Genetic Structure and Diversity of Hatchery and Wild Populations of Yellow Catfish Tachysurus fulvidraco (Siluriformes: Bagridae) from Korea

Yellow catfish Tachysurus fulvidraco is an important commercial fish species in South Korea. However, due to their current declines in its distribution area and population size, it is being released from hatchery populations into wild populations. Hatchery populations also produced from wild broodstocks are used for its captive breeding. We reported 15 new microsatellite DNA markers of T. fulvidraco to identify the genetic diversity and structure of its hatchery and wild populations, providing baseline data for useful resource development strategies. The observed heterozygosity of the hatchery populations ranged from 0.816 to 0.873, and that of the wild populations ranged from 0.771 to 0.840. Their inbreeding coefficient ranged from −0.078 to 0.024. All populations experienced a bottleneck (p < 0.05), with effective population sizes ranging from 21 to infinity. Their gene structure was divided into two groups with STRUCTURE results of K = 2. It was confirmed that each hatchery population originated from a different wild population. This study provides genetic information necessary for the future development and conservation of fishery resources for T. fulvidraco.


Introduction
Aquaculture is increasingly becoming a crucial source of protein, with freshwater aquaculture also serving as an important resource [1].Freshwater ecosystems are critical to maintaining economic significance [2].However, it is the most vulnerable ecosystem due to pressure from human activities [2].Yellow catfish Tachysurus fulvidraco (Richardson, 1846) is one of the most important farmed freshwater fish species in China and South Korea, with China recording production of 565,477 tons and 587,822 tons in 2021 [3] and 2022 [4], respectively.According to the Fisheries Resources Agency of South Korea, domestic releases of T. fulvidraco from 2017 to 2022 amounted to 34,462,711 tons (https://www.fira.or.kr/fira/ main.jsp:accessed on 1 May 2022).
In South Korea, fishery for T. fulvidraco is very popular today (https://www.fira.or.kr/ fira/main.jsp:accessed on 1 May 2022).Commercial fishing of T. fulvidraco has continued in wild populations such as the Hantan River, Seomjin River, and Geum River in the Korean Peninsula, but its catches are decreasing due to ecosystem changes.Therefore, to prevent this decline and sustain the catches, the local government has continued to release T. fulvidraco by purchasing hatchery populations from private hatcheries after captive breeding of broodstock collected from the wild.In this process, its genetic diversity is an important consideration when collecting broodstocks from wild populations [5].Release of hatchery-reared individuals can increase fishery productivity, accelerate the recovery of depleted stocks, and ensure stock survival [5].High genetic diversity affects the hatchability Raw reads were obtained from each individual through whole-genome sequencing (Pf01: 6,231,011, Pf02: 5,705,048, Pf03: 5,123,581, and Pf04: 6,235,796; Table S1).The assembly and screening results of the msDNA markers are shown in Table S1.A total of 100 markers were screened, and the markers that were not amplified and did not show polymorphisms in each population were excluded (Table S2).Multiplex PCR from the final 15 msDNA loci consisted of two sets (Table S3).
The number of alleles (N A ) ranged from 9 to 19, and the polymorphism information content (P IC ) was >0.5, indicating appropriate marker selections (Table S4).Genotyping using the MICRO-CHECKER software (ver.2.2.3) showed no evidence of scoring errors or null alleles (Table S4).
The 15 msDNA loci were analyzed for genetic diversity indices in the eleven populations (Table 1).The values described below are the average values of population genetic diversity information based on them.
N A ranged from 7.60 to 11.00, the observed heterozygosity (H O ) ranged from 0.771 to 0.873, and the expected heterozygosity (H E ) from 0.735 to 0.844.We found that the KS and HT populations followed the Hardy-Weinberg equilibrium value (P HWE ), but the remaining populations deviated from the P HWE .Inbreeding was observed only in SY, YD, and ND populations, but the inbreeding coefficient (F IS ) was not significant.The ND population for resource establishment purposes showed the least genetic diversity with H O = 0.771.The hatchery populations ASS and CT showed a high level of genetic diversity with H O = 0.816 and H O = 0.873, respectively.

