High-quality Chromosome-level genome assembly of male and female greater amberjack (Seriola dumerili) Authors: Lei Chen, Yinjun Jiang, Binbin Tao, Yanlong Song, Yang Huang, Guangli Li, Wei Hu, Chunhua Zhu, Ji Chen Published: 2025 ABSTRACT AbstractThe greater amberjack (Seriola dumerili) is a species of significant commercial value due to the high nutritional value and tender meat of its flesh. The existing availability of genomic data, which is incomplete and discontinuous, limits molecular biology studies on greater amberjack. Furthermore, it is hard to distinguish the genders of amberjack from their appearance, which is a major drawback for artificial propagation. In this study, chromosome-level genomes of male and female greater amberjack were assembled using PacBio HiFi and Hi-C data. The male and female genomes were anchored on 24 chromosomes, with sizes of 637 Mb and 639.4 Mb, respectively. Genome assembly BUSCO scores were 98.5% and 98.71%, respectively. The functionally annotated genes are 24,112 (87.92%) in males and 23,571 (89.03%) in females, with BUSCO assessment scores of 90.23% and 90.31%, respectively. In conclusion, we successfully assembled high-quality chromosome-level genomes for male and female greater amberjack, providing valuable information for future molecular breeding, species conservation and evolutionary research of greater amberjack.Subject terms: Comparative genomics, Genomics Background & Summary The greater amberjack (Seriola dumerili) is classified under the genus Seriola, family Carangidae, and order Perciformes. It is a medium-sized carnivorous marine fish that is distributed in subtropical waters worldwide1. The greater amberjack is also an essential breeding object for marine farming due to its fast-growing rate, flavorful flesh with abundance of unsaturated fatty acids. Recent research has prioritized artificial breeding technology development, focusing on economic traits including reproductive development2–7, immunity8,9, nutrition10,11 and stress adaptation12–14. Among these, gonadal development15 and reproductive characteristics16, as well as the development of efficient oxytocics17,18, have received significant attention. High-quality genomic data is fundamental for modern molecular research on greater amberjack. Additionally, greater amberjacks lack obvious secondary sexual characteristics even at spawning season, making gender discrimination based on morphology challenging. Therefore, constructing high-quality genomes of male and female greater amberjack will lay the foundation for elucidating the sex differentiation mechanism of greater amberjack. The earliest genomic resources (GCA_002260705.1)19,20 exhibit fragmented assemblies (contig N50: 0.25 Mb) and incomplete annotation (Table 1), while the more recent chromosome-level assembly (GCA_041682515.1)21 is limited to a single sex. The prevailing genome assembly paradigm synergistically integrates third-generation sequencing (TGS) for long-read scaffolding, next-generation sequencing (NGS) for short-read polishing of base-level inaccuracies, and high-throughput chromosome conformation capture (Hi-C) to resolve chromosomal architectures—collectively enabling the construction of complete chromosome-scale assemblies through complementary technological advantages22. Complementary long-read isoform sequencing (Iso-seq) and short-read RNA-sequencing (RNA-Seq) further enhanced annotation comprehensiveness23. Table 1Comparison of greater amberjack genome assembly metrics with previous version.GenomeThis studyGCA_002260705.1GCA_041682515.1malefemaleGenome size (Mb)637639.4677.7667.4Contig N50 (Mb)9.457.910.2526.7Scaffold N50 (Mb)28.6328.55.828.5Complete BUSCOs (%)98.598.7193.798.7Number of contig47746741,187184number of scaffolds242434,655139GC percent (%)41414141 Comparison of greater amberjack genome assembly metrics with previous version. In this study, we assembled male and female greater amberjack genomes using a combination of DNBSEQ short reads, PacBio HiFi long reads, and Hi-C scaffolding data, with Iso-seq and RNA-seq integrated for genome annotation (Fig. 1). The resulting genome assemblies spanned 637 Mb (male) and 639.4 Mb (female), both anchored to 24 chromosomes. Assembly quality was validated through multiple metrics: scaffold N50 values of 23.12 