Introduction

The large yellow croaker (Larimichthys crocea) is one of the most economically important marine culture fish mainly cultivated in the coastal areas of Southeast China (Jian and Wu 2003). It also represents the largest yield of single fish species of marine net cage aquaculture of China with annual production of approximate 178,000 tons according to the Fishery Annual Statistics 2017 published by the Ministry of Agriculture, China. With the rapid development of aquaculture industry of large yellow croaker, many problems are emerging and hampering the sustainable development of aquaculture, of which, high mortality rate caused by cryptocaryonosis (white spot disease) is one of the most severe disease problems for large yellow croaker aquaculture in the past decade. Cryptocaryonosis is caused by the most common and destructive protozoan pathogen of marine fish, Cryptocaryon irritans, which infects fish skin and gills, leading to a devastating effect on cultivated large yellow croaker as well as other cultured marine fishes (Niu et al. 2013; Li et al. 2011). Unfortunately, there are few treatments for eradicating cryptocaryonosis effectively, especially for marine cage farm aquaculture. Selective breeding to improve anti-infection resistance in cultured large yellow croaker becomes promising alternatives for controlling cryptocaryonosis that severely affect croakers. With the rapid development of molecular genetics technology, genetic researches on resistant traits of aquatic animals have also made rapid progresses, such as genetic localization of the lymphocystis disease-resistant traits of Japanese flounder (Fuji et al. 2007), Ceratomyxa shasta resistance in salmon (Nichols et al. 2003), resistant trait to sea lice in the Atlantic salmon population (Tsai et al. 2016), expressed sequence tags from the ciliate protozoan parasite (Abernathy et al. 2007), and genetic analysis of resistant traits of Columnaris and Edwardsiella ictaluri in catfish (Zhou et al. 2017; Geng et al. 2015). Thus, it is vital to uncover the molecular markers that genetically linked to disease resistance and subsequently apply marker-assisted selection to expedite selection programs. Traditionally, quantitative trait loci (QTL) mapping with a dense genetic linkage map is an effective approach for dissecting important traits and developing genetically linked markers for selective breeding. In the genomic era, many high-throughput genotyping technologies have been developed, and the genome-wide association study (GWAS) becomes an efficient way for trait localization which does not require pedigree information or mapping family.

Linkage maps are essential tools for genomic and genetic studies, as well as for genetic breeding of economically important species (Peng et al. 2016). Recently, linkage maps have been constructed for many important teleost species including zebrafish (Bradley et al. 2011), medaka (Naruse et al. 2000), catfish (Kucuktas et al. 2009), tilapia (Liu et al. 2013), carps (Zhao et al. 2013; Zhang et al. 2013; Peng et al. 2016), Japanese flounder (Shao et al. 2015), and many others. Moreover, high-density genetic linkage maps provide a framework for QTL localization and facilitate genetic assistant selection and breeding in many aquaculture species (Zhou et al. 2018; Chen et al. 2018; Niu et al. 2017). For instance, growth-related traits have been mapped and fully investigated in many fishes, such as Pacific oyster (Song et al. 2018), catfish (Geng et al. 2017), and Asian seabass (Wang et al. 2015); sex determination traits have been localized in the genetic map in tilapia (Andrey et al. 2006), halibut (Palaiokostas et al. 2013), and Pacific white shrimp (Yu et al. 2017); oxygen tolerance in hybrid catfish (Zhong et al. 2017) and Nile tilapia (Li et al. 2017); salinity tolerance in tilapia (Gu et al. 2018; Lin et al. 2018); and resistance characters in red sea bream (Sawayama et al. 2017), Asian seabass (Wang et al. 2017), and so on.

To date, three draft genomes of the large yellow croaker have been published (Wu et al. 2014; Ao et al. 2015b) (PRJNA491413) and several versions of the large yellow croaker linkage maps have been constructed using various genetic markers including amplified fragment length polymorphism (AFLP), microsatellite, and single nucleotide polymorphism (SNP) markers (Ye et al. 2014; Ao et al. 2015a; Ning et al. 2007), and mapped some growth-related QTL and regions (Xiao et al. 2015). Additionally, a series of studies have been carried out on the life cycle of cryptocaryonosis, the expression and regulation of the anti-insect peptide piscidin, and the resistance gene of large yellow croaker (Ying et al. 2016; Zhou et al. 2014b). However, none of QTL analysis with challenge of C. irritans. Therefore, constructing a high-resolution genetic linkage map is vital for fine-scale mapping of cryptocaryonosis-resistant traits on genomic scale.