Bottleneck Analysis
Using the infinite allele mutation (IAM) model, we identified significant bottlenecks in all populations (p < 0.05).Using the two-phase model (TPM), we identified bottlenecks in the ASS, KS, YD, and HT populations (Table 2).The KK population showed recent mode shifts, showing evidence of a bottleneck.Among the eleven populations, the effective population size was 21-infinity.The effective population size of the ND population was infinity, and that of the CT population was the smallest at 21 (Table 2).The effective population size of the ND population for resource creation purposes was found to be infinity (Table 2).The AS, NW, PT, CT, and HT populations were smaller than the minimum effective population size of 100 required to prevent inbreeding depression.

Population Structure and Genetic Differentiation Analyses
The pairwise genetic differentiation (F ST ) values of the msDNA datasets were all significant (Table 3), with the highest F ST value between the AS and KS populations (F ST = 0.156).The ASS, KK, KS, and YD populations exhibited a range of low F ST values among populations (F ST = 0.012-0.030).The AS, CT, NW, PT, HT, and SY populations also each exhibited a range of low F ST values among populations (F ST = 0.020-0.082).Bayesian clustering analysis maximized the delta population constant (∆K) value for population structure at K = 2 (Figure 1).At K = 2, the first group included the ASS, KK, KS, and YD populations.The second included the AS, CT, NW, PT, HT, ND, and SY populations.Although it did not inhabit the Nakdong River, the artificially produced ND population appeared to belong to the AS, CT, NW, PT, HT, ND, and SY populations.
exhibited a range of low FST values among populations (FST = 0.020-0.082).Bayesian clustering analysis maximized the delta population constant (ΔK) value for population structure at K = 2 (Figure 1).At K = 2, the first group included the ASS, KK, KS, and YD populations.The second included the AS, CT, NW, PT, HT, ND, and SY populations.Although it did not inhabit the Nakdong River, the artificially produced ND population appeared to belong to the AS, CT, NW, PT, HT, ND, and SY populations.
Scatterplots of the discriminant analysis of principal components (DAPC) showed that the first populations were the ASS, KK, KS, and YD populations and the second populations were the AS, CT, NW, PT, HT, ND, and SY populations, which was the same result as that of STRUCTURE (Figure 2).Scatterplots of the discriminant analysis of principal components (DAPC) showed that the first populations were the ASS, KK, KS, and YD populations and the second populations were the AS, CT, NW, PT, HT, ND, and SY populations, which was the same result as that of STRUCTURE (Figure 2).An analysis of molecular variance (AMOVA) for T. fulvidraco was performed on the eleven populations to determine the genetic structure (Table 4).The AMOVA, based on msDNA markers, showed 3.69% variation among populations within the groups, 7.10% variation among groups, and 89.21% variation within the populations.

Genetic Diversity
Genetic diversity actively copes with changing environments and plays an important role in the persistence of species [13].Genetic diversity at the species level in msDNA markers is generally indicated by HO [14].Rita rita, Clarias gariepinus, and Pangasianodon hypophthalmus are commercial fish species belonging to the same family of Siluriformes as T. fulvidraco [15][16][17].The HO range of R. rita was between 0.57 and 0.76, that of C. gariepinus was 0.43-0.61,and that of P. hypophthalmus was 0.42-0.50,suggesting that T. fulvidraco (HO = 0.77-0.87) is a species with higher genetic diversity [15][16][17].In fact, An analysis of molecular variance (AMOVA) for T. fulvidraco was performed on the eleven populations to determine the genetic structure (Table 4).The AMOVA, based on msDNA markers, showed 3.69% variation among populations within the groups, 7.10% variation among groups, and 89.21% variation within the populations.

