The pearl oyster, Pinctada fucata (P. fucata), is one of the marine bivalves that is predominantly cultured for pearl production. To obtain more genetic information for breeding purposes, we constructed a high-density linkage map of P. fucata and identified quantitative trait loci (QTL) for growth-related traits. One F1 family, which included the two parents, 48 largest progeny and 50 smallest progeny, was sampled to construct a linkage map using restriction site-associated DNA sequencing (RAD-Seq). With low coverage data, 1956.53 million clean reads and 86,342 candidate RAD loci were generated. A total of 1373 segregating SNPs were used to construct a sex-average linkage map. This spanned 1091.81 centimorgans (cM), with 14 linkage groups and an average marker interval of 1.41 cM. The genetic linkage map coverage, Coa, was 97.24%. Thirty-nine QTL-peak loci, for seven growth-related traits, were identified using the single-marker analysis, nonparametric mapping Kruskal-Wallis (KW) test. Parameters included three for shell height, six for shell length, five for shell width, four for hinge length, 11 for total weight, eight for soft tissue weight and two for shell weight. The QTL peak loci for shell height, shell length and shell weight were all located in linkage group 6. The genotype frequencies of most QTL peak loci showed significant differences between the large subpopulation and the small subpopulation (P<0.05). These results highlight the effectiveness of RAD-Seq as a tool for generation of QTL-targeted and genome-wide marker data in the non-model animal, P. fucata, and its possible utility in marker-assisted selection (MAS).
Citation: Li Y, He M (2014) Genetic Mapping and QTL Analysis of Growth-Related Traits in Pinctada fucata Using Restriction-Site Associated DNA Sequencing. PLoS ONE 9(11): e111707. https://doi.org/10.1371/journal.pone.0111707
Editor: Gen Hua Yue, Temasek Life Sciences Laboratory, Singapore
Received: April 24, 2014; Accepted: September 29, 2014; Published: November 4, 2014
Copyright: © 2014 Li, He. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by grants from the National Science and Technology Program of China (2012AA10A410), the Funds of Knowledge Innovation program of Chinese Academy of Sciences (ZCX2-EW-Q21), and the Marine Fishery Science and Technology Promotion Program of Guangdong Province, China (A201201A05, A201301A03). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The pearl oyster, Pinctada fucata (P. fucata), is a marine bivalve found in China and Japan, where it is predominantly cultured for pearl production. In China, the regions where it is cultured are mainly distributed along three coastal provinces (Guangdong, Guangxi and Hainan) of the South China Sea. Artificial breeding may no longer be commercially sustainable due to heavy mortality and the decline of pearl quality observed in established breeding programs . Many factors, such as inbreeding , disease outbreak  and environmental threats , have led to the genetic degeneration and high mortality of P. fucata. Solutions to genetic degeneration include the use of conventional breeding methods, selection, crossbreeding and hybridization to improve the quality of aquaculture strains . For P. fucata, breeding with the aim of high quality pearl production, for example the selection of pearl color, weight  or shell traits , has been conducted and gave positive results. Another solution to the degeneration of P. fucata is marker-assisted selection (MAS). Molecular genetic polymorphisms can be integrated with artificial selection of phenotypes by applying MAS to achieve substantial increases in the efficiency of artificial selection . The basic methodology of MAS is high-density linkage map construction and quantitative trait loci (QTL) detection. Genetic linkage maps are constructed by assigning polymorphic DNA markers to chromosome configurations . In shellfish, linkage maps of Crassostrea gigas , Argopecten irradians , Pinctada maxima  and Pinctada martensii  have been constructed. The availability of genetic linkage maps makes it possible to identify QTLs in aquaculture species and assists in the selection of desired traits, such as disease resistance, sex determination and growth traits . Quantitative trait loci in many aquaculture species, such as Crassostrea gigas (disease resistance) , Argopecten irradians (growth)  and Nodipecten subnodosus (orange shell color) , have been identified. However, studies of single nucleotide polymorphisms (SNP) have rarely been reported ,  and no QTLs have been identified in P. fucata. Therefore, it is important to identify new SNPs and QTLs that are associated with growth-related traits, which may contribute to the improvement of pearl production in P. fucata.
The application of restriction fragment length polymorphism, randomly amplified polymorphic DNA, amplified fragment length polymorphism, simple sequence repeats, SNPs and expressed sequence tags has changed the approach to genetic research in the field of aquaculture and has allowed the construction of genetic linkage maps . Among these markers, SNPs are the most abundant type found in the genome and are suitable for automated large-scale genotyping . There are over 30 different methods that are used for the genotyping of SNPs . For next-generation sequencing methods, restriction-site associated DNA sequencing (RAD-Seq) creates a reduced representation of the genome. It combines the use of restriction enzymes and molecular identifiers with Illumina sequencing to facilitate SNP discovery and high-throughput genotyping of large populations , . This technique was first applied to the genetic analysis of lateral plate armor in Gasterosteus aculeatus. Markers at the Eda locus, linked to plate loss, were identified in linkage group (LG) IV . In aquaculture species, RAD-Seq was used to construct a linkage map for Gnathopogon  and to detect SNPs linked to QTLs for infectious pancreatic necrosis resistance in Atlantic salmon . The RAD-Seq data can be readily analyzed without any prior genetic information, which makes the technique applicable to P. fucata for large-scale SNP discovery.