Here, we constructed a double-digest restriction-site associated DNA (ddRAD)-based high-density genetic map of the large yellow croaker. The map comprised 24 linkage groups with a total of 5261 SNP markers. Comparative genomic analysis with medaka, stickleback, European seabass and tilapia genomes provides us new insights of the synteny between L. crocea and models or Perciformes teleosts. We also identified QTL for C. irritans resistance trait on the linkage map, and the candidate genes were then identified from the genome regions of QTL intervals. Our high-density linkage map provides a powerful tool for QTL precise localization and association study of economically important traits, as well as a framework for assembling and improving reference genome of large yellow croaker.

Material and Methods

Mapping Family and DNA Extraction

A F1 full-sib family of large yellow croaker was generated from a maricultural farm (Ningde Fufa Aquatic Breeding Co.) in Ningde, Fujian, China. A total of 136 5-month progenies were randomly collected post hatch from the mapping family for genotyping and challenge experiment. The animal models were infected with a concentration of 2000 C. irritans theronts per fish in 3 L of seawater. Four days after infection, large numbers of tomonts were found to adhere to the bottom of aquarium. The fish were then transferred to another clean aquarium without tomonts. The time of death was recorded in the whole process of the experiment. Genomic DNA was extracted from the dorsal fins using the phenol-chloroform protocol (Sambrook et al. 1989). After quantification by a NanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA) and integrity detection by agarose gel electrophoresis, DNA samples were stored at − 20 °C for further experiments.

ddRAD Library Construction and Sequencing

Genomic DNA of 138 samples (2 parents and 136 offspring) was used to build ddRAD library. The ddRAD library was prepared using the improved published protocol (Peterson et al. 2012). Briefly, an equal amount of DNA (2 μg) of each sample was completely double digested with 20 units of EcoRI and MspI. The reaction mixture was incubated at 37 °C for 13 h and 65 °C for 10 min. Then, the fragments were ligated with adaptors. The P1 adapter with forward amplification primer and barcode was added to the EcoRI overhang, and the P2 adapter with reverse amplification primer was added to the MspI overhang, respectively. The ligation reaction was conducted in a reaction volume of 30 μl at 16 °C for 6 h on a thermocycler with a heated lid set to 65 °C for 10 min. After the heat kill, the solution was cooled down at 2 °C per 90 s until it reaches room temperature. After ligation, 24 individuals with different barcodes were pooled within one index, purifying with AMPure XP beads (Beckman Coulter, Brea, CA, USA). The DNA fragments of 300–400 bp were retrieved on E-Gel (Thermo Fisher Scientific) and then amplified with 30 cycles of PCR with regular forward primer and indexes ligated with reverse primer. Finally, six ddRAD libraries were sequenced on the Illumina HiSeq-PE150 platform with a 150-bp paired-end strategy at Novogene Technology (Tianjin, China).

SNP Discovery and Genotyping

The raw reads were first de-multiplexed according to the specific barcode-index combination of each sample. Reads with low quality and with ambiguous bases were discarded and trimmed to 130 bp because a slight increase of sequencing errors was observed after 130 bp. The cleaned data was filtered using Stacks v1.37 (Catchen et al. 2013). Firstly, we used process_radtags to distinguish individuals and chose BWA to align with the L. crocea genome. The L. crocea genome was our newly resequenced and assembled chromosome-level reference genome (BioProject: PRJNA505758). Secondly, stacks were constructed for each sample with pstacks (− m, 5). All individuals were then used to construct a catalog of loci using cstacks. The program sstacks (− P, 8) was used for matching against the catalog. Finally, the program populations was used for genotyping and only a genotype of the best SNP of each tag was used for further analysis.