Genetic Diversity
Genetic diversity actively copes with changing environments and plays an important role in the persistence of species [13].Genetic diversity at the species level in msDNA markers is generally indicated by H O [14].Rita rita, Clarias gariepinus, and Pangasianodon hypophthalmus are commercial fish species belonging to the same family of Siluriformes as T. fulvidraco [15][16][17].The H O range of R. rita was between 0.57 and 0.76, that of C. gariepinus was 0.43-0.61,and that of P. hypophthalmus was 0.42-0.50,suggesting that T. fulvidraco (H O = 0.77-0.87) is a species with higher genetic diversity [15][16][17].In fact, channel catfish Ictalurus punctatus, which is actively used commercially, had a H O range of 0.72-0.77,which was similar to that of T. fulvidraco [18].Therefore, T. fulvidraco has higher genetic diversity at the species level compared to other siluriform species, which may make it suitable as a commercial species.
In this study, hatchery populations (ASS and CT) showed H O values ranging from 0.816 to 0.873, confirming their relatively higher genetic diversity than wild populations (YD, AS, ND, and SY).We observed that the genetic diversity of the wild populations ranged from 0.771 to 0.838, which was lower than that of the hatchery population (CT).The ND population is a single population released for the purpose of resource creation into a river (Nakdong River) where T. fulvidraco does not exist.The genetic diversity of the ND population was found to be the lowest at 0.771.In general, hatchery populations are reported to have low genetic diversity [15].This is explained by the low number of broodstock used in hatcheries and the loss of genetic variation due to inbreeding [19].However, in this study, the genetic diversity of two hatchery populations, ASS and CT, was relatively high.This may be because the Fisheries Resources Agency of South Korea recommends using diverse wild populations [20].The low genetic diversity of the ND population intended for resource development may be of concern from a resource development perspective.However, it is believed that practical judgments can only be made if genetic diversity, as with the size of the effective population, is taken into consideration [13].
In the hatchery and wild populations, ND, SY, and YD had positive F IS values, suggesting that inbreeding had occurred.However, since it was not significant, it was not judged to be an important implication.Excluding the ND, SY, and YD populations, eight populations (ASS, KK, KS, AS, NW, PT, CT, and HT) showed negative F IS values, suggesting an influx of external populations, although this was not significant (p > 0.05).
Estimating the expected excess of heterozygosity in P IAM is suitable for estimating recent bottlenecks [21].The most recent population bottleneck in P IAM was observed in all populations (p < 0.05).Habitat destruction reduces the population size and creates bottlenecks [13].Wild populations have recently experienced population size declines, and active efforts are needed to conserve the resource.
N e is important for species adaptation to changing environments and maintaining evolutionary potential [13].Small population sizes can accelerate the risk of local extinction due to genetic drift and inbreeding effects [22].Additionally, if populations are too small, managers should strive to preserve interconnected populations that are at least large enough to meet this minimum [23].In this study, the four wild populations (AS, NW, PT, and HT) had a significantly low effective population size (N e < 100).Frankham et al. [24] suggested that the N e should be greater than 100 to avoid inbreeding depression in shortterm generations.However, although the wild populations (AS, NW, PT, and HT) had N e values below 100, there was an influx of external populations.For this reason, it is assumed that it was due to the release of fish from the captive breeding program of the local government.However, if this resource release is not maintained, they are likely to suffer from inbreeding depression in the short term.The effective population size of the ASS population, a hatchery population, was 127, which is higher than 100, but it is believed that the introduction of wild populations is necessary for continued resource expansion.The CT population from hatchery was found to have an effective population size of 21, which was relatively very low compared to the ASS population.It is considered important to improve the effective population size by introducing wild broodstocks from outside.The effective population size of the ND population for resource creation purposes appeared to be infinite, and it was confirmed to have a very large population.This means that the wild population was released only for the purpose of resource generation, and since it is presumed that the offspring of several broodstock populations were released, the effective population size is estimated to be very large.However, this may be due to the small number of samples; therefore, it was judged that a more detailed analysis would be necessary by increasing the sample size in the future.