Methods that are frequently used for QTL mapping include the Kruskal–Wallis (KW) non-parametric test, interval mapping (IM) and composite interval mapping , , . The KW nonparametric test is equal to the nonparametric version of a one-way analysis of variance , which enables the detection of individual association between markers and traits , . An improved method, developed by Lander and Botstein , is IM. This method determines the likelihood for the presence of a segregating QTL at each position in the genome. The genetic effects of the QTL and the residual variance are also calculated when IM is used . In contrast to single-marker association KW nonparametric analysis, IM allows one to detect QTL effects at every map position, rather than just the marker positions . Based on IM, composite interval mapping was developed to eliminate the bias in QTL parameter estimation, reduce the residual error variance and increase the power of QTL detection . The KW nonparametric test is normally used for QTL analysis when no assumptions are made about the probability distribution(s) of the quantitative trait , . For bulk segregant studies (using individuals from the tails of the phenotypic distribution) based on high-throughput sequencing, a statistical framework for QTL mapping was described, and accelerated the identification of QTLs in both model and non-model organisms , .
The aim of this study was to identify a large number of SNP markers from the P. fucata genome using the RAD-seq method, with particular focus on high density genetic linkage map construction and growth-related QTL detection. In this analysis, we built a high-density sex-average linkage map of P. fucata using 1373 SNPs identified from RAD-seq. We also identified 39 QTL-peak loci for growth-related traits and found that the genotype frequencies were significantly different between the large and small subpopulations (P<0.05). These identified QTLs may be valuable resources for MAS in P. fucata.
Materials and Methods
All animal work has been conducted according to relevant national and international guidelines. No specific permissions are required to work with invertebrates in China. Similarly, no specific permissions were required for the collection of P. fucata from sample sites because they were not collected from protected areas of land. The pearl oyster P. fucata, is not an endangered or protected species.
Mapping family and trait measurement
The breeding program of P. fucata was carried out at the Marine Biology Research Station, Daya Bay, Chinese Academy of Sciences, Shenzhen, China. A female and male adult P. fucata were obtained from a wild population in Dapeng Bay (22°41′N–22°61′N, 114°26′E–114°51′E), Shenzhen, China. These were used as founders of an F1 intercross. The hinge length (HL) of 1047 6-month-old F1 progeny of the family was measured. Seventy four of the largest (HL >33 mm) and 74 of the smallest oysters (HL <19.7 mm) were sampled. Each oyster was numbered and shell height (SH), shell length (SL), shell width (SW), HL, total weight (Wt), soft tissue weight (Wf), and shell weight (Ws) were measured. The adductor muscle or the whole soft tissue was also sampled and preserved in 90% ethanol for DNA extraction. A total of 100 individuals, which included the two parents, the 48 largest individuals (defined as the large subpopulation) and the 50 smallest individuals (defined as the small subpopulation) were then used for linkage and QTL mapping.
RAD library construction and genotyping
Genomic DNA of the two parents and the 98 F1 progeny was extracted from the preserved samples using E.Z.N.A mollusc DNA Kit (Omega,USA), in accordance with the manufacturer’s instructions. The concentration of extracted DNA was estimated with a spectrophotometer, using OD260/280, and DNA quality was analyzed using agarose gel electrophoresis. The RAD library construction procedure closely followed the methodology described by Baird et al . Briefly, genomic DNA was digested at 37°C for 30 min with the restriction endonuclease EcoRI (Takara, Dalian, China). The reactions were heat inactivated at 65°C for 20 min. Modified Illumina adapters (P1 adapters) that contained barcode sequences were added to the samples, along with 1 µL of 100 mM rATP (Promega), 1 µL 10× EcoRI buffer, 0.5 µL (1000 U) T4 DNA Ligase (New England Biolabs) and 5 µL H2O. The reactions were incubated at room temperature for 20 min and then heat inactivated at 65°C for 20 min. The ligation products were then combined in appropriate multiplex pools. For each library pool, digested DNA was sheared to an average length of 500 bp. Each of the 22 library samples were then separated by electrophoresis through a 1.3% agarose gel and fragments in the 300–700 bp range were isolated using a MinElute Gel Extraction kit (Qiagen). The Quick Blunting Kit (New England Biolabs) was used to polish the ends of the DNA. The samples were re-purified and 15 U of Exo-Klenow (Enzymatics) were used to generate adenine overhangs on the 3′ end of the DNA, at 37°C. After subsequent purification, 1 µL of 10 µM P2 adapters (a divergent modified Solexa adapter, Illumina) were ligated to the DNA products. Samples were re-purified and eluted in 50 µL. Five microliters of this product were used in PCR amplification with 50 µL of Phusion Master Mix (New England Biolabs), 5 µL of 10 µM modified Solexa Amplification primer mix (Illumina) and 40 µL of H2O. Phusion PCR settings followed product guidelines for a total of 16–18 cycles. Samples were gel purified, to excise the DNA fragments (300–700 bp), and dissolved in Elution Buffer. After quality and quantity tests on the PCR products, the obtained RAD libraries were sequenced on an Illumina Hiseq 2000 platform. For each parent, sequencing data size was set to obtain twice as much of the genome. For each progeny, sequencing data size was set to obtain as much as half of the genome. Raw Illumina reads were filtered to eliminate adapter sequences. The reads with a low quality (Q≤5 E) base number of greater than half of the total number of nucleotides and reads that did not show the enzyme cleaved sequence, “AATTC”, in the first five bases were eliminated. Sequences were sorted to individuals in accordance with the barcode sequences. They were then used for clustering analysis (a minimum stack depth of two) and genotyping by the software stacks . The genotype of a locus was also determined by the maximum likelihood method.
