A specimen of Chaeturichthys stigmatias was collected from the coast of Qingdao, China in November 2018, and sent for high-throughput sequencing. Dozens of C. stigmatias were collected from Zhoushan (August 2019), Qingdao (December 2019), Yantai (December 2019), and Weihai (October 2020), and 24 of them were used for polymorphism detection and genetic diversity analysis in this study. The samples were quickly dissected, and part of the muscle tissues on the caudal peduncle were collected and preserved in 95% alcohol in ice box, and then stored in –80°C for DNA extraction. The traditional phenol–chloroform method was used to extract genomic DNA. The total DNA was treated with RNase, and the DNA with high purity and without RNA contamination was obtained for the detection of SSR primers polymorphism. The extracted DNA was measured using a Nanodrop 2000 (Thermo Scientific, USA) and a Qubit 2.0 (Invitrogen, USA) bioanalyzer system.

After DNA quality inspection, library construction and sequencing were conducted. The steps of RAD library construction (Baird et al. 2008) were as follows: (1) Genomic DNA from the sample was digested at specific sites with a restriction enzyme, and the adapter P1 was ligated to the digested product. The P1 adapter contains forward amplification and Illumina sequencing primer sites, and an individual-specific nucleotide barcode; (2) The adapter-ligated fragments were then pooled, randomly sheared, and size-selected; (3) DNA was then ligated to a second adapter (P2), a Y adapter containing the reverse complement of the reverse amplification primer site, which ensures that lacking P1 adapter-ligated genomic fragments could not be amplified; (4) RAD tags with P1 adapter will be selected and amplified, and the 300–700 bp sequences were recovered. Agilent 2100 and Q-PCR were used to detect the size of library fragments and library quantification to determine whether the library meets the sequencing standard and then sequenced using the Illumina HiSeq2000 platform following the manufacturer’s protocol. To obtain the clean reads, the reads with more than 10% N bases or low-quality bases ≤5, adapter sequences, and duplicated sequences were discarded. The clean reads were used for subsequent analysis.

The software of detecting sequence repeats is “SSR search”, which is a Perl program written by Novogene (Beijing). The detection software is divided into three modules. The first module is used to detect all simple repeats of DNA sequence, the second module is to filter the results of the first module to remove the simple repeats that are too close. The detection criteria were as follows: the length of the SSR repeat unit ranges from 2 to 6 bp; the minimum length of the SSR sequence was 12 bp; the length of the upstream and downstream sequences of the SSR was 100 bp, and 12 bp was the minimum distance between two SSR sequences. The third module is to use Primer3 (a software that designs primers under Linux or UNIX systems) to design primers (Rozen and Skaletsky 2000). The detected microsatellite primer sequences were further screened, and the screening criteria were as follows: the SSR units were repeated more than 6 times; the length of the SSR units ranges from 3 to 10 bp; the expected length of the PCR product was between 130 and 300 bp; the sequence of four consecutive bases was excluded. The selected primers were suitable for synthesizing SSR primers to verify primer polymorphism.

Microsatellites were verified through PCR and electrophoresis. Each 25 μL PCR amplification system contained the following reagents: 17.25 μL ultrapure water, 2.5 μL 10 × PCR buffer, 2 μL dNTPs, 1 μL each primer (5 μmol · L^–1), 0.25 μL Taq polymerase, and 1 μL template DNA. The PCR reactions ran for 5 min at 94°C, followed by 38 cycles of 45 s at 94°C, 45 s at the annealing temperature (Table 1), and 45 s at 72°C in a thermal cycler. Cycling was followed by a final extension step at 72°C for 10 min. The PCR product was incubated at 4°C. The amplified PCR product was electrophoresed on an 8% non-denaturing polyacrylamide gel at 14 W for 3–4 h, and it was shown by silver staining (Lin et al. 2015). The allele size was identified according to the 20 bp DNA ladder.

After statistical analysis, the results were input into Genepop 4.0 (Rousset 2008). The parameters of SSR primers were calculated, including the mean value of effective allele number (N_a), polymorphism information content (PIC), observed heterozygosity (H_o) and expected heterozygosity (H_e), and Hardy–Weinberg equilibrium was also performed.

We have read the policies relating to animal experiments and confirmed this study complied. All procedures performed in this study were approved by the Institutional Animal Care and Use Committee of the Ocean University of China.

A total of 4.682 Gb high-quality data was obtained, and the Q20 and Q30 values were 97.31% and 92.52%, respectively. The RAD-Tag capture rate was 98.03%, and the GC content was 39.43%. Genomic GC content had a significant effect on the randomness of second-generation genome sequencing. Too high (>65%) or too low (<25%) GC content will lead to sequencing bias and seriously affect the results of genomic analysis. The GC content of Chaeturichthys stigmatias was normal, and the sequencing quality was qualified, indicating that the sequencing of the database was successful (Zerbino and Birney 2008).

