BAC library construction and BAC end sequencing

The yellowtail genomic DNA content, represented as C-value, was estimated to be 0.7 pg/cell (data not shown) using flow cytometric analysis and hence the genome size was calculated to be approximately 685 Mb. Of 100 randomly selected BAC clones, 71 (71%) contained inserts, indicating that approximately 78,520 (71% of 110,592 clones) clones had an insert. The size distribution of the 71 clones with inserts was from about 20 kb to 220 kb and average insert length was 140.7 kb (data not shown). Therefore, it is estimated that the total length of the yellowtail BAC library insert DNA was approximately 11 Gb and was 16-times larger than the yellowtail genome. It is known that a minimum of 5-10 × coverage across the entire genome is required for a BAC library to be useful for positional cloning, physical mapping, and genome sequencing [19]. Therefore, the yellowtail BAC library is sufficient for further genomic/genetic analysis except for studies regarding W-linked genes because of our ZZ male derived DNA source [20].

By sequencing both ends of randomly-selected 2,960 BAC clones, a total of 5,920 raw reads were obtained, and of those reads, 4,956 reads (2,471 in the SP6 side and 2,485 in the T7 side) were qualified for subsequent repeat identification and BLAST search (GA867436 - GA872391). Total length of the qualified BESs was 3,074,133 bp with an average size of 620 bp, representing approximately 0.45% of the yellowtail genome. The GC content was estimated to be 41.36%, which is almost the same as other fishes (Takifugu rubripes: 45.46%; Gasterosteus aculeatus: 44.60%; Oryzias latipes: 40.46%; Tetraodon nigroviridis: 46.43%) (http://esper.lab.nig.ac.jp/genome-composition-database/).

Preliminary survey of genomic content using BESs

Genetic linkage map

Out of the 743 primer pairs designed from the qualified BESs, 373 primer pairs (27 mononucleotide repeats, 285 dinucleotide repeats, 31 trinucleotide repeats, 26 tetranucleotide repeats, 3 pentanucleotide repeats and 1 hexanucleotide repeat) produced amplicons. In addition to the 464 microsatellite markers derived from the genomic library Ohara et al. developed [2, 3], 837 markers in total were included in the yellowtail genetic linkage maps (Additional file 1). No segregation distortion was observed in any markers and hence lethal allele-linked markers were not included in our marker set.

Resultant yellowtail female and male genetic map consists of 715 and 702 markers including 232 and 271 framework markers, spanning 1,026.65 and 1,057.83 cM Kosambi with an average interval 4.96 and 4.32 cM on 24 linkage groups respectively (Table 4, Figure 1). The number of chromosomes in yellowtail has been reported to be 2n = 48 and hence the SSR markers we developed are distributed throughout the yellowtail genome [28]. The “gaps” observed in Squ21 and 24 in male and both map respectively might be caused by “recombination hot-spots” where recombination occurs frequently (Figure 1). The genome length was estimated to be 1,274.64 (L ₁ ) and 1,284.34 (L ₂ ) cM in the female and 1,282.35 (L ₁ ) and 1,285.45 (L ₂ ) cM in the male map by the two different methods respectively (see Materials and Methods). Using formula c = 1 – e^-2dn/L and estimated genome length L, coverage of the female and male map is estimated to be 83.3 to 83.9% respectively (Table 4). Considering the average interval less than 10 cM and the genome coverage, we concluded that the yellowtail genetic map was sufficient for further QTL studies [29].

