Chromosome-level genome assembly and annotation of the kuruma shrimp Marsupenaeus japonicus Authors: Yiming Wei, Hao Xu, Zhixiong Zhou, Panpan Wang, Wenzhi Cheng, Huiyang Huang, Shaoxiong Ding, Yong Mao Published: 2025 ABSTRACT The kuruma shrimp Marsupenaeus japonicus is one of the most economically important shrimp species in the world. Here, we constructed a chromosome-level genome assembly of M. japonicus by combining PacBio long reads, Illumina short reads and Hi-C scaffolding. The genome size was 1.64 Gb with a scaffold N50 length of 40.61 Mb, and 97.83% (1.60 Gb) of the assembled sequences were anchored to 43 chromosomes. The genome contained 62.79% repeat sequences and 21,172 protein-coding genes, of which 83.20% were functionally annotated. The completeness of M. japonicus genome assembly is highlighted by a BUSCO score of 91.0%. Evolutionary analysis indicated that M. japonicus was closely related to Litopenaeus vannamei and Penaeus monodon, with an estimated divergence time from their common ancestor of 88.33 million years ago. In sum, the chromosome-level genome assembly and annotation provide fundamental resources for genetic breeding and molecular mechanism related studies of M. japonicus. Background & Summary The kuruma shrimp Marsupenaeus japonicus is the only species in Marsupenaeus genus, which belongs to the family Penaeidae and the order Decapoda1. It is widely distributed across the South China Sea, Japan, the Malay Archipelago, and from the western Indian Ocean to eastern South Africa, primarily inhabiting bays and inland seas2. The M. japonicus is divided into two morphologically similar varieties (Form I and II) based on carapace banding patterns1, with most experimental research conducted on variety I. Due to its desirable traits such as good taste, high economic value, rapid growth rate, and remarkable low oxygen tolerance that allows survival out of water for over 10 hours3,4, M. japonicus has become one of the most economically important penaeid shrimp species in global fisheries and aquaculture5. Artificial spawning and hatching were first achieved in 19336, and industrial-scale larval rearing technology was established in the 1960s. In 2023, annual production in China reached 45,968 tons7. Crustaceans serve dual ecological and economic roles, encompassing not only a significant number of commercially valuable species that provide high-quality protein sources, but also function as critical bioindicators of aquatic ecosystem health and keystone species in marine trophic networks8. But limited by sequencing technology9, crustacean’s genomics research lags other species and started to break through until 201910. Decapods are the highest and largest orders of crustaceans which have relatively developed body structure and complex physiological function11. The genomes of Decapods are more difficult to assemble because of their large size, high heterozygosity, and high level of repeated sequences. In 2019, the first whole-genome sequence of penaeid shrimp was reported, presenting a 1.66 Gb (scaffold N50 605.56 kb) genome of the Pacific white shrimp Litopenaeus vannamei12. Then the genome assembly for other economically invertebrates are successively carried out, such as the black tiger shrimp Penaeus monodon13, Chineses shrimp Fenneropenaeus chinensis14, and Portuguese oyster Crassostrea angulata15. Although a chromosome-level genome of M. japonicus has been previously published16,17, several critical limitations necessitate a new, high-quality reference genome. First, the published genome assembly (1.54 Gb) was anchored to only 42 pseudo-chromosomes¹⁵, whereas cytogenetic studies consistently indicate that the species has a diploid chromosome number of 2n = 8618–20. This discrepancy suggests possible errors in chromosome-scale scaffolding or unresolved haplotypic