured, cycled, and digested to obtain single-stranded circular DNA. Single-stranded circular DNA was amplified by rolling circle amplification (RCA), further producing DNA nanoball (DNB). Illumina sequencing was performed on DNBSEQ-T7 sequencer after the DNA libraries were qualified.

2.5. Whole-Genome Resequencing Data Quality and Filtering

To exclude bias from low-quality reads that arise from the process of base-calling or adapter contamination, the quality of the raw data obtained by whole-genome resequencing was evaluated until the value of Q₃₀ was over than 85%. The clean reads were used for subsequent bioinformatics analysis. For further analysis, we downloaded previously published genomic data of Citrus reticulata from the NCBI (GenBank accession number ASM325862v1). We mapped high-quality data per individual to the reference Citrus reticulata genome using Burrows–Wheeler Aligner (BWA) software [21], then the sequencing read depth and genomic coverage of each sample were counted, and the variation was detected.

2.6. SNP and InDel Calling

SNPs and InDels can be called by mapping the unitigs against a reference genome. The main calling procedures are as follows: (1) for the results of BWA comparison, Mark Duplicate tool of Picard software is used to remove the duplication and shield the influence of PCR-duplication; (2) the Genome Analysis Toolkit (GATK) software [22] is used to perform InDels realignment, with local realignment of the sites near the alignment result with insertion-miss alignment and correction of alignment errors due to insertion-miss alignment; (3) GATK software was used for base recalibration to calibrate the base masses; (4) variant calling of SNPs and InDels was performed by GATK software; and (5) SNPs and InDels with any of the following features were filtered: two SNPs within 5 bp; SNPs within 5 bp near InDel; and two InDels within 10 bp [23].

3.1. Quality and Concentration of the Extracted DNA

In this work, Genomic DNA was extracted from tender leaves of 22 Citrus reticulata “Chachi” samples by using the plant DNA extraction kit, and OD value (A260/280) and DNA concentration are shown in Table 3. The results showed that the concentration of DNA could be used in subsequent experiments.

3.2. Sequence Features and Homologous Analysis of DNA Barcodings

According to the agarose gel electrophoresis result of PCR amplification products (Figure 1), ITS2, rbcl, matK, and psbA-trnH produced amplification bands of approximately 750 bp, 750 bp, 1000 bp, and 500 bp, respectively. The electrophoresis bands of each sample were uniform, bright, and nonspecific heterozygous, indicating that the success rate of sequence amplification was 100%, which could be further analyzed by sequencing.

Four barcodes (ITS2, rbcl, matK, and psbA-trnH) were analyzed by DNAMAN software for the length and base composition of each sequence fragment and further identified by BLAST in Genebank. The success rates of PCR amplification and sequencing of 3 DNA barcodings (ITS2, rbcl, and matK) of 22 samples were 100%. However, due to the large number of fragments missing in the sequencing of two samples (A16 and A21) of the psbA-trnH barcode, 20 psbA-trnH sequences were actually obtained in the experiment. PCR success rate, barcoding length, GC content, variable site, and BLAST rate of 4 DNA barcodings are listed in Table 4. The aligned partial sequences had lengths of 232 bp, 680∼682 bp, 1059 bp, and 537∼538 bp for ITS2, rbcl, matK, and psbA-trnH, respectively. Among them, the ITS2 barcode had the advantages of shorter sequence length and higher GC content (71.60%), followed by the psbA-trnH barcode having shorter sequence length.

Moreover, homologous analysis of 4 DNA barcodings about 22 batches of Citrus reticulata “Chachi” samples was carried out by DNAMAN software. The homologous analysis result showed that the homology of ITS2 of 22 samples was 100.00%, which indicated that ITS2 maintained a high degree of genetic stability in Citrus reticulata “Chachi” of different propagation methods, different tree ages, different varieties, and different rootstocks. Also, the homology of rbcl, matK, and psbA-trnH of 22 samples was 99.97%, 99.99%, and 99.81%, respectively.

