Samples of large yellow croaker for BAC library construction and somatic chromosome preparation were collected from the breeding section of Jimei university in Ningde, Fujian, China. Spinyhead croaker for somatic chromosome preparation were previously collected from the Sansha Bay (26°42′33″N, 119°46′49″E), Ningde city, Fujian province, China (Zhang et al., 2018). The information of the samples was shown in Supplementary Table 1 . Metaphase chromosomes were prepared by using head kidneys according to the methods described previously (Gold et al., 1990). In addition, part of the fin from each fish was fixed in absolute ethyl alcohol to extract genomic DNA.

The BAC library of large yellow croaker was constructed following the methods previously published (Luo and Wing, 2003; Luo et al., 2006). The main steps of the procedure included: i) extracting high molecular weight genomic DNA from a pooled sample of brain tissue from 2 male individuals; ii) partial digesting DNA with restriction enzyme HindIII (0.8 U/μL) (Fermentas); iii) recovering DNA fragments with 110 kb - 220 kb and 220 kb - 300 kb from gel after being separated with pulsed-field gel electrophoresis (PFGE) with a CHEF Mapper (Bio-Rad) for two times; iv) ligating the size-selected DNA fragments with HindIII-digested pHZAUBAC1 vector (Eight Star Bio-tech Co., Ltd., Wuhan, China); v) transforming the ligation products into Escherichia coli strain DH10B T1 Phage-Resistant (Invitrogen); vi) cultivating the transformants on LB plates with appropriate chloramphenicol, X-gal, and IPTG; vii) picking clones and arraying them into 384-well plates containing LB media; viii) incubating the clones in plates for 16 h and freezing them at -80°C.

Eighty clones were randomly selected, and the BAC plasmids were extracted by conventional alkaline lysis method and digested with I-Sce I enzyme (Fermentas) as described previously (Shi et al., 2011). The size of the inserted fragments and the nulling rate were examined by pulsed-field gel electrophoresis. The genome coverage of the BAC library was estimated based on the size of the genome, the number of clones in the library, the average insert size and the null rate.

Totally, 590 BAC clones were selected randomly from the BAC library, and paired end sequenced by Sanger sequencing with Applied Biosystems 3730 DNA Analyzer. The quality of BAC end sequences (BESs) was evaluated with Phred (http://www.phrap.org/) after removing the vector sequence by Seqclean (https://sourceforge.net/projects/seqclean). The resulted high-quality BESs were aligned to the reference genome of large yellow croaker (GenBank assembly accession: GCA_003711585.2) by Blast (Altschul et al., 1997), with an E-value threshold of 1 × 10⁻⁵. The high-quality BESs were also aligned to the assembled genome of spinyhead croaker [SCMI00000000.1 (Cai et al., 2019)]. The results of in silico mapping were visualized with Mapchart v2.30.

The insert sequences of mapped BAC clones were deduced from the genomic sequences between the paired BESs, and subjected to estimate the size and the repetitive sequence content. The repetitive sequences were found by using Repeatmasker (http://www.repeatmasker.org/RepeatMasker/). For each chromosome, 3 candidate chromosome-specific BAC clones were selected for further validation according to the following criteria: 1) with insert size around the estimated average; 2) containing as few repetitive sequences as possible.

The correspondence between the selected BAC clones and pseudochromosomes was verified by amplifying fragments of the expected insert sequence. A pair of primers for each BAC clone was designed using Primer Premier 5 software according to the deduced sequence of BAC insert. The primers were synthesized by BGI Genomics Inc. PCR was carried out with a commercial kit (Tiangen Biotech) according to the manual. The conditions of PCR amplification were as following: pre-denaturation at 94 ℃ for 3 min; 40 cycles of denaturation at 94 ℃ for 30 s, annealing at Tm for 30 s, and extension at 72 ℃ for 30 s. The number and the size of PCR amplification products were checked with 1% agarose gel electrophoresis.

After PCR validation, the candidate chromosome specific BAC clones were further tested by FISH. The manipulation of FISH was essentially the same as previously described (Zhang et al., 2018). BAC plasmid DNA was extracted from Escherichia coli using a BAC/PAC DNA extraction kit (Omega). Probes labeled with biotin-11-dUTP or digoxigenin-11-dUTP were prepared with the BAC plasmid DNA by using nick translation kit (Roche). The labeled probe was added into hybridization solution (50% formamide deionized/2 × SSC) and denatured. Chromosome preparation slides were denatured and dehydrated in gradient ethanol solution. The probe solution hybridized with denatured chromosome on the slides at 37°C overnight. After stringent wash, biotin-labeled probes were detected with Avidin-Alexa fluor-488 (Invitrogen), and digoxigenin-labeled probes were detected with Anti-digoxi-Rhodamine (Invitrogen). And then, slide was dyed by DAPI (4’,6-diamidino-2-phenylindole). Metaphase spreads were observed using an Olympus BX53 epifluorescence microscope. Gray-scale images were captured for each color channel with a digital image capture system (Olympus DP 80), and then merged with Cellsens digital image software (Olympus). Background subtraction and image feature intensification were conducted by using the “curve” and “layer overlay” command in Adobe photoshop CC 2017. Chromosome paring and alignment were performed by using the “cut and paste” and “rotation” commands in Adobe photoshop CC 2017.

After PCR and FISH validation, 1 to 3 chromosome-specific BAC clones for each pseudochromosome were selected to formulate a probe cocktail. All selected BAC clones were divided into two groups labeled with biotin-11-dUTP or digoxigenin-11-dUTP, and then pooled together for dual-color FISH on metaphase chromosome spreads of large yellow croaker. The chromosomes were identified according to signal pattern, length, and morphology the chromosomes. Chromosome lengths were measured by using “measurement” tool in Adobe photoshop and used to estimate the relative lengths.