	Female map					Male map
	No. of markers			Genome length		No. of markers			Genome length
	All	Framework	Length (cM)	L 1 (cM)	L 2 (cM)	All	Framework	Length (cM)	L 1 (cM)	L 2 (cM)
Squ1	33	11	48.19	58.11	57.83	34	15	49.28	57.91	56.32
Squ2	49	14	49.27	59.19	56.85	48	18	43.45	52.09	48.56
Squ3	29	10	40.17	50.09	49.10	29	18	54.97	63.60	61.43
Squ4	28	9	42.82	52.74	53.52	25	3	5.57	14.20	11.14
Squ5	38	9	59.51	69.43	74.39	39	17	50.24	58.87	56.52
Squ6	30	5	50.31	60.23	75.47	23	11	42.62	51.26	51.15
Squ7	29	14	50.66	60.58	58.45	32	11	40.41	49.04	48.49
Squ8	26	8	54.83	64.75	70.49	23	12	51.60	60.24	60.99
Squ9	37	10	30.22	40.14	36.94	36	10	52.89	61.53	64.65
Squ10	35	7	40.60	50.52	54.13	36	7	28.60	37.24	38.14
Squ11	10	1	0.00	9.92	0.00	11	6	59.19	67.82	82.86
Squ12	26	12	26.74	36.66	31.60	27	10	52.09	60.73	63.67
Squ13	26	10	48.03	57.95	58.71	25	10	42.63	51.27	52.10
Squ14	34	9	56.36	66.28	70.45	33	11	45.39	54.02	54.47
Squ15	35	14	37.23	47.15	42.96	36	11	55.44	64.07	66.52
Squ16	25	8	36.12	46.04	46.44	27	14	51.66	60.29	59.61
Squ17	20	10	40.32	50.24	49.28	22	7	44.55	53.18	59.39
Squ18	37	10	45.12	55.04	55.15	33	13	57.00	65.64	66.51
Squ19	30	9	40.52	50.44	50.65	26	14	58.45	67.08	67.44
Squ20	22	9	51.21	61.13	64.01	24	10	40.38	49.02	49.36
Squ21	29	12	44.90	54.82	53.06	3	2	1.11	9.75	3.33
Squ21′	N/A	N/A	N/A	N/A	N/A	25	5	7.78	16.42	11.68
Squ22	26	12	45.87	55.79	54.21	26	11	54.03	62.67	64.84
Squ23	25	7	41.67	51.59	55.56	26	10	44.03	52.67	53.82
Squ24	4	4	15.67	25.59	26.12	29	11	17.79	26.42	21.34
Squ24′	32	8	30.31	40.23	38.97	4	4	6.68	15.31	11.13
Average	29	9	41.07	50.99	51.37	27	10	40.69	49.32	49.44
Total	715	232	1,026.65	1,274.64	1,284.34	702	271	1,057.83	1,282.35	1,285.45
Genome coverage c (%)				83.56	83.33				83.88	83.81

Identification orthologous chromosomes with other fishes

In addition to the BAC or whole genome sequence, comparative genome analysis especially conserved synteny would be helpful for fine-scale QTL analyses and/or understanding the genetic basis of the traits [30, 31]. BLAST searches of the 818 mapped yellowtail loci against medaka, Tetraodon, stickleback, fugu and zebrafish proved that 25.7, 23.0, 42.2, 24.4 and 9.4% of the loci were mapped to each genome sequence. Oxford grids showed that eighteen linkage group pairs between yellowtail and stickleback retained a one-to-one relationship, and another three stickleback and six yellowtail linkage groups had a one-to-two relationship, implying that chromosomal fusions or breakages occurred after divergence from ancestor of both species (Figure 2). Nevertheless, the result suggests conserved synteny between yellowtail and stickleback and hence the stickleback genome data would be useful as a reference of yellowtail genome.

Ethics statements

Field permits are not required for this species in Japan. Since all fish treatments were performed in Goto Branch of Seikai National Fisheries Research Institute of Fisheries Research Agency, fish handling, husbandry and sampling methods were approved by Institutional Animal Care and Use Committee of National Research Institute of Aquaculture (IACUC-NRIA No. 03).