The VCF file generated from populations was extracted and converted to Ped/Map format that served as an input file to PLINK software. All individuals with missing SNP genotypes (> 25%) and minor allele frequency (< 5%) were removed using the “--geno” and “--maf” parameters and manually removed with the low genome coverage. Internal Perl scripts were used for data cleaning. Only the SNPs that were heterozygous in at least one parent and in accordance with the Mendelian inheritance were used for further linkage analysis.

Linkage Map Construction

A genetic linkage map was constructed using JoinMap 4.1 software (Van Ooijen 2011) with CP-type population, which is designed to handle F1 population data containing two various heterozygous and homozygous diploid parents. The logarithm of odds (LOD = 19.0) was used for distributing SNP markers to different linkage groups (LGs). The recombination frequency of the same LG markers was transformed into map distance (cM) via the Kosambi mapping method (Kosambi 1943) in the regression algorithm. All genetic linkage maps were drawn using MapChart (Voorrips 2002).

Comparative Genomic Analysis

Comparative genomics was analyzed between large yellow croaker and each of the two perciforms, European seabass (Dicentrarchus labrax) and Nile tilapia (Oreochromis niloticus), as well as between large yellow croaker and each of the two model teleost species, medaka (Oryzias latipes) and stickleback (Gasterosteus aculeatus). Here, we used our newly assembled chromosome-level reference genome of L. crocea (BioProject No. PRJNA505758) and the four closely related species were downloaded from NCBI databases (https://www.ncbi.nlm.nih.gov/genome). The mapping and alignment were conducted using ddRAD tags with an e value cutoff of 1e−5 using BLASTN (Zhang et al. 2000). Among all the results obtained, we extracted non-overlapping sequences, each of which showed the lowest e value and the longest sequence length among the hits to a given region. There were no special requirements for the linkage group and chromosome number. Finally, syntenic regions defined by the top hits of the homology search were visualized using the program Circos v0.67 (Krzywinski and Schein 2009).

QTL Mapping Analysis and Association Study of Disease-Resistant Traits

QTL mapping with death time phenotypes was performed using the MapQTL6 software package (Van Ooijen 2011) with composite interval mapping (IM) and restricted multiple QTL model (MQM) mapping algorithms. LOD score significance thresholds were calculated by 1500 permutation tests for the trait. The genome-wide LOD threshold at a P value < 0.05 was considered to be significant. As a complementary method for QTL mapping, the genome-wide association study was performed between genotypes and traits using genome-wide efficient mixed model association (GEMMA) (Zhou and Stephens 2012) with a univariate linear mixed model (LMM) of phenotype on genotype. Combining the results from both QTL mapping and association analysis, we identified SNP loci that significantly correlated with the traits and extracted candidate genes cross these SNP loci to identify candidate genes associated with cryptocaryonosis-resistant traits. Briefly, we screened the ± 50,000-bp genome regions surrounding the significant SNPs based on the L. crocea reference genome and annotated the candidate genes accurately by BLAST against the non-redundant protein database.

Results

ddRAD-Seq Library Construction and SNP Quality Control

A total of 6 ddRAD sequencing (ddRAD-seq) libraries from 2 parents and their 136 offspring were constructed and sequenced by the HiSeq PE150 platform (BioProject No. PRJNA513258). After trimming, filtering the lower quality reads, assessing the genotyping quality, and removing samples according to sequencing mean depth (< 25), a total of 109,503 SNP markers were obtained from 2 parents and 124 progenies. Then, we continue to filter the SNPs with missing SNP genotypes (> 25%) and minor allele frequency (< 5%), and a total of 30,540 SNP markers were retained for linkage analysis. According to the Mendelian inheritance together with all tags which only hold the single version, 6226 SNPs were heterozygous in either one of the parents and were genotyped for adequate numbers of progeny (Supplementary File). Finally, only 5261 SNP markers were used for constructing the linkage map based on the LOD threshold. Details of each sample are shown in Table S1.