Mb (male) and 23.48 Mb (female), BUSCO (v5.8.3)24 completeness scores of 98.5%(male) and 98.7%(female), and short-read mapping rates of 99.47%(male) and 99.59%(female). Functional annotation revealed 24,112 (87.92%, male) and 23,571 (89.03%, female) protein-coding genes with homologs in public databases. We present the first high-quality, chromosome-level genome assemblies for both male and female greater amberjack. Our assemblies achieve comparable quality to the most recent chromosome-level assembly of greater amberjack-GCA_041682515.121 (Table 1, Figs. 2, and 3). These high-quality annotated genomes provide a foundation for molecular breeding, species conservation, and evolutionary studies of greater amberjack. Workflow overview for the greater amberjack chromosome-level genome assembly. Synteny plot of the male greater amberjack genome against the GCA_041682515.1. Synteny plot of the female greater amberjack genome against the GCA_041682515.1. Methods Sample collection Three male and three female wild greater amberjack were caught off the coast of Fujian Province, China. A portion of the gonads was used for paraffin sections for sex identification25 (Fig. 4). Tissue samples including muscle, heart, liver, spleen, kidney, and gonad were collected from the 6 fish. These samples were immediately frozen in liquid nitrogen for RNA sequencing. A portion of the muscle from a male and a female were for DNA sequencing, respectively. All experiments were conducted under the approval and regulations of the Institutional Animal Care and Use Committee of the Institute of Hydrobiology, Chinese Academy of Science (keshuizhuan08529). The morphological structure of the testis in the male and the ovary in the female greater amberjack. Library construction, sequencing and data preparation High-quality genomic DNA was isolated from muscle tissue for PacBio HiFi library construction. The genomic DNA extraction was performed following the standard protocol provided with the TIANamp Genomic DNA Kit (TIANGEN, Beijing, China). DNA quality was assessed by 1% agarose gel electrophoresis and NanoDrop spectrophotometry, with quantification performed using a Qubit fluorometer. An initial 150-bp short-insert library was generated via DNBSEQ for genome survey analysis (Table 2). After obtaining 95.11 Gb (male) and 94.12 Gb (female) of high-quality filtered data, genomic DNA was sheared into 13–16 kb fragments using a Megaruptor system. Size selection was conducted with a SageELF instrument to construct PacBio HiFi libraries. DNA ends were repaired and ligated to SMRTbell adapters to generate 15-kb insert libraries. Finally, 23.70 Gb (male) and 24.69 Gb (female) of HiFi reads were produced on the PacBio Sequel II platform in circular consensus sequencing (CCS) mode (Table 3). Table 2DNBseq Sequencing Statistics.DNBseqClean readsClean bases(Gb)Read Length(bp)Q20(%)GC rate (%)Male634,063,88895.1115095.57%40.85%Female627,466,57294.1215094.83%40.82% DNBseq Sequencing Statistics. Table 3HiFi Sequencing Statistics.HiFiClean readsClean bases (Gb)Clean data N50 (bp)Mean read qualityMedian read qualityMale1,627,30023.7014,56728.428.5Female1,643,90824.6915,13829.129.2 HiFi Sequencing Statistics. For Hi-C libraries, muscle cells were crosslinked with formaldehyde and digested with MboI restriction endonuclease. Digested DNA ends were labelled with biotin-14-dCTP, followed by proximity ligation, de-crosslinking, and phenol-chloroform purification. Biotinylated fragments were enriched using streptavidin magnetic beads for library preparation. Valid Hi-C read pairs were identified with HiC-Pro (v3.1.0)26, generating 109.97 Gb (male) and 104.16 Gb (female) of high-quality valid pairs (Table 4). Table 4Hi-C Sequencing Statistics.Hi-CClean reads (Gb)Clean bases (Gb)Read Length(bp)Q20(%)GC rate (%)Male733.12109.9715094.95%42.37%Female694.42104.1615095.01%42.39% Hi-C Sequencing Statistics. For comprehensive genome annotation, total RNA was extracted from five tissues (heart, liver, spleen, kidney, and gonad) using TRIzol reagent following the manufacturer’s protocol. RNA integrity was verified by Agilent Bioanalyzer 2100 analysis, followed by DNase I treatment to remove genomic DNA contamination. RNA-seq data were obtained using the Illumina NovaSeq 6000 platform (Table 5), while Iso-seq