BAC construction and BAC-end sequencing

The BAC library was constructed according to Katagiri et al. with some modifications [14]. Briefly, at first, approximately 5 × 10⁷ frozen sperm cells taken from one male yellowtail were embedded in agarose plugs, digested with proteinase K overnight at 37°C and stored in 0.5 M EDTA following proteinase K inhibitor treatment until use. The plugs were dialyzed in 0.5 × TE, partially digested with Mbo I and size fractionated by pulse-field electrophoresis. The fraction containing 150 to 250 kb genomic DNA was excised from the gel and was recovered as high molecular weight (HMW) genomic DNA. The HMW genomic DNA was then integrated into Bam HI site of pBACe3.6 vector and reactions were transfected to E. coli DH10B strain. Finally, a total of 110,592 recombinant BAC clones were picked and stored in 288 384-well microtiter plates. The length of the insert DNAs was estimated by analyzing 100 BAC inserts digested with Not I.

The BESs were obtained from eight 384-well plates containing 3,072 clones. The BAC DNAs extracted by conventional alkaline lysis method were sequenced from SP6 and T7 sides with BigDye Terminator v3.1 Cycle Sequencing Kit (Life Technologies) following the manufacturer’s instructions and reactions were electrophoresed with Applied Biosystems 3730 DNA Analyzer (Life Technologies). All raw reads were processed using PHRED software with default parameters except for the trimming error probability was set at 0.01 [34, 35], and vector and bacterial sequences were masked by CROSS_MATCH implemented in PHRAP software. The masked BESs of more than 100 bp in length, hereafter called “qualified BES”, were extracted using our in-house perl script. The GC content of the extracted BESs was estimated using the geecee program included in the EMBOSS package [36].

Sequence data analysis

Repetitive DNA elements in the qualified BESs, such as transposable elements and SSRs, were identified and masked using Crossmatch search engine (v1.090518), “teleostei” repeat database implemented in Repbase RepeatMasker Edition (20120418) and RepeatMasker program (see http://www.repeatmasker.org/ for details).

The repeat-masked qualified BESs were subjected to homology search. The eight fish proteome data sets (Atlantic cod: gadMor1.70, Tetraodon: TETRAODON8.70, medaka: MEDAKA1.70, Nile tilapia: Orenil1.0.70, platyfish: Xipmac4.4.2.70, stickleback: BROADS1.70, fugu: FUGU4.70 and zebrafish: Zv9.70) were obtained from Ensembl (ftp://ftp.ensembl.org/). Only the longest protein for each gene was extracted and used for in-house database construction. For construction of genomic sequence database, the genomic sequences of stickleback (gasAcu1) and fugu (fr3) were downloaded from UCSC genome browser (http://hgdownload.soe.ucsc.edu/) and others from Ensembl.

BLAST searches were performed with qualified BESs as query with cut-off e-value e⁻⁹. The top hit query-subject pairs were extracted using in-house perl script under the criteria in which if multiple query-subject pairs were observed and were overlapped each other, only the most significant pair was considered significant.

Development of SSR markers

The SSR motifs and primer pairs in the qualified BESs were searched by WebSat online application (http://wsmartins.net/websat/) with default settings except for product size, which was set to 100–200 bp. For each BAC clone, the SP6-side BES was at first analyzed and if no SSR motif or primer-binding site was found in the sequence, T7-side BES of the same clone was alternatively used. In both sides, the SSR motifs containing over six repeats were considered as real SSRs. In the case where more than two SSR motifs were found in one read, the longest one was used as a representative. Finally, 743 SSRs (89 mono-, 550 di-, 68 tri-, 31 tetra-, 4 penta- and 1 hexa-nucleotide repeats) were selected for primer design.

In addition to the BES-derived SSRs, we also developed microsatellite markers from the genomic library constructed by Ohara et al. [2, 3]. The microsatellites containing CA/GT repeat motifs were isolated according to the protocol of Ohara et al. and primers were designed as described above [2, 3].