High-Density Genetic Map Construction

A high-density ddRAD-based SNP genetic map of large yellow croaker based on a pseudo-testcross population was constructed using JoinMap 4.1. The remaining putative 5261 SNPs were successfully grouped into 24 LGs (Table 1, Fig. 1). The total genetic map length was 1885.67 cM with an average LG that contained 219 effective loci spanning 78.570 cM. Specifically, LG3 had the most numbers of loci spanning 119.994 cM, whereas LG23 had the least number of effective loci spanning 61.092 cM. The inter-locus distance estimated based on the unique marker positions ranged from 0.25 cM in LG6 to 0.50 cM in LG24 with an average inter-locus distance of 0.36 cM in the map. Details of SNP names and positions on the 24 LGs of the integrated genetic map are listed in Table S2.

Table 1 Characteristics of the genetic map of the large yellow croaker
Full size table
Fig. 1
figure 1

The high-density consensus linkage genetic map for large yellow croaker

Full size image

Comparative Genomics

The high-density linkage map represented a valuable resource for future comparative genome analysis with closely related model fishes. We repasted the similarity of loci (1.0) filtered from JoinMap 4.1 onto the corresponding SNP chromosome and eventually obtained 5755 SNPs. The loci (1.0) meant that the two SNPs were completely consistent in all individuals. To compare chromosome framework among teleosts, we used the 5755 ddRAD tags searched against the chromosomes of medaka, stickleback, European seabass, and tilapia. The Circos plots were drawn based on the ortholog pairs between four species.

For comparison of the croaker-medaka and croaker-European seabass, 878 (15.27%) and 3277 (56.94%) of 5755 1:1 best ortholog pairs had been mapped on those paired chromosomes (Fig. 2a, b), respectively. Although we observed that parts of the large yellow croaker LGs correspond to several chromosomes in medaka or seabass, the major markers in the croaker chromosomes tended to be represented in only one corresponding chromosome, which built perfect chromosome-scale syntenic boxes.

Fig. 2
figure 2

Genomic comparisons as shown by the Circos diagram between large yellow croaker and medaka (a), European seabass (b), stickleback (c), and tilapia (d). Each colored arc represents an orthologous match between two species. We messed up the order of the four closely related species genome chromosomes on the image to better illustrate our results

Full size image

Similarly, comparative genomic analysis was also performed between croaker and stickleback genomes. A total of 1423 (24.73%) 1:1 best ortholog pairs were identified. The comparison between the croaker and stickleback (Fig. 2c) revealed that 18 of the 24 croaker chromosomes have chromosome-level conserved syntenic blocks with stickleback chromosomes. Except the 1:1 correspondence of chromosomes in the two species, we also identified the portion 2:1 chromosome relationship between croaker and stickleback. Chromosome 1 of stickleback ortholog occurred in croaker linkage groups LG13 and LG18. Likewise, stickleback chromosome 4 was examined to be co-orthologous to croaker LG7 and LG23 as well as to croaker LG10 and LG24 orthologs, which were distributed in stickleback chromosome 7. This interesting 1:2 synteny suggested that the chromosome fusion events might occur post the divergence of croaker and gasterosteus.

Moreover, comparative genomic analysis of large yellow croaker with tilapia was also performed, which revealed that a total of 1914 (33.26%) 1:1 best ortholog pairs were dentified in tilapia (Fig. 2d). The extensive syntenic blocks were quite obviously for some chromosomes/linkage pairs, e.g., LC_LG1-ON_chr6, LC_LG2-ON_chr8, and LC_LG21-ON_chr13, but there were also few exceptions in conditions of the croaker linkage group. The croaker LG4 was made of two corresponding tilapia chromosomes ON_chr17 and ON_chr23. Conversely, LC_LG17 and LC_LG20 was mainly syntenically related to one tilapia linkage group, ON_chr7. This syntenic pattern would imply at least three inter-chromosomal rearrangements between the two species.