Genetic Structure
In this study, we analyzed the genetic structure of eleven populations of T. fulvidraco that inhabit the Hantan River, Seomjin River, and Geum River, as well as the Nakdong River population formed for the purpose of resource creation (ND), and identified the origin of hatchery broodstock.The populations showed significant genetic differentiation in the msDNA dataset.Here, the genetic structure was divided into two groups (ASS, KK, KS, and YD vs. AS, CT, NW, PT, HT, ND, and SY).However, the results of this study are different from the pattern of genetic differences among river basins, mainly due to geographical differences.In general, freshwater fish in the Korean Peninsula show genetic differences depending on the geographical water system, and there is little genetic difference within the same river basin [25].Interestingly, the PT, AS, and YD populations were geographically close to each other, so no differences in genetic structure were expected among them.However, it was confirmed that the YD population showed the same genetic structure as the ASS, KS, and KK populations.The first hypothesis is that the YD population may not have existed originally but became a stock-building population through release from a hatchery population.The second hypothesis is that this population was originally present in the area and that the released individuals may have been captured due to continued release from the hatchery.The first hypothesis is considered to have low probability because T. fulvidraco was continuously caught even before release [26].The second hypothesis implies that since continuous fish offspring release was conducted in local government from 2017 to 2022, the fish seed population is likely to be captured.Although more samples are needed to investigate the YD population before release, it is necessary to secure more samples through future research and to perform sampling by season.
It was supported through STRUCTURE that the population for resource creation purposes was not genetically different from the AS, CT, NW, PT, HT, and SY populations.DAPC is a non-model method and shows higher reliability compared to STRUCTURE [27].The DAPC results showed that the genetic structure of the eleven populations was divided into two groups (ASS, KK, KS, and YD vs. AS, CT, NW, PT, HT, ND, and SY).This shows a clear division, similar to the STRUCTURE results.Additionally, the F ST values likewise supported significant genetic differences between the two groups.The hatchery populations ASS and CT were shown to have different genetic structures.Therefore, these significant differences in genetic structure can be important basic data for planning when creating resources [28].
The ND population is a group that did not originally exist in the water system but was created for the purpose of resource creation.Looking at the genetic structure results, it is suggested that the group was created from broodstocks of the AS, NW, CT, PT, HT, and SY populations.
According to the local government, the hatchery populations, ASS and CT, are believed to be produced from broodstocks of different river basins.This may explain the differences in genetic structure.It is presumed that the ASS population was mainly produced from broodstocks of the KK, KS, and YD populations, and the CT population is presumed to have been produced from broodstocks of the AS, NW, PT, HT, and SY populations.The use of broodstocks in hatchery populations is recognized as important.This requires consideration of the importance of genetic differences, genetic diversity, and effective population size upon release from the hatchery.Therefore, it is recommended that fisheries use an offspring release program (for local government), with an offspring population using KK, KS, SY, and YD populations with large effective population sizes as parents for conservation of the genetic diversity of the natural population in South Korea.In addition, we suggest using ND and SY populations for commercial aquaculture production.

Sampling and DNA Extraction
Tachysurus fulvidraco is used as an inland fisheries resource for commercial fishing and aquaculture form in the Republic of Korea; therefore, animal ethics approval was waived.As wild populations of T. fulvidraco were released from hatchery populations, we sampled the hatchery populations (ASS and CT).We also sampled wild populations using stocking records (KK, KS, YD, AS, NW, PT, HT, and SY) and sampled the ND population (for stock composition purposes).A total of eleven populations of T. fulvidraco were sampled (location and latitude-longitude details are provided in Figure 3 and Table 5).Each sampling was conducted between 2021 and 2023, and fin tissue from the fish was collected and soaked in 99% ethanol.Genomic DNA (gDNA) was extracted using the DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions.The extracted gDNA was stored at −20 • C after dilution to 50 ng/µL to amplify msDNA loci.
Tachysurus fulvidraco is used as an inland fisheries resource for commercial fishin and aquaculture form in the Republic of Korea; therefore, animal ethics approval w waived.As wild populations of T. fulvidraco were released from hatchery populations, w sampled the hatchery populations (ASS and CT).We also sampled wild populations u ing stocking records (KK, KS, YD, AS, NW, PT, HT, and SY) and sampled the ND pop lation (for stock composition purposes).A total of eleven populations of T. fulvidraco we sampled (location and latitude-longitude details are provided in Figure 3 and Table 5 Each sampling was conducted between 2021 and 2023, and fin tissue from the fish w collected and soaked in 99% ethanol.Genomic DNA (gDNA) was extracted using th DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany) according to the manufactu er's instructions.The extracted gDNA was stored at −20 °C after dilution to 50 ng/µL amplify msDNA loci.  5.   5.