duplication. Second, the contiguity of previous assemblies remains suboptimal, with contig N50 values of 229.97 kb17, limiting their utility for studying genomic structural variation and repetitive regions. To address these gaps, we generated a new chromosome-level genome assembly using a combination of PacBio long-read sequencing, Illumina short-reads, and Hi-C scaffolding (Fig. 1). This high-quality M. japonicus reference genome will provide an important genomic resource for genetic breeding and molecular mechanism related studies.Fig. 1Characterization of assembled M. japonicus genome. From the outside to inside rings: (I) Chromosomes (Mb in length); (II) Gene density distribution within 100 kb nonoverlapping windows; (III) Genome GC content distribution within 100 kb nonoverlapping windows; (IV) Distribution of repeat sequences within 100 kb nonoverlapping windows. Characterization of assembled M. japonicus genome. From the outside to inside rings: (I) Chromosomes (Mb in length); (II) Gene density distribution within 100 kb nonoverlapping windows; (III) Genome GC content distribution within 100 kb nonoverlapping windows; (IV) Distribution of repeat sequences within 100 kb nonoverlapping windows. Methods Ethics statement The M. japonicus specimen utilized in this study was collected from Xiamen City, Fujian Province, China. All experimental procedures were conducted in strict compliance with the Guidelines for The Care and Use of Laboratory Animals and were approved by the Laboratory Animal Ethics Committee of the School of Ocean and Earth Sciences at Xiamen University. Biological material and genomic DNA extraction One healthy adult male M. japonicus (variety I) was collected from a natural marine environment in Xiamen, China. Muscle tissue was dissected, rapidly frozen in liquid nitrogen, and subsequently used for DNA extraction. Genomic DNA (gDNA) for both PacBio and Illumina paired-end library construction was isolated using the QIAGEN DNeasy Blood & Tissue Kit (Shanghai, China) according to the manufacturer’s instructions. DNA quality was evaluated by 1.0% agarose gel electrophoresis, while purity and concentration were determined using a Nanodrop system (OD 260/280 ratio) and the Qubit® DNA Assay Kit on a Qubit® 3.0 Fluorometer (Invitrogen, USA), respectively. Library construction and sequencing The gDNA was fragmented with a Covaris® g-TUBE® Shearing Device, and the sheared fragments were pureed using 0.45X AMPure beads. A pair-end library with an insert size of 350 bp was prepared and sequenced on the Illumina HiSeq platform, while Pacbio libraries were generated and sequenced on the Pacbio Sequel II platform at Novogene (Tianjin, China). After quality filtering, approximately 285 Gb of Illumina short-insert reads and 265 Gb of PacBio reads were obtained, yielding a contig-level assembly coverage of 107.33 × . Additionally, a Hi-C library was constructed following the Rao et al.21 methodology and sequenced on the Illumina HiSeq-2500 platform, yielding approximately 280 Gb of data for chromosome-level assembly (Table 1).Table 1Sequencing data used for the genome assembly of M. japonicus.LibraryInsert sizesRaw data (Gb)Sequence coverage (×)Illumina350 bp285115.43PacBio15–20 kb265107.33Hi-c350 bp280113.40Total—830336.16 Table 1Sequencing data used for the genome assembly of M. japonicus.LibraryInsert sizesRaw data (Gb)Sequence coverage (×)Illumina350 bp285115.43PacBio15–20 kb265107.33Hi-c350 bp280113.40Total—830336.16 Sequencing data used for the genome assembly of M. japonicus. Genome assembly Before genome assembly, the genome characteristics of M. japonicus were assessed through K-mer analysis22 based on Illumina sequencing data. Using SOAPdenovo23, a K-mer frequency distribution (K = 17) with a total count of 117,488,892,845 and a dominant peak at depth 47 was obtained. This