Fine Mapping of QTL and Related Genes for Disease-Resistant Traits

We employed restricted MQM mapping to detect QTLs for resistant traits. The data of death time is shown in Table S3. Because of the small sample size, our phenotype distribution could not accurately satisfy the Gaussian mixture model. This was a preliminary result, and for the complex traits of C. irritans resistance, large samples were needed for more accurate QTL mapping. In general, seven significant QTLs under C. irritans challenge were distributed on LG5, LG6, LG7, and LG19 of large yellow croaker (Table 2 and Fig. 3). The most significant QTL CIR-5 located on LG19 at 26.640–26.902 cM presented the highest LOD value of 13.69, explaining the highest percentage of the phenotypic variations (32.7%). Another QTL CIR-6 and CIR-7 on LG19 were detected at position 65.201–66.877 with LOD values of 9.56–11.41 and contributions to a phenotypic variation of 32.1–32.4%. On LG5, CIR-1 and CIR-2 were located at 25.118–25.718 cM and 62.835–63.934 cM, explaining a 23.3–24.5% phenotypic variation. The QTL CIR-3 located on LG6 at 50.395–51.885 cM explained the lowest percentage of the phenotypic variations (3.1%) with a LOD value of 7.64 while QTL CIR-4 located on LG7 at 30.555–30.564 cM explained 30.8% of the total phenotypic variations with a LOD value of 7.43. We also conducted genome-wide association analysis between SNP genotypes and C. irritans challenge phenotypes as a complementary. The results revealed similar distribution patterns with QTL mappings. Integrating the results of both approaches, we certified that the identified SNP loci were significantly associated with disease-related traits.

Table 2 Summary of the C. irritans disease resistance QTLs in large yellow croaker
Full size table
Fig. 3
figure 3

QTL mapping and associate analysis of C. irritans resistance traits in large yellow croaker. The black horizontal line represents the genome-wide significance threshold where logarithm of odds (LOD) was 7.1

Full size image

Several genes related to disease resistance were identified from QTL mapping of the high-density map based on the large yellow croaker genome annotation. A total of 36 genes (Supplementary Table S4) were identified from the QTL regions, of which 29 (Table 3) were immune or disease-related genes, such as inhibitor of nuclear factor kappa-B kinase subunit epsilon (ikbke), nectin-2 (CD112), interferon alpha/beta receptor 1 (ifnar1), and interferon gamma receptor 2 (ifngr2). Inhibitor of nuclear factor kappa-B kinase subunit epsilon (IKBKE) is a non-canonical member of the IκB kinase (IKK) family and regulates immune response (Shih et al. 2011), which is identified on LG5 (Fig. 4a). Further, we identified nectin-2 on LG6 (Fig. 4b) which were immunoglobulin (Ig)-like cell adhesion molecules (CAMs) that play key roles in a variety of the formation of cell-cell adhesions and novel regulators of cellular activities, including cell polarization, differentiation, movement, proliferation, and survival (Takai et al. 2008). Interferon receptors (ifnrs) were located on QTL CIR-7 of LG19 (Fig. 4c) and belonged to the class II helical cytokine receptor family (CRF2) which could activate the responsive elements in the nucleus and induce the transcription and expression of antiviral proteins, playing an important role in regulating the immune response against foreign pathogens (Zhou et al. 2014a; Krause and Pestka 2005). Finally, the information on the mechanisms of ifnr, ikbke, and nectin-2 actions against C. irritans is visualized in Fig. 5.

Table 3 Summary of the candidate genes from QTLs in large yellow croaker
Full size table
Fig. 4
figure 4

a–c QTL region and candidate genes for C. irritans resistance traits in large yellow croaker. Genes are extracted in the ± 50,000-bp genome regions around the significant SNPs. Rectangles represent the candidate genes, and red rectangles were the highlighted genes

Full size image
Fig. 5
figure 5

Pathways of interferon receptor, inhibitor of nuclear factor kappa-B kinase subunit epsilon, and nectin-2 induction and effect in C. irritans infection. When C. irritans infect, NK cell with DNAM-1 and nectin-2 produce high levels of inflammatory cytokines, while ikbke promotes phosphorylation of interferon (ifn) regulatory factor 3 (irf3) and irf7 which bind to the response elements, eventually promoting expression of type I ifn genes. Type I ifns bind to respective receptors, activate the JAK-STAT pathway, and then induce the overlapping ISGs. IFN-γ activates STAT1 and induces a partially overlapping but distinct ISGs