Whole-Genome Sequencing and msDNA Marker Screening
Whole-genome sequencing was performed using a total of four individuals (Pfu01-Pfu04).It was performed using an Illumina MiSeq ® System (Illumina, San Diego, CA, USA) and included 300 bp paired-end library construction.For screening msDNA markers, contigs assembled from Geneious Prime ver.2020.02(Biomatters Ltd., Auckland, New Zealand) were selected for di-, tri-, tetra-, penta-, and hexa-sequences with more than five repetitions using the MISA tool [29].Using Primer3 [30], we set the length of the appropriate primer to 17-25 bp, the size of the amplification product to 150-400 bp, and the melting temperature (Tm) value to 50-60 • C.

msDNA Genotyping
One hundred msDNA loci were randomly selected, and PCR was performed using an Applied Biosystems™ ProFlex™ PCR System (Thermo Fisher Scientific, Foster City, CA, USA).PCR was performed using 20 ng gDNA; 0.5 units AccuPower ® PCR PreMix (Bioneer, Daejeon, Republic of Korea); 0.5 µM forward primer labeled fluorescent dyes FAM, HEX, TAMRA, and ATTO565; and 0.5 µM reverse primer.PCR condition was as follows: predenaturation at 95 • C for 5 min, denaturation at 95 • C for 20 s, annealing at 55 • C for 20 s, and extension at 72 • C for 20 s.After 35 repetitions, the final extension was performed at 72 • C for 10 min, and the temperature was held at 8 • C. Each amplified PCR product was electrophoresed on 2% agarose gel to confirm the presence or absence and size of the amplified fragment.The PCR fragments were prepared by mixing a GeneScan™ 500 ROX size standard ladder (Thermo Fisher Scientific) and HiDi™ formamide and performing denaturation at 95 • C for 2 min, followed by termination at 4 • C. The allele sizes were determined using an Applied Biosystems™ ABI 3730xl DNA Analyzer (Thermo Fisher Scientific).Genotyping was performed using GeneMapper ver. 5 software [31].

Population Genetic Structure Analysis
ARLEQUIN software (ver.3.05) [34] was used to analyze the differences in genetics between groups, as well as AMOVA.STRUCTURE software (ver.2.3) [40] was used to perform genetic structure clustering analysis based on the Bayesian method model.To estimate the most suitable population, we set K to 1-10, and a suitable admixture model was applied to the mixture of water systems.The burn-in period was repeated 10 times with 10,000 iterations, and Markov chain Monte Carlo simulation was used with 100,000 iterations.To estimate a population-appropriate constant, we analyzed a study by the cluster results corresponding to K using STRUCTURE SELECTOR [41].A discriminant analysis of principal components (DAPC) of the microsatellite dataset was performed on the population using the R package ADEGENET (ver.2.1.3)[42], a non-model-based genetic clustering method.

Conclusions
Tachysurus fulvidraco is an important commercial fish species in the Republic of Korea.They are released from hatchery populations into wild populations through artificial propagation.The hatchery populations ASS and CT use broodstock collected from wild populations.In this study, 15 new msDNA markers of T. fulvidraco were found.We sought to identify the genetic diversity and structure of hatchery and wild populations and provide baseline data for useful resource development strategies.The observed heterozygosity in hatchery populations ranged from 0.816 to 0.873, and the observed heterozygosity in wild populations ranged from 0.771 to 0.840.The hatchery population CT had the highest genetic diversity but the lowest effective population size of 21.Supplementation of wild mothers is thus required to supplement the population size.The genetic rescue was divided into two groups with a STRUCTURE result K = 2.The first group included the ASS, KK, KS, and YD populations, and the second group included the CT, AS, NW, PT, HT, and SY populations.Therefore, it was confirmed that the ASS and CT populations originated from different populations.This study provides genetic information necessary for the future development and conservation of fishery resources for T. fulvidraco.

