*3.2. Genetic Diversity and Population Structure*

The genetic diversity parameters were calculated according to the original SNPs, and the statistical results were obtained. The genome-wide nucleotide diversity (π) of the ZB population and MJ population was 0.02154 and 0.02478, respectively. The inbreeding coefficient (Fis) of the ZB population and MJ population was −0.18729 and 0.03256, respectively. The genetic differentiation (Fst) between the ZB and MJ subpopulations was 0.00255102. The expected heterozygosity (He) of individuals from ZB and MJ was 0.33585 and 0.22098, respectively. The observed heterozygosity (Ho) of individuals from the ZB population and MJ population was 0.46834 and 0.23103, respectively (Table 3).

**Table 3.** Comparison of population genetic parameters in the ZB population and MJ population.


The heatmap showed positive correlations between the individuals of *L. kasmira*. In general, the correlations between individuals within the ZB population were the strongest, followed by those between the ZB population and MJ population, and the correlations between individuals within the MJ population were the weakest (Figure 2). The POP revealed the genetic distances between the 30 individuals, clustering MLGs based on genetic distance. The distances between individuals from the ZB population and MJ population were within the range of 0.002~0.004, indicating relatively short distances, consistent with other analysis results. Each MLG is a node, and the distance between the nodes represents the genetic distance between individuals (Figure 3). The phylogenetic tree revealed no significant clustering of the 30 *L. kasmira* samples (Figure 4). The population structure of *L. kasmira* was analyzed by Admixture software. Assuming that the number of groups (K value) is 1–10, the cross-validation error is maximal whenK=7 (Figure 5a,b). Therefore, the 30 samples of *L. kasmira* collected in this study were not from the same population. The genetic backgrounds of the ZB and MJ populations were complex, and there was admixture between the populations, which might be explained by the coexistence of artificial breeding and wild resources. The cross-validation results of the clustering showed that when K=1

(Figure 5a), the error rate of cross-validation was minimal, indicating that the optimal clustering number was 1. It is speculated that *L. kasmira* individuals in the ZB and MJ populations came from the same primitive ancestor (Figure 5b). The results of PCA showed that the 30 *L. kasmira* individuals were tightly clustered (Figure 6a,b) and the genetic heterogeneity between individuals was small, which was consistent with the phylogenetic tree results.

**Figure 2.** The distance matrix heatmap of 30 individuals of *L. kasmira* created using the R package poppr.

**Figure 3.** The minimum spanning network of 30 *L. kasmira* individuals.

**Figure 4.** Phylogenetic tree of 30 *L. kasmira* individuals based on SNP loci created using the neighborjoining method.

**Figure 5.** The group structure diagrams. (**a**) The cross-validation error for *L. kasmira* according to the admixture value K; (**b**) results of Bayesian cluster analysis of *L. kasmira* based on SNP loci using ADMIXTURE software for K = 1–10 clusters.

**Figure 6.** Population relationships of the collected *L. kasmira* individuals. (**a**) Principal component analysis (PCA) plot of 30 *L. kasmira* individuals based on all SNP loci between the ZB and MJ populations. A and B represent the MJ population and ZB population, respectively; (**b**) three−dimensional PCA clustering map of 30 *L. kasmira* individuals based on all SNP loci between the ZB and MJ populations. A and B represent the MJ population and ZB population, respectively.

#### **4. Discussion**

RAD-seq is a simplified genome sequencing technology based on whole-genome restriction sites developed on the basis of second-generation sequencing [11]. The number of RAD markers developed by this method is 10 times higher than that of traditional molecular marker development technology [11], with high accuracy and high data utilization [25]. RAD-seq shortens the marker development cycle compared to that of traditional markers and reduces experimental costs [26]. The technology can also screen for SNPs on a large scale in species without a reference genome [27]. SNPs are a new type of DNA molecular marker with broad application prospects [28]. SNPs occur at a high frequency and have a high marker density in most genomes. Compared with simple sequence repeat (SSR) markers, SNPs have higher genetic stability and are easier to automate [29]. RAD-seq technology can also be used to construct linkage maps [30]. For example, Palaiokostas et al. used RAD-seq technology to construct the first linkage map of Dicentrarchus labrax based on high-density SNPs [31].