Linkage map construction and analysis
The SNP markers were categorized into “lm×ll”, “nn×np” and “hk×hk” types, which represent heterozygosity in the maternal, paternal and both parents, respectively. Markers that were genotyped in less than 85% of progeny and markers that showed significantly distorted segregation (χ2 test, P<0.05) were excluded. All of the marker information was input into JoinMap 4.0  and the genetic analysis was carried out under the cross pollination population type. Similar markers were eliminated using the function “Similarity of Loci” and loci that segregated only in the paternal or maternal population were used for sex-specific linkage map construction, using the function “Create Maternal and Paternal Population Nodes”. The threshold independence logarithm of the odds (LOD) scores was set to 3.0–5.0 for the maternal map and 3.0–7.0 for the paternal map. The “lm×ll”, “nn×np” and “hk×hk” RAD-tag markers that were genotyped in more than 85% of progeny and segregated in a Mendelian manner were used to create the sex-average linkage map by JoinMap 4.0, with the threshold independence LOD score set to 4.0–6.0. Map distances (in cM) were calculated using Kosambi’s mapping function . To calculate genome coverage of the linkage maps, observed genome length (Goa) and expected genome length (Ge) need to be established. The Goa was taken as the total length of all markers on the framework map and included the triplets and doublets . Two approaches were used to estimate Ge of the linkage map: (1) Ge1 was calculated by adding 2 s (s was the average spacing of the genetic linkage map) to the length of each genetic linkage group, to account for chromosome ends . (2) Ge2 was calculated by multiplying the length of each genetic linkage group by (m+1)/(m−1), where m was the number of loci in each genetic linkage group . The average of Ge1 and Ge2 was used as Ge, to describe the genome length. Genome coverage (Coa) was denoted by Goa/Ge .
QTL detection and genotype frequency comparison
Pearson correlation coefficients were used to determine the linear correlation between pairs of growth-related traits, using software SPSS version 16.0 (SPSS Inc., USA). The mean size of growth-related traits was compared in the large and small subpopulation using the Student’s t-test in SPSS version 16.0. The significance level was set to 0.05. The QTL mapping was performed using MapQTL 5 software  and the KW non-parametric test was performed to determine the significant relationship between the regions of the genome and growth-related traits. We used the nonparametric KW analysis because the growth-related traits of the individuals used in this study deviated from a normal distribution, and no assumptions were being made about the probability distributions of the quantitative traits. The KW test ranks all individuals in accordance with the quantitative trait and classifies them in accordance with their marker genotype. A KW value larger than the thresholds given by the KW test (χ2 test, P<0.005) and a degree of correlation between loci and traits that was equal to or greater than four asterisks were used to identify QTL-peak loci . The genotype frequency difference between the large subpopulation and the small subpopulation was compared using the Chi-square test in SPSS version 16.0. The significance level was set to 0.05.
RAD sequencing and genotyping
There are few known SNP markers available for the linkage mapping of P. fucata. Hence, the RAD-seq analysis was employed for DNA polymorphism detection and genotyping. The Illumina Hiseq 2000 sequencing yielded a total of 84.36 gigabases (Gb) of clean bases from the 100 samples. The average number of clean bases per individual was 0.84 Gb. Sample 9 was the individual with least number of clean bases (0.45 Gb) and sample 2 was the individual with maximum number of clean bases (1.93 Gb) (Table S1). A total of 1956.53 million clean reads was generated and 86,342 candidate RAD loci were identified. After 84,336 candidate RAD loci were discarded (were scored in less than 85% of the progeny or were types of marker that had “ab×cd” for four alleles and “ef×eg” for three alleles), 2006 “lm×ll”, “nn×np” and “hk×hk” type of SNP markers, which were scored in enough individuals, were retained for further analysis. Amongst the 2006 markers, 625 showed deviation from Mendelian segregation (P<0.05). A total of 1381 SNP markers, which included 562 “lm×ll” type markers, 611 “nn×np” type markers and 208 “hk×hk” type markers, were retained for linkage map construction.
A total of 562 “lm×ll” type SNP markers were used for the female linkage map construction. After four similarity markers were discarded, 558 markers were located on the female linkage map. Fourteen linkage groups were identified and the number of loci per linkage group varied from nine to 94, with a mean of 39.8. The female linkage map spanned 1024.31 cM, with an average spacing of 3.19 cM and a Coa of 93.51% (Table 1, Figure S1).
For the construction of the male linkage map, 611 “nn×np” type SNP markers were used and four similarity markers were excluded. The remaining 607 markers were all located on the linkage map. Seventeen linkage groups were identified, which included two triplets and one doublet. The number of loci per linkage group varied from two to 158, with a mean of 35.7. The male linkage map covered 928.00 cM, with an average spacing of 4.03 cM and a Coa of 90.67% (Table 2, Figure S2).