Mapping panel

The mapping panel consists of ninety progenies produced by artificial fertilization. Parent fish were caught off Goto Island, Nagasaki Prefecture, Japan and reared in a sea cage until they were matured with approximate body weight 7 kg. Human chorionic gonadotropin (ASKA Pharmaceutical) was intramuscularly administered to the parent fish at 600 IU/kg body weight and eggs and sperm were taken at 45 hours after administration. Fertilized eggs were kept in 500 L seawater at approximately 19°C with 0.5-1 L/min aeration until hatching. The juvenile fish were reared in 500 L seawater at 20-25°C until their body length reached 10 cm. The caudal fin was partially clipped from each progeny as the DNA source and kept in absolute ethanol until use. Genomic DNA of each fish was extracted using DNeasy Blood and Tissue Kit (Qiagen) according to manufacturer’s instructions.

Data acquisition

Genotyping was performed in an 11 μl reaction volume containing 0.5 pmol/μl of unlabelled primer, 0.05 pmol/μl of fluorescence-end-labeled primer with [5’-TET], 1 × buffer, 2.0 mM MgCl₂, 0.2 mM dNTP, 1.1 μg of BSA, 0.025 U of EX Taq DNA polymerase (Takara) and 25 ng of template DNA. PCR was performed on a GeneAmp® PCR System 9700 (Applied Biosystems), and the program conditions were 95°C for 2 min for initial denaturation, followed by 30 cycles of 30 sec at 95°C, 1 min at the annealing temperature (52-55°C), 1 min at 72°C and 10 min at 72°C for final extension. Amplification products were mixed with an equal volume of loading buffer (98% formamide, 10 mM EDTA, 0.05 w/v% bromophenol blue), heated for 10 min at 95°C and then immediately cooled on ice. 2 μl of each sample was loaded onto a 6% PAGE-PLUS gel (Amresco) containing 8 M urea and 0.5 × TBE buffer. Electrophoresis was performed in 0.5 × TBE buffer, and after electrophoresis, the gel was scanned and imaged using a FLA-9000 image scanner (GE Healthcare).

Linkage map construction

Genotype data obtained above were subjected for linkage analysis for the male and the female meiosis independently. Marker genotypes were analyzed with LINKMFEX ver. 2.3 (http://www.uoguelph.ca/~rdanzman/software.htm). Linkage analysis was performed using genotype data converted to a backcross format. As the grandparent genotypes were unknown, pairwise analyses were performed, and markers were sorted in linkage groups at a minimum LOD score of 4.0. A goodness-of-fit for Mendelian segregation distortion was tested for all alleles using the chi-square test (p < 0.05, d.f. = 1). Finally, the marker order was determined and double recombination events were checked with MapManagerQTX version 2.0 [37]. The resultant genetic map was visualized using MapChart version 2.2 [38].

The genome length L was estimated using two different methods following Fishman et al. [39]. In the first method (L ₁ ), average marker interval was estimated by dividing the summed length of all linkage groups by the number of intervals, and twice the average marker interval was added to each linkage group. In the second method (L ₂ ), the length of the each linkage group was multiplied by the factor (m + 1)/(m - 1), where m is the number of framework markers on the linkage groups. Finally, genome coverage c of the linkage map was estimated by calculating c = 1 - e^-2dn/L, where d is the average interval of markers, n is the number of markers, and L is the genome length estimated above.

Identification orthologous chromosomes with other fishes

The flanking sequences obtained from all SSR markers assigned to the yellowtail linkage groups were used for the BLASTn search against genomic sequences of medaka, Tetraodon, stickleback, fugu and zebrafish with a cut-off e-value of 0.01. The top hit query-subject was extracted using in-house perl script. In the case where multiple hits were obtained, we defined orthology as follows; let us consider only the first, second and third top hit, if query position of the first and second/third top hit is overlapped each other and quotient of e-value of the first hit divided by that of the second/third hit is greater than 10⁻³, the hit is considered to be an unclear orthologous pair and rejected. The substantial hits were processed for constructing Oxford grid using Grid Map ver. 3.0a (http://cbr.jic.ac.uk/dicks/software/Grid_Map/).