analysis estimated the genome size to be approximately 2,469.11 Mb. Additionally, the repeat sequence content and heterozygosity rate were calculated as 77.97% and 0.49%, respectively (Table 2).Table 2K-mer frequency and genome size evaluation of M. japonicus genome.K-mer numberK-mer DepthGenome Size (Mb)Heterozygous Ratio (%)Repeat (%)117,488,892,845472,469.110.4977.97 Table 2K-mer frequency and genome size evaluation of M. japonicus genome.K-mer numberK-mer DepthGenome Size (Mb)Heterozygous Ratio (%)Repeat (%)117,488,892,845472,469.110.4977.97 K-mer frequency and genome size evaluation of M. japonicus genome. The draft genome was de novo assembled from PacBio long reads using WTDBG (v2)24 with the following parameters: wtdbg2 -A -k 0 -p 21 --node-drop 0.20 --node-len 1536 --node-max 200 -s 0.05 -e 3 -t 10”. Subsequent polishing was performed with NextPolish (v1.1.0)25, utilizing both Pacbio and Illumina sequencing data. Due to the high proportion of homozygous SNPs, we used NextPolish (v1.1.0)25 to correct the genome again based on all the Illumina sequencing data. The final contig-level genome size was 1.64 Gb, including 15,814 contigs in total, with a Contig N50 of 324.18 kb (Table 3).Table 3Statistics for Hi-C assisted assembly.TotalContig NumContig N50 (bp)Scafold NumScafold N50 (bp)ProportionGC-percentHi-C assisted pre-assembly1,638,807,78915,814324,180————Hi-C-assisted assembly1,603,178,92711,385324,1824340,613,94297.8334.15 Table 3Statistics for Hi-C assisted assembly.TotalContig NumContig N50 (bp)Scafold NumScafold N50 (bp)ProportionGC-percentHi-C assisted pre-assembly1,638,807,78915,814324,180————Hi-C-assisted assembly1,603,178,92711,385324,1824340,613,94297.8334.15 Statistics for Hi-C assisted assembly. Hi-C scaffolding To achieve a chromosome-level genome assembly, a hybrid strategy integrating both second- and third-generation sequencing technologies was employed. The third-generation assembly results were used as the draft to construct contigs, while second-generation reads assisting in error correction. Finally, genome was assembled to the chromosome level using the Hi-C library with ALLHiC26. After chromosome-level scaffolding, we obtained 43 chromosomes with a total length of approximately 1.60 Gb (97.83% of the contig assembly), resulting in a scaffold N50 of 40.61 Mb (Table 3). The length of 43 chromosome-level sequences ranged from 4.11 Mb to 58.06 Mb, with an average of 37.28 Mb (Fig. 2, Table 4).Fig. 2Genome-wide Hi-C interaction mapping of chromosome sections. The color indicates chromatin interactions quantified based on the count of Hi-C reads.Table 4Summary information for each chromosome of M. japonicus.Chromosome IDLength (bp)Number of Contigschr158,060,486337chr251,474,019307chr353,607,860310chr449,883,850327chr539,939,977267chr650,291,297360chr747,532,289345chr842,719,050292chr945,623,537293chr1040,674,941255chr1145,264,469281chr1247,194,139336chr1349,812,099326chr1436,801,798152chr1538,387,483260chr1640,836,015259chr1732,789,775187chr1841,916,231277chr1940,613,942266chr2044,531,641271chr2134,153,371265chr2242,074,583298chr2340,855,826337chr2440,132,394305chr2539,342,561296chr2636,734,213274chr2736,626,005187chr2833,858,489301chr2935,306,632294chr3037,827,447281chr3131,552,034208chr3229,886,336223chr3327,189,388211chr3425,894,584244chr3529,967,719224chr3625,358,992230chr3726,833,977215chr3824,283,769219chr3925,785,014262chr4026,887,208243chr4127,611,239214chr4222,949,287236chr434,112,961110Total1,603,178,92711,385 Genome-wide Hi-C interaction mapping of chromosome sections. The color indicates chromatin interactions quantified based on the count of Hi-C reads. Table 4Summary information for each chromosome of M. japonicus.Chromosome IDLength (bp)Number of