Five hundred and fifty-eight “lm×ll” type SNP markers, 607 “nn×np” type SNP markers and 208 “hk×hk” type SNP markers were used for construction of the sex-average map. All of the 1373 markers (Table S2) were located on the linkage map. Fourteen linkage groups were formed and the number of markers per linkage group varied from 18 to 265, with a mean of 98. The sex-average linkage map covered 1091.81 cM, with an average spacing of 1.41 cM. The length of the linkage groups ranged from 63.22 cM to 115.79 cM, with an average of 77.98 cM. (Table 3, Figure 1). The expected genome length, Ge, was 1122.7 cM for the sex-average linkage map, with an expected genome length, Ge1, of 1114.3 cM and Ge2 of 1131.1 cM. The genetic linkage map coverage, Coa, was 97.24%. Given the estimated genome size of 1150 Mb for P. fucata , the average recombination rate across all of the linkage groups was 0.97 cM/Mb.
The map is composed of 14 linkage groups, with 1373 SNP markers, and spans 1091.81 cM. The map distances (in cM) are indicated on the left of the chromosomes and the names of the SNP markers are shown on the right.
The sex-average linkage map contained all the markers located in the female and male linkage maps. Some LGs in the sex-average map contained mainly SNP markers from the corresponding female and male LGs. For example, LG 3 of the sex-average linkage map contained 64 SNP markers, which included 27 markers from LG 3 (28 markers) of the female map and 23 markers of LG 3 (24 markers) of the male map. The SNP markers in LG 3 of the sex-average linkage map also presented a similar order of markers as in the female and male linkage maps.
Phenotypic values and correlation between growth-related traits
Pearson correlation coefficients between the growth-related traits ranged from 0.935 to 0.994 (Table 4), showing the growth-related trait was significantly correlated each other (P<0.01), and the mean value of each growth-related trait in the large subpopulation was significantly larger than that in the small subpopulation (P<0.0001) (Table 5).
To identify QTL-peak loci for growth-related traits, the phenotypic and genotypic data from the selected individuals were analyzed using the MapQTL5 software . The statistical analysis was conducted by a nonparametric genomic scan based on the KW non-parametric test, where markers were assessed for an association with trait data. Thresholds given by the KW statistic for the QTL detection of SH, SL, SW, HL, Wt, Wf and Ws (P<0.005) were 10.913, 12.579, 11.082, 10.885, 7.898, 8.115 and 11.303, respectively. A total of 39 QTL-peak loci for growth-related traits were identified, which included three for SH, six for SL, five for SW, four for HL, 11 for Wt, eight for Wf and two for Ws (Table 6). All of the QTL-peak loci for SH, SL and Ws were located on LG 6 and Pearson correlation coefficients between pairs of the three growth-related traits were higher than between most of the other trait pairs (Table 4). The QTL detected for other traits were located on LG 1, LG 5, LG 6, LG 7 and LG 10. Three loci were found to be consensus QTLs for five growth-related traits, which included locus 423886 (SH, SL, SW, HL and Wt), locus 288691 (SH, SL, SW, HL and Wf) and locus 99925 (SL, SW, HL, Wt and Ws) (Table 6).
Genotype frequency comparison
The genotype frequencies of all QTL-peak loci (except locus 196733 and locus 411871) showed significant differences (P<0.05) between the large subpopulation and the small subpopulation (Table 7). For example, the genotype frequencies for QTL-peak locus 423886 in the large subpopulation were 58% for the “np” type and 25% for the “nn” type. In the small subpopulation, the frequency for the “np” type was 26%, whilst the frequency for the “nn” type was 60%. The genotype frequencies of QTL-peak locus 466758 in the large subpopulation were 58% for the “hk” type, 12% for the “hh” type and 21% for the “kk” type. In the small subpopulation, the genotype frequencies were 32% for the “hk” type, 38% for the “hh” type and 16% for the “kk” type.
Linkage map construction and analysis
The sex-average linkage map has fourteen linkage groups, which is consistent with the number of haploid chromosomes of P. fucata . For the sex-specific maps, fourteen linkage groups were identified in the female map, whilst seventeen linkage groups were constructed in the male map. This discrepancy between the number of linkage groups and the haploid number was also found in Argopecten irradians  and zebrafish . The increase in number of linkage groups over the haploid number indicates that additional markers in several areas of the genome or a larger mapping population are required to reduce the linkage groups to the haploid chromosome number . Linkage maps for many bivalve species, such as Chlamys farreri  and Ostrea edulis L , have been constructed. Although they provide valuable information for the breeding of bivalve species, their use for QTL identification may be limited by their low resolution (5 to 20 cM). The sex-average linkage map for P. fucata consists of 1373 markers, with an average spacing of 1.4 cM. It is denser than the linkage maps for Pinctada martensii and Pinctada maxima , . Linkage maps have been constructed for Pinctada martensii using amplified fragment length polymorphism and microsatellite markers from an F1 family of two parents and 78 progeny . The female map has an average interval of 14.9 cM and the male map has an average interval of 16.1 cM. The sex-average linkage map for Pinctada maxima was constructed from 887 SNPs, with an average spacing of 2.0 cM, using 335 individuals that belonged to eight families. Genotyping was performed using Illumina 3 k iSelect custom arrays . The sex-average linkage map coverage, Coa, observed here was 97.24%, which is similar to the linkage map of Chlamys farreri (99.5%, constructed by a 2b-restriction site-associated DNA method, using two parents and 96 progeny) . Such a dense linkage map contains detailed information on the genome of P. fucata and can be used as the basis for a better understanding of genetic structure in shellfish species.