Spinyhead croaker is a close relative of large yellow croaker with a fused Y chromosome, which was speculated to originate from the fusion of Chr. 1 and Chr. 7 by a bioinformatic method (Xiao et al., 2020). To explore the Y origin of spinyhead croaker and to test cross-species adaptability of the developed chromosome-specific BAC-FISH probes of large yellow croaker, two BAC clones corresponding to Chr.7 and 18S rDNA corresponding to Chr.1 were labeled with biotin-11-dUTP and digoxigenin-11-dUTP, respectively; and then pooled together for a dual-color FISH on metaphase chromosome spreads of spinyhead croaker.

We constructed a BAC library of large yellow croaker with genomic DNA extracted from brain tissues. The library contained a total of 41,472 BAC clones. To evaluate the quality of library, eighty clones were randomly selected and examined with pulsed field gel electrophoresis ( Supplementary Figure 1 ). Among the eighty clones, there was only one empty clone, indicating an empty load rate of 1.25%. The size of the insert fragments ranged from 120 kb to 150 kb, with an average of 133 kb ( Supplementary Figure 2 ). Based on the genome size, the number of clones in the library, the average size of inserted fragments, and the rate of empty load, the library was estimated to cover the genome of large yellow croaker for 8.1 times.

A total of 590 BAC clones were randomly selected from the constructed BAC library for paired end sequencing. After processing, 1100 high-quality BESs were obtained, covering 0.13% of genome ( Table 1 ). By using BLAST, there were a total of 435 BAC clones with the paired BESs aligned to the same pseudochromosome. The genomic positions of the BES were recorded and used to estimate the expected length of the BAC insert. A total of 342 BAC clones with putative insert sizes around the average (133 kb ± 30 kb) were subjected to further analysis ( Supplementary Table 2 ). Thus, 11 to 32 candidate chromosome-specific BAC clones were obtained for each chromosome, with an average of 18 ( Figure 1 ).

The repeat content of BAC insertions was estimated to range from 4.24% to 93.5%, with an average of 18.8% ± 10.9% ( Supplementary Table 2 ). Three BAC clones for each pseudochromosome, totally 72 BAC clones, with low content of repeat sequence were selected for further PCR and FISH validaton, which were highlight in color in the ideogram of pseudochromosomes ( Figure 1 ; Supplementary Table 2 ).

For PCR validation, primers were designed according to putative sequence of the insertions ( Supplementary Table 3 ). The PCR results showed that single product with expected length was amplified with each primer pairs. The representative results of PCR were shown in Supplementary Figure 3 . Thus, the accuracy of in silico location of 72 selected BAC clones was preliminarily proved with PCR.

For FISH validation, 72 selected BAC clones were labeled with biotin or digoxigenin. Probes from the same pseudochromosome were pooled for dual-color FISH on metaphase chromosome spreads to examine their location and specificity ( Supplementary Figure 4 ). Figure 2 summarized the FISH results for each pseudochromosome. For 19 chromosome, all 3 selected BAC clones passed FISH validation as they generated 3 specific FISH signals with similar collinearity to the pseudochromosomes. For the other 5 chromosomes including Chr.5, Chr.6, Chr.18, Chr.23 and Chr.24, 2 out of 3 selected BAC clones passed FISH validation. Totally, 67 BAC clones could produce specific signals at the expected positions ( Supplementary Table 4 ). Thus, all chromosomes of large yellow croaker could be identified individually with chromosome-specific FISH probes ( Figure 2 ; Supplementary Figure 4 ).

FISH results also showed that the distance between FISH signals on chromosome was related to its distribution on pseudochromosome. For example, on Chr.17, 102-K23 and 101-N03 were located in adjacent positions on both pseudochromosome and metaphase chromosome, while 101-H16 was distributed at a more distant position ( Figure 2 ).

On the basis of chromosome specific BACs, we developed a probe cocktail for dual-color FISH to identify the whole set of chromosomes in metaphase cells of large yellow croaker. For each chromosome, 1 to 3 chromosome-specific BACs were selected to produce probes and mixed into two pools according to the scheme shown in Figure 3A . After FISH with the cocktails, red and green signals were generated and formed distinct patterns for each chromosome. Combining FISH signal patterns and chromosome morphology, chromosomes corresponding to pseudochromosomes were all identified ( Figure 3B ).

Relative lengths of chromosome were estimated and compared them to the lengths of pseudochromosomes ( Supplementary Table 5 ). The results of correlation analysis showed that they were only weakly correlated ( Supplementary Figure 5 ). Karyotyping was performed by arranging chromosomes in descending order of relative length, linking chromosomes in classic karyotyping and pseudochromosomes in assembled genome ( Supplementary Figure 6 ; Supplementary Table 5 ).

By using BLAST, 218 BAC clones were anchored to pseudochromosomes of spinyhead croaker genome ( Supplementary Table 6 ; Supplementary Figure 7 ). Two BAC clones corresponding to Chr.7 were selected to prepare probes for a dual-color FISH on metaphase spreads of spinyhead croaker, along with probes derived from 18S rDNA that corresponds to Chr.1 (Zhang et al., 2018). In females, Chr.7-specific BAC clones and 18S rDNA were located on two pairs of chromosomes, respectively. While in males, they were located on three chromosomes including a male-specific metacentric chromosome, where Chr. 7-specific BAC clones and 18S rDNA were located, respectively ( Figure 4 ). These results provided cytogenetic evidence that the Y chromosome of spinyhead croaker originated from the fusion of Chr. 1 and Chr. 7.