The sequences were clustered and assembled. The total contig base was 113 171 723 bp, and the total contig number was 337 800. The mean value of contig length of the assembly sequences was 335 bp, and N50 length was 393 bp. The GC content of the assembly result was 39.04%, which was consistent with the GC content of the sequencing clean data, indicating that the assembly result was true and reliable (Wang et al. 2017; Gao et al. 2018). Subsequently, the variation detection was carried out on the assembly results. The number of heterozygous SNPs in the detected SNPs was 142 307, and the heterozygous rate was 82.47%. The high heterozygous SNP and the low homozygous SNP values also indicated the reliability of the assembly results.

Based on the RAD-seq, the total number of identified microsatellites was 5829. Among them, there were 5631 microsatellite loci containing primer fragments (Table 2). The trinucleotide repeats were dominant (57.56%), followed by dinucleotide repeats (23.30%), tetranucleotide repeats (11.79%), pentanucleotide repeats (4.14%), and hexanucleotide repeats (3.21%).

The previous studies showed that the dominant repeating unit was discrepant. Some fish species were dinucleotide, such as Megalobrama amblycephala Yih, 1955 and Larimichthys crocea (Richardson, 1846) (see Wang et al. 2012; Zeng et al. 2013; Li unpublished²), while some fish were trinucleotide, such as Acanthogobius hasta (Temminck et Schlegel, 1845) and mollusk, Ruditapes philippinarum (see Yan et al. 2015; Song et al. 2019). The previous studies have suggested that enzymes and other proteins involved in various aspects of DNA processing and chromatin remodeling may be responsible for the taxonomic specificity of microsatellite abundance. This was manifested in that not only the repetitiveness of the genome varies, but also the dominant microsatellite types are different. This might indicate that SSRs play an important role in genome evolution, and the process responsible for the generation and fixation of SSR has also changed during evolution (Toth et al. 2000). In this study, trinucleotide repeats have absolute quantitative advantages, and the number of dinucleotide repeats was less than half. We speculate that a genetic mutation might occur during the evolution of Chaeturichthys stigmatias. Further comparative investigations including more species are needed to clarify this point.

The distribution and frequency of microsatellite motifs were presented in Fig. 1. The AT repeat motif (300) was the most frequent among all 11 types of dinucleotide repeat, whereas GC was the least frequent, with only one microsatellite locus. The AAT repeat motif (396) was the most frequent among all 60 types of a trinucleotide repeat. The AAAT repeat motif (55) was the most frequent among all 104 types of tetranucleotide repeat. The AATTG repeat motif (56) was the most frequent pentanucleotide repeat, and the ATTCTG (35) was the most frequent hexanucleotide repeat. Because the repeat types of trinucleotide, tetranucleotide, pentanucleotide, and hexanucleotide were too dispersed, only the top 30 types of loci were selected for illustration in order to show the results more clearly. All detailed data was provided in Suppl. material 1: Appendix 1 (online resource).

Figure 1. 

The distribution and frequency of microsatellite motifs of Chaeturichthys stigmatias from China. (A) Frequency of different dinucleotide microsatellite motifs; (B) Frequency of different trinucleotide microsatellite motifs; (C) Frequency of different tetranucleotide microsatellite motifs; (D) Frequency of different pentabase microsatellite motifs; (E) Frequency of different hexanucleotide microsatellite motifs.

In terms of the frequency of repeating units, there were only four distinct types of repeats detected in pentanucleotide and hexanucleotide, and all of them were predominant at a frequency of a 4-fold repeat. Seven types were identified and 4-fold repeat was predominant in all tetranucleotide repeats. The types of repetition frequency detected in dinucleotide and trinucleotide were not less than 10 types, and 5-fold repeat and 6-fold repeat were the main components in dinucleotide and trinucleotide respectively (Figs 1, 2).

In this study, the frequency distribution of the repetition units of dinucleotide, trinucleotide, tetranucleotide, pentanucleotide, and hexanucleotide microsatellite were mainly 4–10 times, 4–7 times, 4–5 times, 4 times, and 4 times, respectively (Fig. 2). This result showed that the frequency of tandem repeats decreased exponentially with the increase of repetition unit length, and this was consistent with the conclusion proposed by Chen et al. (2010) that the number of repeating units was negatively correlated with the length of repeating units. According to previous studies, slipped-strand mispairing is a major mechanism for DNA sequence evolution (Levinson and Gutman 1987), and the results of this study can be explained as microsatellites with a large number of repetitions may be more unstable due to the increase of sliding possibility (Ellegren 2004). It was generally believed that there was a certain positive correlation between the variation frequency of SSR sites and the number of repetition units (Schlötterer 2000). Katti et al. (2001) also reported that the mutation rate increases gradually with the increase of the length of repeating units in eukaryotes.