The estimated genome length of the integrated linkage map for P. fucata was 1122.7 cM. This is shorter than the expected genome length of a consensus map for Pinctada martensii (1415.9 cM for the female map, with 110 markers, and 1323.2 cM for the male map, with 98 markers) . Maps of low resolution are commonly longer than maps of high density, with low marker density being the most likely cause of the overestimation of genome size , . In comparisons of high resolution genetic linkage maps, those with more markers may result in a longer genome length. The consensus map of Chlamys farreri (1551.9 cM, with 3806 markers)  is longer than linkage maps of P. fucata, in this study, and Pinctada maxima . The RAD-seq method was available for high-resolution linkage map construction of P. fucata and this research has provided foundation materials for the utility of RAD-seq in non-model animals.
In this study, 2006 SNPs were scored in enough individuals and belongs to “lm×ll”, “nn×np” and “hk×hk” type of SNP markers. Amongst the 2006 markers, 625 showed deviation from Mendelian segregation. One of the problems for linkage mapping is marker-segregation distortion, which has also been observed in Ostrea edulis and Crassostrea gigas , . The deviation from Mendelian expectations could be the result of duplicated genes, deleterious alleles, unusual meiotic segregation distortions and transposable elements –. The cause of SNP loci segregation distortion in P. fucata may be the same as that observed in Crassostrea gigas, in which Mendelian segregation occurs during early developmental stages but is distorted during later development. These distortions are largely attributable to selection against recessive deleterious mutations of fitness genes that are closely linked to the markers . However, markers with moderate segregation distortion have little effect on marker order or the length of the linkage map , . To be conservative, only the markers that segregated in a Mendelian manner (P<0.05) were used for the linkage map construction of P. fucata.
The IM method from the MapQTL5 software  was initially used to identify QTLs for P. fucata. The use of this method will provide high significance thresholds when the phenotype distribution is skewed , . In this study, a P. fucata family with 1047 progeny were produced, and individuals at the extremes of the phenotype distribution (largest individuals with the HL >33 mm and smallest individuals with the HL <19.7 mm) were selected for QTL mapping. The growth-related traits of the selected individuals deviated from a normal distribution, which meant that the permutation test of the IM method was not justified and the LOD value of significance thresholds was too large. The use of the IM method in this analysis failed to identify any regions that reached genome-wide significance. The KW nonparametric test for QTL detection does not assume a normal distribution of quantitative traits , . It has been used for QTL analysis in many species, such as the investigation of total biomass and shell width of Pacific lion-paw scallop (Nodipecten subnodosus) , plant persistency of red clover , cane splitting of red raspberry (Rubus idaeus)  and low-temperature tolerance of wheat . It ranks all individuals by the quantitative trait value and classifies them by their genotype. A segregating QTL (with a large effect) that is closely linked to the tested marker will result in large differences in average rank of the marker genotype classes . To take full advantage of the RAD-seq data, a statistical framework for QTL mapping ,  may be adopted for QTL identify in Pinctada fucata when a reference genome is available.
Although QTL analyses have been carried out for economically important traits in over 20 aquaculture species, most QTLs were only mapped in large spaces between markers . Using randomly selected individuals or all individuals of a family as mapping population, the identified QTLs are usually evenly distributed across the linkage groups , . To identify QTLs, the family structure is extremely important in the experimental design. For oysters with large full-sibling families, generally only individuals with extreme phenotypes are genotyped . The selection of individuals with extreme traits focuses the analysis on the most genetically informative individuals, reduces the overall cost of genotyping and increases the experimental power , , . Here, we used the extreme phenotype trait individuals for QTL mapping to see whether the identified QTLs will be concentratedly distributed across the LGs. Growth-related traits of P. fucata are of particular interest to researchers due to their correlations with pearl production. In this study, 39 QTL-peak loci for growth-related traits were identified and most of the QTL-peak loci were distributed on LG 6. Many QTL-peak loci, such as locus 423886, locus 288691 and locus 99925, were also found to be consensus QTLs for growth-related traits. The consensus QTLs may correspond to distinct, closely linked QTLs or one QTL that acts upon several quantitative characters involved in the same metabolic pathway . Consensus QTLs have previously been observed in soybean, in which several QTLs for agronomic traits were mapped to the same loci . The detection of concentratedly distributed and consensus QTLs in P. fucata highlight their further application for MAS.
Genotype frequency comparison
Marker-trait association analysis is an alternative approach to QTL mapping that we used to identify QTLs for growth . In this methodology, association is inferred by the comparison of allelic frequency differences between selected populations . The large subpopulation and small subpopulation selected from the mapping family were used for allele frequency comparison. All growth-related traits had significantly higher mean values in the large subpopulation than in the small subpopulation (P<0.0001) and the genotype frequencies of most QTL-peak loci showed significant differences (P<0.05) between the two subpopulations. For example, the genotype frequencies at locus 423886 and locus 466758 showed significant differences between the two subpopulations. The frequency of the “np” genotype of locus 423886, plus “hk” and “kk” of locus 466758, were higher in the large subpopulation than in the small subpopulation. That may be caused by the accumulation of genotype frequency at marker-trait association loci during the selection of extreme individuals. These identified marker-trait association loci may be valuable resources for molecular selection breeding in P. fucata.