In previous studies, genetic diversity was studied mainly by SSR analysis and D-loop sequence analysis [32], but these techniques have inherent difficulties in obtaining highquality DNA from wild individuals [33]. By referring to the study of 29 Rhinopithecus roxellana by Zhang et al. [34], we used RAD-seq to detect SNP markers in 30 *L. kasmira* individuals in this study and further analyzed the genetic diversity levels of different populations of this species. Based on a study of *L. kasmira* in the South China Sea by microsatellite analysis [32], further exploration was performed. A comparison of the results of this study with those of other studies shows that the degree of variation of *L. kasmira* in the ZB and MJ areas is low. This reflects its p ability to adapt to environmental change and respond to natural selection. Therefore, it is necessary to investigate the wild population as soon as possible and determine how to improve its genetic diversity. A larger sample size would produce more statistically robust results. More samples should be collected in different regions for more comprehensive genetic diversity analysis in the future.

In the phylogenetic tree based on RAD-seq data, the 30 samples of *L. kasmira* did not form obvious groups, and the population structure diagram also showed that they came from the same ancestor. Similar results were obtained with PCA, which were mutually confirmed with the results of the phylogenetic tree. The genome-wide nucleotide diversity (π) of the ZB population and MJ population was 0.02478 and 0.02154, respectively, indicating high nucleotide diversity. The expected heterozygosity (He) of individuals from ZB and MJ was 0.33585 and 0.22098, respectively. However, the observed heterozygosity (Ho) of individuals from the ZB population and MJ population was 0.46834 and 0.23103, respectively. The He of the two populations was lower than the Ho, indicating a low degree of variation for *L. kasmira* in the two areas and a certain degree of heterozygote deficiency. The genetic differentiation (Fst) between the MJ and ZB subpopulations was 0.00255102. The Fst value was consistent with the results of the heatmap and MSN. The distance between individuals was short, and differentiation was not obvious. Although π was high, He and Ho were low. The overall analysis showed that the population structure of *L. kasmira* was simple and had a low degree of variation, which might lead to poor adaptability to the environment. Furthermore, the genetic diversity of the MJ population was lower than that of the ZB population.

These results may be due to overfishing and global warming, which have reduced the genetic diversity of wild populations. Therefore, there is an urgent need to protect the population genetic resources of *L. kasmira*, and we can improve its genetic diversity through fishing restriction protection measures and artificial breeding of superior varieties. It is suggested that effective population size analysis and population history tracing should be performed for *L. kasmira*.

#### **5. Conclusions**

In this study, SNP data obtained by RAD-seq technology were used to analyze the genetic diversity in two populations of *L. kasmira*, providing a new approach for genetic diversity evaluation. The results showed that the genetic diversity of the two populations was relatively low at the genome level. To ensure adequate survival of the species, it is necessary to protect existing diversity and take measures to improve genetic diversity. To promote the healthy and sustainable development of germplasm resources of *L. kasmira*, much attention should be given to improving and maintaining its population genetic diversity, to implementing methods to increase gene flow, and to breeding excellent varieties by applying molecular breeding techniques.

Finally, our results indicate that the RAD-seq technique can detect SNP markers and apply them to the research of genetic diversity, with the benefits of a huge number of markers, low costs, and simple automation. This study is useful for the conservation of aquatic germplasm resources as well as for the development and utilization of high-quality resources.

**Author Contributions:** Conceptualization, D.Z. and S.J.; methodology, F.Z. and L.G.; software, L.G.; validation, F.Z. and J.Y.; formal analysis, F.Z.; investigation, K.Z.; resources, N.Z.; data curation, F.Z.; writing—original draft preparation, F.Z.; writing—review and editing, D.Z.; visualization, F.Z.; supervision, H.G. and B.L. (Bo Liu); project administration, B.L. (Baosuo Liu). All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Financial Fund of the Ministry of Agriculture and Rural Affairs, P. R. of China (NHYYSWZZZYKZX2020) and National Marine Genetic Resource Center, and China-ASEAN Maritime Cooperation Fund.

**Institutional Review Board Statement:** All applicable international, national, and institutional guidelines for the care were followed by the authors.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patients to publish this paper.

**Data Availability Statement:** All data generated or analyzed during this study are included in this published article.

**Conflicts of Interest:** The authors declare no conflict of interest.