Figure 2. 

Frequency of repeating units of different types of microsatellites of Chaeturichthys stigmatias from China.

A gradient PCR experiment was performed on the synthesized 148 pairs of primers, and the optimal temperature of each pair of primers was screened. The results showed that a total of 97 pairs of primers were successfully amplified. Then after the PCR product was subjected to polyacrylamide gel electrophoresis experiments, a total of 30 primers with polymorphism were screened out. A total of 312 alleles were detected for 24 individuals at 30 polymorphic loci, and the number of alleles per locus ranged from 6 to 14, with the mean value of effective alleles was 10.4. The mean value of expected heterozygosity was 0.870, the observed heterozygosity was 0.349, and the mean value of polymorphic information content was 0.836 (Table 1). All the polymorphic sites deviated significantly from Hardy–Weinberg equilibrium (P < 0.05). The PIC was between 0.749 and 0.905, and all loci showed high polymorphism (PIC > 0.5) (Botstein et al. 1980).

In this study, 30 primers with polymorphism were screened out as dinucleotide and trinucleotide repeats, without tetranucleotide, pentanucleotide, and hexanucleotide repeats. Kong et al. (2019) observed that compared with trinucleotide and tetranucleotide repeats, dinucleotide repeats had a higher screening efficiency and polymorphism (Kong et al. 2019). However, in recent years, it has also been found that the trinucleotide and tetranucleotide repeats have higher screening efficiency and polymorphism than dinucleotide repeats in Ctenopharyngodon idella, Hypophthalmichthys molitrix (Valenciennes, 1844), and Cyprinus carpio Linnaeus, 1758 (see Fang et al. 2018). The higher repetition unit length has the disadvantages of lower repeats, lower sequence richness, and lower mutation rate. The better trinucleotide and tetranucleotide repeats polymorphisms obtained in other experiments may be related to the genome doubling in the long-term evolution of this species (Lu et al. 2009; Fang et al. 2018). The different results may be related to the specificity of species, the randomness of the number and type of primers selected in the experiment, and the number of samples. The above-mentioned results indicated that the SSR repeats which had higher screening efficiency and polymorphism may be species-dependent, and the most probable SSRs were dinucleotide and trinucleotide repeats. In terms of polymorphism, the mean PIC values of dinucleotide and trinucleotide repeats microsatellite primers screened in this study were 0.836 and 0.835, respectively, with little difference in polymorphism. Therefore, differences in screening efficiency and polymorphism may be caused by species differences or other factors (Kong et al. 2019).

The higher heterozygote ratio reflects the stability of the genetic structure of the population. We found that the observed heterozygosity (H_o) of 30 polymorphic sites was lower than the expected heterozygosity (H_e), showing a relative lack of heterozygosity. It was generally believed that the loss of heterozygosity was caused by geographical isolation, decreased gene exchange between populations, and increased inbreeding (Zhao et al. 2009). The samples used in this study were collected from the Yellow Sea, the Bohai Sea, and the East China Sea. The low heterozygosity of Chaeturichthys stigmatias may be due to geographical isolation and excessive intraspecific hybridization. It was commonly accepted that the expected heterozygosity (H_e) was a more accurate reflection of the genetic diversity of a population than the observed heterozygosity (H_o) (Nei 1978). Therefore, the mean value of observed heterozygosity of 0.870 in this study showed a high population diversity.

At the same time, according to Hardy–Weinberg equilibrium analysis, all the 30 microsatellite loci discussed in this study showed significant imbalance, which was a common phenomenon in fish populations, such as Siniperca scherzeri Steindachner, 1892 and Lutjanus peru (Nichols et Murphy, 1922) (see Dou et al. 2015; Paz-García et al. 2017). This result also confirmed that these populations did not mate randomly, and non-random sampling was also the reason for the deviation of Hardy–Weinberg equilibrium. It was worth noting that inbreeding, subgroup structure, genetic drift, overfishing, Wallund effect, and ineffective alleles should also be considered (Bergh and Getz 1989; Lu et al. 2017; Song et al. 2018). The above results indicated that the microsatellite markers identified in this study have high polymorphisms and can be used as effective molecular markers to analyze the genetic diversity and phylogenetic relations among C. stigmatias.