Supplementary Materials:
The following are available online at: https://www.mdpi.com/article/10.3390/ijms25073923/s1.Funding: This research was conducted with support from the National Institute of Fisheries Science, "Research on environments and fisheries resources in inland fisheries resources protection area (R2024011)".

Institutional Review Board Statement:
The study was conducted in accordance with the National Institute of Fisheries Science guidelines and regulations.In addition, all experiments were performed in accordance with ARRIVE-related guidelines and the regulations of the National Institute of Fisheries Science.Population samples were collected according to the guidelines of the National Institute of Fisheries Science.

Figure 1 .
Figure 1.Population genetic structure of Tachysurus fulvidraco (K = 2).The ASS and CT populations are hatchery populations and the remaining populations are wild populations.The appropriate

Figure 1 .
Figure 1.Population genetic structure of Tachysurus fulvidraco (K = 2).The ASS and CT populations are hatchery populations and the remaining populations are wild populations.The appropriate delta K information is represented for population constants.A single histogram represents the probability that an object and a particular color are assigned to a particular cluster.
Int. J. Mol.Sci.2024, 25, x FOR PEER REVIEW 5 of 12 delta K information is represented for population constants.A single histogram represents the probability that an object and a particular color are assigned to a particular cluster.

Figure 2 .
Figure 2. Scatterplots of discriminant analysis of principal components (DAPC) for Tachysurus fulvidraco.The colors shown on the plot are the population.Each color represents a different genetic cluster, and population abbreviations are provided for each cluster.The upper left graph represents the contribution of the eigenvalues of the selected principal components, and the upper right graph represents the variance explained by the eigenvalues of the two discriminant functions in the scatterplot.

Figure 2 .
Figure 2. Scatterplots of discriminant analysis of principal components (DAPC) for Tachysurus fulvidraco.The colors shown on the plot are the population.Each color represents a different genetic cluster, and population abbreviations are provided for each cluster.The upper left graph represents the contribution of the eigenvalues of the selected principal components, and the upper right graph represents the variance explained by the eigenvalues of the two discriminant functions in the scatterplot.

Figure 3 .
Figure 3. Sampling location of eleven populations of Tachysurus fulvidraco.Abbreviations for po ulations are given in Table5.

Figure 3 .
Figure 3. Sampling location of eleven populations of Tachysurus fulvidraco.Abbreviations for populations are given in Table5.

Table 1 .
Average values of genetic diversity information for populations based on 15 microsatellite DNA loci of Tachysurus fulvidraco.

Table 2 .
Summary of information regarding bottleneck signature and effective population size for Tachysurus fulvidraco populations.
N: number of samples; P IAM : p-value of bottleneck test using infinite allele mutation model; P TPM : p-value of bottleneck test using two-phase mutation model (10% variance and 90% proportions of SSM); P SMM : p-value of bottleneck test using stepwise mutation model; N e

Table 3 .
F ST among populations according to microsatellite DNA data of Tachysurus fulvidraco.

Table 3 .
FST among populations according to microsatellite DNA data of Tachysurus fulvidraco.

Table 4 .
Summary of information for analysis of molecular variance (AMOVA) for populations of Tachysurus fulvidraco.
(Two groups based on the DAPC data (ASS, KS, KK, and YD vs. CT, ND, SY, AS, NW, PT, and HT populations) d.f.: degrees of freedom; *** p < 0.001.FST is based on standard permutation across the full dataset.

Table 4 .
Summary of information for analysis of molecular variance (AMOVA) for populations of Tachysurus fulvidraco.Two groups based on the DAPC data (ASS, KS, KK, and YD vs. CT, ND, SY, AS, NW, PT, and HT populations) CT , F SC and F ST were based on standard permutation across the full dataset.

Table 5 .
Details of the sampling information of Tachysurus fulvidraco.