Full size image

Discussion

Genetic maps which were constructed using sequence-based SNP markers were useful in genetic studies including comparative genomics, functional gene mapping, candidate gene positional cloning, and genome assembly (Kujur et al. 2015; Yue 2014; Xu et al. 2014a). Linkage maps have been constructed for many important aquaculture species since the first report of the linkage map of tilapia in 1998 (Kocher et al. 1998). To date, SNPs are a particularly attractive approach for genetic map construction due to the most common type of DNA polymorphism in the genome (Berthier-Schaad et al. 2010). With the progress of the next-generation sequencing technologies, various sequence-based SNP genotyping technologies have been detected, providing rapid and cost-effective high-throughput SNP genotyping platforms for linkage mapping. Additionally, RAD sequencing approaches come into vogue for linkage mapping in non-model species (Gonen et al. 2014; Shao et al. 2015; Le et al. 2015). In the present study, we used the ddRAD-seq approach to discover and genotype SNPs in large yellow croaker. Among the identified SNPs, 5261 ddRAD-based SNP markers were successfully genotyped and assigned to 24 LGs corresponding to the 24 chromosome pairs of the genome. The total length of the linkage map was 1885.67 cM with an average marker interval of 0.36 cM. Many genetic maps of large yellow croaker have been previously constructed using various markers, but the density of the maps varies considerably. Our new map was the first ultra-high-density linkage map, representing a significant improvement over the previous linkage maps constructed by microsatellite markers (0.76 cM) (Ye et al. 2014) and RAD-based SNP markers (0.54 cM) (Ao et al. 2015a), although the method (Lep-MAP (Rastas et al. 2013) vs JoinMap 4.1) and data type (SNP vs microsatellite) were different.

Comparative genomics is a powerful tool to transfer genomic information from model species to related non-model species. Teleost-specific genome duplication and accompanying genome rearrangements were reported to lead to teleost species with a higher rate of gene linkage disruption and lineage divergence than mammals (Kasahara et al. 2007; Meyer and Van de Peer 2005). Whole-genome duplication (WGD) and subsequent asymmetric changes in duplicated genes were thought to have an important role in genome evolution (Xu et al. 2014b). In a relatively short period of ~ 50 Myr after the WGD event (336~404 Myr ago), the MTZ ancestor (the last common ancestor of medaka, Tetraodon, and zebrafish) had 24 chromosomes and undergone 8 major inter-chromosomal rearrangements (2 fissions, 4 fusions, and 2 translocations) and only the medaka genome has preserved its ancestral genomic structure without undergoing major inter-chromosomal rearrangements for more than 300 Myr (Kasahara et al. 2007; Takeda 2008). In our study, we performed the chromosome-level comparisons between the croaker and two model species, medaka and stickleback. The results provided evidence that the croaker genome is more closely related to the medaka genome than the genome of stickleback. The comparison between the croaker and European seabass indicated that their chromosomes present 1:1 chromosome-level syntenic blocks, suggesting their close phylogenic relationship in perciforms.

In contrast, 2:1 syntenic correspondences were found in large yellow croaker linkage groups with stickleback (21) (Peichel et al. 2001) and tilapia (22) (Guyon et al. 2012), indicating that dramatic genome rearrangements occurred after these species diverged. This observation of a few exceptions of the 2:1 phenomenon suggests the genome plasticity and the underlying evolutionary constraints may not be evenly distributed across the genome, as large yellow croakers LG13 and LG23 both undergo inter-chromosomal rearrangements in stickleback and tilapia. Meanwhile, large-scale chromosomal rearrangements were rare but small fusion and translocations were subsistent. Moreover, all linkage groups in large yellow croaker mainly corresponded to one or two chromosomes in the four model fishes (Fig. 2), implying the high evolutionary conservation of chromosomes in these five species. Finally, large yellow croaker linkage group 4 corresponded to two chromosomes in tilapia, suggested that the group may easily occur independent of chromosomal fission events. Our findings could facilitate the functional inference of genes in croaker and further benefit the studies of karyotype evolution in teleosts.