The identification of high growth-related traits that may contribute to the improvement of pearl production is the main goal of genetic breeding programs of P. fucata. The RAD-seq method was successfully utilized for the identification of a large number of SNPs and the construction of a high-density genetic linkage map of P. fucata. The choice of extreme phenotype individuals from a large full-sib P. fucata family as the mapping population was able to identify QTLs by the KW non-parametric test. Many QTL-peak loci were found to be consensus loci for growth-related traits and were located at similar regions of the linkage groups. Significant differences in genotype frequency, for most of the QTL-peak loci, were also observed between the large and the small subpopulations. The choice of extreme phenotype individuals from a large full-sib P. fucata family as the mapping population could yield QTLs that concentrate distribution across the linkage groups. The location of QTLs at similar regions of the linkage groups and QTLs that act upon several quantitative characters highlight their further application for MAS in P. fucata.
The female linkage map for Pinctada fucata. The map is composed of 14 linkage groups, with 558 markers, and spans 1024.3 cM. The map distances (in cM) are indicated on the left of the chromosomes and the names of the SNP markers are shown on the right.
The male linkage map for Pinctada fucata. The map is composed of 17 linkage groups, with 607 markers, and spans 928.0 cM. The map distances (in cM) are indicated on the left of the chromosomes and the names of the SNP markers are shown on the right.
Statistics of sample information. Detailed statistics on 100 samples used as the mapping family. Q20-rate means the percentage of bases with a quality value ≥20 and Q30-rate means the percentage of bases with quality value ≥30.
Conceived and designed the experiments: MXH. Performed the experiments: YGL MXH. Analyzed the data: MXH YGL. Contributed reagents/materials/analysis tools: MXH YGL. Contributed to the writing of the manuscript: YGL MXH. Reared the pearl oysters and collected the phenotype data: MXH.
- 1. He MX, Guan YY, Yuan T, Zhang HY (2008) Realized heritability and response to selection for shell height in the pearl oyster Pinctada fucata (Gould). Aquac Res 39: 801–805.
- 2. Wada KT, Komaru A (1994) Effect of selection for shell coloration on growth rate and mortality in the Japanese pearl oyster, Pinctada fucata martensii. Aquaculture 125: 59–65.
- 3. Miyazaki T, Goto K, Kobayashi T, Kageyama T, Miyata M (1999) Mass mortalities associated with a virus disease in Japanese pearl oysters Pinctada fucata martensii. Dis Aquat Organ 37: 1–12.
- 4. Liu WG, Huang XD, Lin JS, He MX (2012) Seawater acidification and elevated temperature affect gene expression patterns of the pearl oyster Pinctada fucata. PLoS One 7: e33679.
- 5. Hulata G (2001) Genetic manipulations in aquaculture: a review of stock improvement by classical and modern technologies. Genetica 111: 155–173.
- 6. Wada KT, Komaru A (1996) Color and weight of pearls produced by grafting the mantle tissue from a selected population for white shell color of the Japanese pearl oyster Pinctada fucata martensii (Dunker). Aquaculture 142: 25–32.
- 7. Lande R, Thompson R (1990) Efficiency of marker-assisted selection in the improvement of quantitative traits. Genetics 124: 743–756.
- 8. Kakioka R, Kokita T, Kumada H, Watanabe K, Okuda N (2013) A RAD-based linkage map and comparative genomics in the gudgeons (genus Gnathopogon, Cyprinidae). BMC genomics 14: 32.
- 9. Hubert S, Hedgecock D (2004) Linkage maps of microsatellite DNA markers for the Pacific oyster Crassostrea gigas. Genetics 168: 351–362.
- 10. Qin YJ, Liu X, Zhang HB, Zhang GF, Guo XM (2007) Genetic mapping of size-related quantitative trait loci (QTL) in the bay scallop (Argopecten irradians) using AFLP and microsatellite markers. Aquaculture 272: 281–290.
- 11. Jones DB, Jerry DR, Khatkar MS, Raadsma HW, Zenger KR (2013) A high-density SNP genetic linkage map for the silver-lipped pearl oyster, Pinctada maxima: a valuable resource for gene localisation and marker-assisted selection. BMC genomics 14: 810.
- 12. Shi YH, Kui H, Guo XM, Gu ZF, Wang Y, et al. (2009) Genetic linkage map of the pearl oyster, Pinctada martensii (Dunker). Aquac Res 41: 35–44.
- 13. Yue GH (2014) Recent advances of genome mapping and marker-assisted selection in aquaculture. Fish Fish (Oxf) 15: 376–396.
- 14. Sauvage C, Boudry P, De Koning DJ, Haley CS, Heurtebise S, et al. (2010) QTL for resistance to summer mortality and OsHV-1 load in the Pacific oyster (Crassostrea gigas). Anim Genet 41: 390–399.
- 15. Petersen JL, Baerwald MR, Ibarra AM, May B (2012) A first-generation linkage map of the Pacific lion-paw scallop (Nodipecten subnodosus): Initial evidence of QTL for size traits and markers linked to orange shell color. Aquaculture 350: 200–209.
- 16. Shi YH, Yu CC, Gu ZF, Zhan X, Wang Y, Wang AM (2013) Characterization of the pearl oyster (Pinctada martensii) mantle transcriptome unravels biomineralization genes. Mar Biotechnol 15: 175–187.