In addition, among the 22 batches of CRP samples, 2 SNP sites were identified in the matK barcode and 1 Indel site was identified in the psbA-trnH barcode. The results showed that there were still some variations within the species of Citrus reticulata “Chachi,” making us realize molecular breeding of Citrus reticulata “Chachi” and the distinction of Citrus reticulata “Chachi” and related species need more valuable molecular markers. Therefore, whole-genome resequencing was also performed on 22 Citrus reticulata “Chachi” samples of different propagation methods, different tree ages, different varieties, and different rootstocks, which provided more scientific basis for molecular breeding.

Previous studies have reported that the ITS2 region was selected for discrimination of the four CRP cultivars including Citrus reticulata “Chachi,” Citrus reticulata “Dahongpao,” Citrus reticulata “Unshiu,” and Citrus reticulata “Tangerina,” while ITS, trnH-psbA, and rbcL could not distinguish these CRP samples [18]. Different from the existing studies, this work focuses on the hereditary stability of 4 barcodes including ITS2, rbcl, matK, and psbA-trnH in Citrus reticulata “Chachi” with different varieties, propagation methods, grafting rootstocks, and tree ages. Because of DNA degradation in moderately or highly processed products with time, PCR amplification of standard-length (around 650 bp) barcodings is a huge challenge [24]. Combined with existing research and the result in this work, ITS2 was considered to be a useful DNA barcoding to distinguish Citrus reticulata “Chachi” from other varieties, which presented not only low variability within a certain taxa but also a high level of interspecies variability. Also, this work indicated that matK was not considered because of its long length and variable sites within taxa, while rbcl and psbA-trnH had the potential to distinguish Citrus reticulata “Chachi” from other varieties. Actually, combining DNA barcodes in the ribosome and in the chloroplast makes it more convincing in species identification of plants [25, 26].

3.3. Quality Analysis of Whole-Genome Resequencing Data

A total of 22 Citrus reticulata “Chachi” sample genomes were sequenced, which generated 158 Gb raw data. Base coverage depth distribution curve and coverage distribution curve indicated that the coverage depth of the bases on the genome was evenly distributed. The statistical results of insert fragment distribution with a single peak show that insert fragment distribution fits the normal distribution and the construction of DNA libraries was reliable. The chromosome coverage depth map showed that the genome was evenly covered, indicating good randomness of sequencing.

Summary of clean sequencing data results about 22 Citrus reticulata “Chachi” samples is given in Table 5. The size of reference genome Citrus reticulata is 344.27 Mb (assembly level: scaffold). In this work, the average coverage depth was 16X, and the value of Q₃₀ reached 88.94%. The average mapped ratio and genome coverage of all the samples were 98.97% and 93.55%, respectively. The average GC content of Citrus reticulata “Chachi” was 38.68% in line with reference genome Citrus reticulata.

3.4. SNP and InDel Calling of Citrus reticulata “Chachi”

In this study, the variant site detection of Citrus reticulata “Chachi” for the national geographical indicated product CRP was firstly analyzed by SNP and InDel calling from whole-genome resequencing data (Table 6). Except for the high genetic stability of 4 barcodes (ITS2, rbcl, matK, and psbA-trnH), 22 Citrus reticulata “Chachi” samples showed its genetic diversity between different propagation methods, different tree ages, different varieties, and different rootstocks as well.

A total of 2652697 SNP sites were excavated between 22 Citrus reticulata “Chachi” samples, among which 1741507 SNP sites were transition (T_i), 902182 SNP sites were transversion (T_v), and 9008 SNP sites were transition or transversion. These SNP sites were with a T_i/T_v ratio of 1.93, which is in line with general rules of base mutation in natural organisms [27]. In the course of evolution about Citrus reticulata “Chachi,” transition happens much more frequently than transversion, which means that evolution of Citrus reticulata “Chachi” tends to accept the substitution between purines and purines or the substitution between pyrimidines and purines, the substitution between purines and pyrimidines causes bad things to happen, and that substitution has mostly been eliminated by evolution.

In addition, InDel sites, as codominant molecular markers, are widely distributed in the genome with high density, which are suitable for genome-wide molecular marker exploration. A total of 533906 InDel sites were detected between 22 Citrus reticulata “Chachi” samples, among which 275380 InDel sites were insertion, 241768 InDel sites were deletion, and 9008 InDel sites were insertion or deletion. Unlike SNP sites, insertions and deletions of InDel sites are equally likely to occur.