libraries were sequenced and processed using the PacBio Sequel II system (Table 6). Table 5RNA-seq Sequencing Statistics.SampleRead lengthCleand readsClean basesRaw_Q20Raw_Q30M-liver150;15095,284,04014,292,606,00097.03;96.1692.59;90.60M-spleen150;15074,494,91811,174,237,70097.01;96.4192.53;91.17M-Kidney150;15080,500,13812,075,020,70097.01;96.3292.51;90.94M-heart150;15094,013,39414,102,009,10096.97;96.4192.45;91.20M-gonad150;15081,997,42812,299,614,20096.93;96.4592.40;91.31F-liver150;15097,090,94014,563,641,00096.98;96.4892.47;91.31F-spleen150;15065,607,8089,841,171,20096.95;96.4192.51;91.26F-Kidney150;15064,871,3509,730,702,50097.01;96.6192.59;91.69F-heart150;15087,679,38813,151,908,20096.92;96.5292.36;91.47F-gonad150;15091,090,99413,663,649,10096.93;96.2692.45;91.02 RNA-seq Sequencing Statistics. Table 6ISO-seq Sequencing Statistics.SampleCCS numberCCS bases (bp)min_lenavg_lenmax_lenN50 lengthM-iso1,649,2329,351,134,810505,67058,7505,503F-iso2,171,82510,372,515,582484,77659,3455,351 ISO-seq Sequencing Statistics. Genome survey A k-mer survey analysis of the greater amberjack genome was conducted using the Jellyfish (V2.1.4)27. Initially, k-mer frequency distributions were calculated for k-values ranging from 17 to 31. The optimal k-mer size was determined to be 19 based on minimal k-mer multiplicity anomalies (Fig. 5). Genome characteristics were subsequently estimated by modeling the k-mer spectrum with GenomeScope (v2.0)28. Results indicated a male genome size of 636.21 Mb with 0.375% heterozygosity and 14.94% repetitive sequence content. The female genome was estimated at 630.31 Mb, exhibiting 0.46% heterozygosity and 14.48% repetitive sequences (Table 7). 19-Kmer analysis for the genome size evaluation in greater amberjack. Table 7Genome survey results.KmerDepthNk-merGenome size (Mb)Heterozygous rate (%)Repeat rate (%)Male19147.982,806,520,413636.210.37514.94Female19148.082,143,315,007630.310.4614.48 Genome survey results. Chromosome-level genome assembly of Greater amberjack Chromosome-level genome assembly of Greater amberjack Chromosome-level de novo assemblies of male and female greater amberjack genomes were generated by integrating PacBio HiFi long reads with Hi-C scaffolding data (Table 8). Contig-level assemblies were produced using HiFiasm (v0.16.1)29. Hi-C reads were aligned to draft contigs with BWA (v0.7.17-r1188)30 using the mem module and processed through Juicer (v1.5.6)31 to calculate mapping efficiency metrics, including the proportion of valid/invalid interaction pairs. Chromosomal scaffolding was performed with 3D-DNA (v180922)32, followed by manual curation. The resulting 637 Mb (male) and 639.4 Mb (female) assemblies resolved 24 chromosomes (Fig. 6). Contact frequency matrices were generated by partitioning genomes into 1-kb bins and quantifying Hi-C read pairs between bin pairs, visualized as interaction heatmaps (Fig. 7). Table 8Statistics for Hi-C assisted assembly.totalContig NumContig N50Scaffold NumScaffold N50Male659,581,8522819,457,1292428,629,000Female662,033,0002627,911,9852428,502,500 Statistics for Hi-C assisted assembly. Hi-C interaction heat map for genome assembly of greater amberjack. The level of interaction density is calculated according to the number of supporting Hi-C reads and visualised using a color gradient ranging from white, indicating low density, to dark red, indicating high density. The circos plot illustrates the (a) GC content, (b) gene number, (c) Copia, and (d) Gypsy in male and female greater amberjack genomes. Repeat annotation Repetitive sequences were annotated through homology-based screening and de novo prediction. For homology detection, the RepBase (v21.12)33 library was employed via RepeatMasker (v4.0.7)34 to identify known repetitive elements. De novo prediction is based on sequence comparison, which involves creating a library of de novo35 repetitive sequences using LTR_FINDER (v1.06)36 and RepeatModeler (v2.0.1)34, and then predicting them with RepeatMasker34. The Tandem Repeats Finder (v4.09)37 was used to identify tandem repeats in the genome. Comprehensive repetitive element annotations are summarized in Table 9. Ultimately, repetitive sequences constituted 23.12% of the male genome and 23.48% of the female genome. Table 9Statistics of the annotated repeat