The availability of mapping QTL is to understand the numbers and effects of genes that determine the traits on the whole genome and to assist in selective breeding to accelerate genetic improvement of important traits. Disease resistance is one of the most important researched traits in QTL studies as diseases represent one of the major challenges and bottlenecks in aquaculture. In our study, we identified 29 immune- or disease-related genes. Interferons have been shown to be a potent regulator of adaptive immunity, which now becomes clear that a broad range of viruses, bacteria, and even parasites express ligands capable of stimulating a growing number of signal pathways that must be specifically bound to the corresponding interferon receptors in order to perform biological functions (Smith et al. 2005). Thus, interferon receptors as molecular switches play a decisive role in the host immune response. Interferon receptor knockout models have made significant contributions to elucidating their associated immune mechanisms in mice (Jabbar et al. 2013; De et al. 2013). Some studies showed interferons, especially type I interferons, have been described as the master factors that bridge innate and adaptive immunity (Ivashkiv and Donlin 2014; Schoggins et al. 2010); clearly, the ifnar1 functions as the signal locks. In addition, disruption of interferon (IFN) or its receptor causes increased susceptibility to viral, bacterial, and parasitic infections in animal models (Smith et al. 2005). A study found IFN-gamma receptor null-mutant mice were more susceptible to herpes simplex virus type 1 infection than IFN-gamma ligand null-mutant mice (Cantin et al. 1999). Herpes simplex virus is a common human infectious skin disease, and after infection C. irritans, large yellow croaker also had the largest skin damage, which demonstrated the importance and reliability of the candidate gene. Interestingly, when fish is infected with C. irritans, interferon as one of the second lines of defense against pathogens may rapidly eliminate invasive parasites. Also, interferon binding to interferon receptors could trigger the induction of the expression of hundreds of IFN-inducible genes via the Janus kinase (JAK) and signal transduction and activators of transcription (STAT) signaling pathway (Sun et al. 2014). Additionally, in response to inflammatory factor and viral infection, IKBKE was activated and subsequently phosphorylates interferon response factors 3 and 7 (IRF3 and IRF7, respectively) and signal transducer and activator of transcription 1 (STAT1) (tenOever et al. 2007), this suggest ikbke and ifn may interact. Moreover, nectin-2 has some other important biological functions besides the structure features of an extracellular region with three Ig-like loops, a single transmembrane region, a cytoplasmic tail region (Takai et al. 2008), and cell adhesion functions, and it was also an important immune molecule. Some studies showed nectin-2 could activate the activity of natural killer (NK) cells and trigger a variety of immunological effects, including graft rejection, virus infection defense, and killing of tumor cells (Taharahanaoka et al. 2006; Baury et al. 2003). However, there were no reports about the functional domains and precise sites. Further studies stated nectin-2 has been identified as a ligand of DNAX accessory molecule 1 (DNAM-1) in both humans and mice (Taharahanaoka et al. 2005; Pende et al. 2005). Evidence from in vitro and in vivo studies revealed that the interaction of DNAM-1 with its ligands is involved in the functions of a variety of immune cells (Iguchimanaka et al. 2008; Gilfillan et al. 2008), so nectin-2 also plays an important role in immune cells. Nectin-2 was expressed in blood cells (Mori et al. 2014) and epithelial cells (Takai et al. 2008), and it may form a stable barrier on the surface of the skin through the cell-cell adhesion system to prevent the invasion of C. irritans or play roles in transient cell-cell contacts, such as those between macrophages and lymphocytes and between leukocytes and vascular endothelial cells, but the exact roles of nectin-2 in blood cells remain unknown.

Overall, our results provided some immune-associated SNP markers, which will facilitate the molecular marker-assisted selection design of resistant populations of large yellow croaker. Also, further verification of functional experiments of the mapped genes in this study is necessary.

Conclusions

Here, a high-density genetic linkage map was constructed with 5261 SNPs using the ddRAD-seq approach. The map consisted of 24 linkage groups, spanning 1885.67 cM with an average marker interval of 0.36 cM. Comparative genomic analyses have been conducted between large yellow croaker and four closely related species, indicating that high evolutionary conservation of chromosomes existed in these fish species as well as inter-chromosomal rearrangements with certain fusion, translocation, or fission events. Through SNP mapping analysis, we identified seven significant QTLs under C. irritans challenge and functional genes underlying these traits, which will be significant to improve breeding in large yellow croaker. Our work provides excellent useful genomic resources for the improvement of genome sequence assembly and molecular dissection of disease resistance, and lays a foundation in the future for molecular assisted breeding of new varieties of C. irritans resistance of large yellow croaker.