- 17. Huang XD, Wu SZ, Guan YY, LI YG, He MX (2013) Identification of sixteen single-nucleotide polymorphism markers in the pearl oyster, Pinctada fucata, for population genetic structure analysis. J Genet 93: e1–e4.
- 18. Liu ZJ, Cordes J (2004) DNA marker technologies and their applications in aquaculture genetics. Aquaculture 238: 1–37.
- 19. Lindblad-Toh K, Winchester E, Daly MJ, Wang DG, Hirschhorn JN, et al. (2000) Large-scale discovery and genotyping of single-nucleotide polymorphisms in the mouse. Nat Genet 24: 381–386.
- 20. Perkel J (2008) SNP genotyping: six technologies that keyed a revolution. Nat Methods 5: 447–454.
- 21. Houston RD, Davey JW, Bishop SC, Lowe NR, Mota-Velasco JC, et al. (2012) Characterisation of QTL-linked and genome-wide restriction site-associated DNA (RAD) markers in farmed Atlantic salmon. BMC genomics 13: 244.
- 22. Baird NA, Etter PD, Atwood TS, Currey MC, Shiver AL, et al. (2008) Rapid SNP discovery and genetic mapping using sequenced RAD markers. PLoS One 3: e3376.
- 23. Sánchez-Molano E, Cerna A, Toro M A, Bouza C, Hermida M, et al. (2001) Detection of growth-related QTL in turbot (Scophthalmus maximus). BMC genomics 12: 473.
- 24. Ulloa M, Hutmacher R B, Roberts P A, Wright SD, Nichols RL, et al. (2013) Inheritance and QTL mapping of Fusarium wilt race 4 resistance in cotton. Theor Appl Genet 126: 1405–1418.
- 25. Van Ooijen JW (2004) MapQTL 5. Software for the mapping of quantitative trait loci in experimental populations. Kyazma BV, Wageningen.
- 26. Atienza SG, Satovic Z, Petersen KK, Dolstra O, Martin A (2003) Identificationof QTLs influencing agronomic traits in Miscanthus sinensis Anderss. I. Total height, flag-leaf height and stem diameter. Theor Appl Genet 107: 123–129.
- 27. Lander ES, Botstein D (1989) Mapping mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121: 185–199.
- 28. Alonso-Blanco C, Koornneef M, van Ooijen JW (2006) QTL analysis. Methods Mol Biol 323: 79–99.
- 29. Gao HJ, Yang RQ (2006) Composite interval mapping of QTL for dynamic traits. Chin Sci Bull 51: 1857–1862.
- 30. Wang C, Ulloa M, Mullens TR, John ZY, Roberts PA (2012) QTL analysis for transgressive resistance to Root-Knot Nematode in interspecific cotton (Gossypium spp.) progeny derived from susceptible parents. PLoS One 7: e34874.
- 31. Magwene PM, Willis JH, Kelly JK (2011) The statistics of bulk segregant analysis using next generation sequencing. PLoS Comput Biol 7: e1002255.
- 32. Yang ZM, Huang DQ, Tang WQ, Zheng Y, Liang KJ, et al. (2013) Mapping of quantitative trait loci underlying cold tolerance in rice seedlings via high-throughput sequencing of pooled extremes. PLoS One 8(7): e68433.
- 33. Catchen J, Hohenlohe PA, Bassham S, Amores A, Cresko WA (2013) Stacks: an analysis tool set for population genomics. Mol Ecol 22: 3124–3140.
- 34. Van Ooijen JW (2006) JoinMap 4. Software for the calculation of genetic linkage maps in experimental populations. Kyazma BV, Wageningen, Netherlands.
- 35. Kosambi D (1943) The estimation of map distances from recombination values. Ann Eugen 12: 172–175.
- 36. Cervera MT, Storme V, Ivens B, Gusmao J, Liu BH, et al. (2001) Dense genetic linkage maps of three Populus species (Populus deltoides, P. nigra and P. trichocarpa) based on AFLP and microsatellite markers. Genetics 158: 787–809.
- 37. Fishman L, Kelly AJ, Morgan E, Willis JH (2001) A genetic map in the Mimulus guttatus species complex reveals transmission ratio distortion due to heterospecific interactions. Genetics 159: 1701–1716.
- 38. Chakravarti A, Lasher LK, Reefer JE (1991) A maximum likelihood method for estimating genome length using genetic linkage data. Genetics 128: 175–182.
- 39. Liao M, Zhang L, Yang G, Zhu M, Wang D, et al. (2007) Development of silver carp (Hypophthalmichthys molitrix) and bighead carp (Aristichthys nobilis) genetic maps using microsatellite and AFLP markers and a pseudo-testcross strategy. Anim Genet 38: 364–370.
- 40. Takeuchi T, Kawashima T, Koyanagi R, Gyoja F, Tanaka M, et al. (2012) Draft genome of the pearl oyster Pinctada fucata: a platform for understanding bivalve biology. DNA Res 19: 117–130.
- 41. Wada K (1976) Number and gross morphology of chromosomes in the peal oyster, Pinctada fucata (GOULD), collected from two regions of Japan. Jap J Malac 35: 9–14.
- 42. Postlethwait JH, Johnson SL, Midson CN, Talbot WS, Gates M, et al. (1994) A genetic linkage map for the zebrafish. Science 264: 699–703.