sequences.TypeMaleFemaleLength(bp)Rate(%)Length(bp)Rate(%)Class I: Retrotransposon29,177,1134.4232,016,3564.84LTR-Retrotransposon9,247,2171.4010,142,9151.53LTR/Copia304,8650.05295,3150.04LTR/Gypsy5,943,4430.906,803,9811.03LTR-other2,998,9090.453,043,6190.46Non-LTR Retrotransposon19,929,8963.0221,873,4413.30SINE994,9800.151,098,5160.17LINE18,934,9162.8720,774,9253.14Class II: DNA Transposon72,515,88610.9975,722,13511.44EnSpm/CACTA4,581,6770.694,891,3170.74Harbinger892,5020.14904,0810.14Helitron3,139,8340.481,900,2220.29MuDR825,3570.13805,0440.12Tcl/Mariner9,388,4611.429,238,0111.40hAT8,362,1351.2710,610,1261.60DNA-other45,325,9206.8747,373,3347.15Low Complexity2,403,5800.362,395,2650.36Tandem repeat27,802,3764.2124,203,7253.66Unclassified20,648,8683.1321,129,8813.19Total content152,547,82323.12155,467,36223.48Genome Length659,710,352662,152,000 Statistics of the annotated repeat sequences. Protein-coding gene prediction and annotation To predict gene structure, we combined de novo and evidence-based approaches (Table 8). De novo annotation was performed via Augustus (v3.3.2)38. Evidence-directed annotation combined transcriptomic evidence, homology evidence and cross-species protein sequences. Transcriptomic evidence (RNA-seq and Iso-seq data) aligned via Blat39 (http://genome.ucsc.edu/cgi-bin/hgBlat) (Table 10). Homology evidence from closely related species, such as Seriola rivoliana40, Seriola lalandi Dorsalis41, Danio rerio42, Trachinotus ovatus43, Seriola quinqueradiata44 and Seriola dumerili20. Integrated gene models were generated using Maker (v2.31.10)45. Table 10Predicted gene basic structure statistics.TypeCountMaleFemalegene count2742526476gene median/AVG (bp)6134/9366.256403/9617.61mRNA count3481534038mRNA median/AVG (bp)2003/2642.792091/2709.03mRNA count per gene median/AVG1/1.271/1.29CDS count3481534038CDS median/AVG (bp)1209/1679.251248/1712.91protein median/AVG (bp)403/559.75416/570.97exon median/AVG (bp)134/248.34134/248.19exon count median/AVG8/10.648/10.91UTR3 median/AVG (bp)542/903.61522.5/885.19UTR5 median/AVG (bp)111/180.22111/181.20Intron median/AVG (bp)321/816.28319/812.50Intron count median/AVG7/9.647/9.91 Predicted gene basic structure statistics. Initial filtration excluded proteins with >50% repetitive content or length <50 amino acids. The final gene set was refined through expression profiling and cross-species conservation analysis. Finally, the genes were functionally annotated in GO, KEGG46, COG, NR, swissprot47 and other databases by eggNOG48 and BLAST. This yielded 24,112 (87.92%) functionally annotated genes in male and 23,571 (89.03%) in female genomes. Table 11 displays the statistical information for functional annotation. Table 11Functional annotation of the predicted protein-coding genes.DatabaseMaleFemaleNumberPercent (%)NumberPercent (%)nr count2409987.872355788.97sw count2201980.292179982.33zebrafish count2267982.692240284.61go count1487154.221477455.8KEGG count1649360.141633661.7cog count2239881.672218683.8eggnog count2239881.672218683.8pfam count2026873.92013276allanno count2411287.922357189.03total2742526476 Functional annotation of the predicted protein-coding genes. High synteny between male and female greater amberjack genomes To compare the chromosomal structural differences between the male and female genomes, we performed a whole-genome synteny analysis between them. Using the JCVI toolkit, we first identified orthologous gene blocks between the two genomes based on gene annotation results (parameters: jcvi.compara.catalog ortholog --cscore = 0.7 --no_strip_names), and further filtered reliable syntenic blocks (parameters: jcvi.compara.synteny screen --minspan = 30). The results revealed a high degree of synteny between the male and female genomes (Fig. 8). Most chromosomal regions showed one-to-one syntenic correspondence, and no large-scale structural variations (such as chromosomal translocations, inversions, or duplications) were detected. This indicates that the macrochromosomal structure is highly conserved between male and female greater amberjack, with no obvious heteromorphic sex chromosomes formed. Whole-genome synteny between male and female greater amberjack assemblies. Data Records All raw sequencing dataset of greater amberjack have been uploaded to the National Center for Biotechnology Information (NCBI) database under the BioProject accession number PRJNA1309863. PacBio HiFi data, DNB-seq, Hi-C data, RNA-seq and Iso-seq data of male and female greater amberjack have been deposited in the SRA49. Genome assembly data of greater amberjack have been deposited in the Genome Warehouse (GWH) in National Genomics Data Center (NGDC) under accession number GWHFKZW00000000.1 (male)50 and GWHFKZX00000000.2 (female)51. The genome assembly and annotation data were deposited in the Figshare database52. Technical Validation Genome evaluation BUSCO (v5.8.3)21 is a Python-based software that evaluates the quality of transcriptome and genome assembly. In this study, the actinopterygii_odb12 library was utilized to evaluate assembly integrity. We used BUSCO to determine the completeness of the scaffold-level genome of male greater amberjack, which was 98.5%. The BUSCO assessment for the male annotated gene set was 90.23%. The BUSCO assessment for the female scaffold-level genome was 98.71%, and the annotated gene set was 90.31% (Tables 12, 13). The BUSCO evaluation shows high-quality genome assembly and gene prediction. Table 12BUSCO assessment results in the scaffold-level genome.scaffold-level genomeMaleFemaleGCA_041682515.1TypeNumberRate (%)NumberRate (%)NumberRate (%)Complete BUSCOs7,09998.507,11498.717,11398.70Complete and single-copy BUSCOs7,09398.427,10798.617,10998.64Complete and Duplicated BUSCOs60.0870.1040.06Fragmented BUSCOs (%)440.61430.60440.61Missing BUSCOs (%)640.89500.69500.69Total BUSCO groups searched7,207100.007,207100.007,207100.00 BUSCO assessment results in the scaffold-level genome. Table 13BUSCO assessment of the male and female greater amberjack genome annotation.Type (male)MaleFemaleNumberRate (%)NumberRate (%)Complete BUSCOs6,50390.236,50990.31Complete and single-copy BUSCOs6,46489.696,46889.75Complete and duplicated BUSCOs390.54410.57Fragmented BUSCOs3104.303184.41Missing BUSCOs3945.473805.27Total BUSCO groups searched7,207100.007,207100.00 BUSCO assessment of the male and female greater amberjack genome annotation. Date availability All raw sequencing dataset of greater amberjack have been uploaded to the National Center for Biotechnology Information (NCBI) database under the BioProject accession number PRJNA1309863. PacBio HiFi data (male: SRR35104904; female: SRR35104887), DNB-seq (male: SRR35104894; female: SRR35104901), Hi-C data (male: SRR35104903; female: SRR35104902), RNA-seq(male: SRR35104893, SRR35104892, SRR35104891, SRR35104890 and SRR35104889; female: SRR35104899, SRR35104898, SRR35104897, SRR35104896 and SRR35104895) and Iso-seq (male: SRR35104888,female: SRR35104900) data of greater amberjack have been deposited in the SRA. Genome assembly data of greater amberjack have been deposited in the Genome Warehouse (GWH) in National Genomics Data Center (NGDC) under accession number GWHFKZW00000000.1 (male) and GWHFKZX00000000.2 (female). The genome assembly and annotation data were deposited in the Figshare database (10.6084/m9.figshare.29380682). Acknowledgements This research was supported by the Fund of Southern Marine Science and Engineering Guangdong Laboratory (Zhanjiang) (ZJW-2019-06). Author contributions Ji Chen and Chunhua Zhu conceived the project and designed the study. Lei Chen and Ji Chen performed the data analysis and generated figures and tables. Yinjun Jiang, Binbin Tao and Yanlong Song, Yang Huang, Guangli Li performed the sampling. Lei Chen wrote the manuscript. Ji Chen and Wei Hu revised the manuscript. All authors read and approved the final manuscript. Code availability The software utilised in this study is in the public domain and is described in details in the Methods section. If no detailed parameters for the software are available, the developers’ recommended default parameters are employed. No custom script or code was employed. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Chunhua Zhu, Email: zhu860025@163.com. Ji Chen, Email: chenji@ihb.ac.cn. References Associated Data Data Citations Data Availability Statement The software utilised in this study is in the public domain and is described in details in the Methods section. If no detailed parameters for the software are available, the developers’ recommended default parameters are employed. No custom script or code was employed.