- 43. Semagn K, Bjørnstad Å, Skinnes H, Marøy AG, Tarkegne Y, et al. (2006) Distribution of DArT, AFLP, and SSR markers in a genetic linkage map of a doubled-haploid hexaploid wheat population. Genome 49: 545–555.
- 44. Wang LL, Song LS, Chang YQ, Xu W, Ni DJ, et al. (2005) A preliminary genetic map of Zhikong scallop (Chlamys farreri Jones et Preston 1904). Aquac Res 36: 643–653.
- 45. Lallias D, Beaumont AR, Haley CS, Boudry P, Heurtebise S, et al. (2007) A first-generation genetic linkage map of the European flat oyster Ostrea edulis (L.) based on AFLP and microsatellite markers. Anim Genet 38: 560–568.
- 46. Jiao WQ, Fu XT, Dou JZ, Li HD, Su HL, et al. (2014) High-resolution linkage and quantitative trait locus mapping aided by genome survey sequencing: building up an integrative genomic framework for a bivalve mollusc. DNA Res 21: 85–101.
- 47. Wada H, Naruse K, Shimada A, Shima A (1995) Genetic linkage map of a fish, the Japanese medaka Oryzias latipes. Mol Mar Biol Biotechnol 4: 269–274.
- 48. Yu ZN, Guo XM (2003) Genetic linkage map of the eastern oyster Crassostrea virginica Gmelin. Biol Bull 2003 204: 327–338.
- 49. Naciri Y, Vigouroux Y, Dallas J, Desmarais E, Delsert C, et al. (1995) Identification and inheritance of (GA/TC) n and (AC/GT) n repeats in the European flat oyster Ostrea edulis (L). Mol Mar Biol Biotechnol 4: 83–89.
- 50. Launey S, Hedgecock D (2001) High genetic load in the Pacific oyster Crassostrea gigas. Genetics 159: 255–265.
- 51. Gut IG, Lathrop GM (2004) Duplicating SNPs. Nat Genet 36: 789–790.
- 52. Martínez-Navarro B, Palmqvist P, Shabel AB, Pérez-Claros JA, Lorenzo C, et al. (2008) Reply to Gibert, et al.(2008) on the supposed human phalanx from Cueva Victoria (Cartagena, Spain). J Hum Evol 54: 157–161.
- 53. Hubert S, Higgins B, Borza T, Bowman S (2010) Development of a SNP resource and a genetic linkage map for Atlantic cod (Gadus morhua). BMC genomics 11: 191.
- 54. Knox MR, Ellis THN (2002) Excess heterozygosity contributes to genetic map expansion in pea recombinant inbred populations. Genetics 162: 861–873.
- 55. Hackett CA, Broadfoot LB (2003) Effects of genotyping errors, missing values and segregation distortion in molecular marker data on the construction of linkage maps. Heredity 90: 33–38.
- 56. Guo X, Li Q, Wang QZ, Kong LF (2012) Genetic mapping and QTL analysis of growth-related traits in the Pacific oyster. Mar Biotechnol 14: 218–226.
- 57. Welch S, Gebhart SSP, Bergman RN, Phillips LS (1990) Minimal model analysis of intravenous glucose tolerance test-derived insulin sensitivity in diabetic subjects. J Clin Endocrinol Metab 71: 1508–1518.
- 58. Klimenko I, Razgulayeva N, Gau M, Okumura K, Nakaya A, et al. (2010) Mapping candidate QTLs related to plant persistency in red clover. Theor Appl Genet 120: 1253–1263.
- 59. Woodhead M, Williamson S, Smith K, McCallum S, Jennings N, et al. (2013) Identification of quantitative trait loci for cane splitting in red raspberry (Rubus idaeus). Mol Breeding 31: 111–122.
- 60. Båga M, Chodaparambil SV, Limin AE, Pecar M, Fowler DB, et al. (2007) Identification of quantitative trait loci and associated candidate genes for low-temperature tolerance in cold-hardy winter wheat. Funct Integr Genomics 7: 53–68.
- 61. Li HJ, Liu X, Zhang GF (2012) A consensus microsatellite-based linkage map for the hermaphroditic bay scallop (Argopecten irradians) and its application in size-related QTL analysis. PLoS One 7: e46926.
- 62. Massault C, Bovenhuis H, Haley C, de Koning DJ (2008) QTL mapping designs for aquaculture. Aquaculture 285: 23–29.
- 63. Lebowitz M (1987) Experiments with incremental concept formation: Unimem. Mach Learn 2: 103–138.
- 64. Dirlewanger E, Moing A, Rothan C, Svanella L, Pronier V, et al. (1999) Mapping QTLs controlling fruit quality in peach (Prunus persica (L.) Batsch). Theor Appl Genet 98: 18–31.
- 65. Zhang WK, Wang YJ, Luo GZ, Zhang JS, He CY, et al. (2004) QTL mapping of ten agronomic traits on the soybean (Glycine max L. Merr.) genetic map and their association with EST markers. Theor Appl Genet 108: 1131–1139.
- 66. Lu X, Wang HX, Liu BZ, Xiang JH (2013) Three EST-SSR markers associated with QTL for the growth of the clam Meretrix meretrix revealed by selective genotyping. Mar Biotechnol 15: 16–25.
- 67. Stuber CW, Moll RH, Goodman MM, Schaffer HE, Weir BS (1980) Allozyme frequency changes associated with selection for increased grain yield in maize (Zea mays L.). Genetics 95: 225–236.