Contigschr158,060,486337chr251,474,019307chr353,607,860310chr449,883,850327chr539,939,977267chr650,291,297360chr747,532,289345chr842,719,050292chr945,623,537293chr1040,674,941255chr1145,264,469281chr1247,194,139336chr1349,812,099326chr1436,801,798152chr1538,387,483260chr1640,836,015259chr1732,789,775187chr1841,916,231277chr1940,613,942266chr2044,531,641271chr2134,153,371265chr2242,074,583298chr2340,855,826337chr2440,132,394305chr2539,342,561296chr2636,734,213274chr2736,626,005187chr2833,858,489301chr2935,306,632294chr3037,827,447281chr3131,552,034208chr3229,886,336223chr3327,189,388211chr3425,894,584244chr3529,967,719224chr3625,358,992230chr3726,833,977215chr3824,283,769219chr3925,785,014262chr4026,887,208243chr4127,611,239214chr4222,949,287236chr434,112,961110Total1,603,178,92711,385 Summary information for each chromosome of M. japonicus. Repeat annotation Repeat annotation was performed using an integrated approach combining homology-based alignment and de novo prediction. Tandem repeats were identified ab initio with TRF v4.09.127. The homolog prediction commonly used Repbase28 database employing RepeatMasker (open-4.0.9) software and its in-house scripts (RepeatProteinMask) with default parameters to extracted repeat regions. De novo repetitive elements database were constructed using LTR_FINDER (v1.0.5)29, RepeatScout (v1.0.5)30, RepeatModeler31, retaining sequences longer than 100 bp and containing less than 5% gap ‘N’ to form a raw transposable element (TE) library. A custom library (a combination of Repbase and our de novo TE library which was processed by UCLUST to yield a non-redundant library) was supplied to RepeatMasker for DNA-level repeat identification. The results indicated that repetitive sequences constitute 62.79% of the M. japonicus genome. Long terminal repeats (LTRs) were the most abundant, covering over 924 Mb (56.41% of the genome). Other major repeat classes included long interspersed nuclear elements (LINEs), comprising 28.77 Mb (1.76%), and DNA transposons, accounting for 10.52 Mb (0.64%) (Table 5).Table 5Classification of repetitive sequences of M. japonicus genome.TypeDenovo + RepbaseTE ProteinsCombined TEsLength(bp)Percentage (%)Length(bp)Percentage (%)Length(bp)Percentage (%)DNA10,325,2830.63249,9450.0210,522,9910.64LINE6,496,9670.4024,647,7191.5028,766,1051.76SINE57,2000.000057,2000.00LTR923,839,81756.375,160,7940.31924,499,84656.41Unknown22,587,7021.380022,587,7021.38Total950,976,24458.0330,056,7511.83955,035,38258.28 Table 5Classification of repetitive sequences of M. japonicus genome.TypeDenovo + RepbaseTE ProteinsCombined TEsLength(bp)Percentage (%)Length(bp)Percentage (%)Length(bp)Percentage (%)DNA10,325,2830.63249,9450.0210,522,9910.64LINE6,496,9670.4024,647,7191.5028,766,1051.76SINE57,2000.000057,2000.00LTR923,839,81756.375,160,7940.31924,499,84656.41Unknown22,587,7021.380022,587,7021.38Total950,976,24458.0330,056,7511.83955,035,38258.28 Classification of repetitive sequences of M. japonicus genome. Protein-coding gene prediction and annotation Structural annotation of the M. japonicus genome incorporates homology-based prediction, ab initio prediction and RNA-Seq assisted prediction, was used to annotate gene model. Sequences of homologous proteins were downloaded from Ensembl/NCBI/others. Protein sequences were aligned to the genome using TblastN (v2.2.26; E-value ≤ 1e−5)32 based on Drosophila melanogaster, Daphnia pulex, L. vannamei and Portunus trituberculatus, and then the matching proteins were aligned to the homologous genome sequences for accurate spliced alignments with GeneWise (v2.4.1)33. For gene predication based on Ab initio, we used Augustus (v3.2.3)34, Geneid (v1.4)35, Genescan (v1.0)36, GlimmerHMM (v3.04)37 and SNAP (https://github.com/KorfLab/SNAP) in our automated gene prediction pipeline. The non-redundant reference gene set was generated by merging genes predicted by these methods with EvidenceModeler (EVM, v1.1.1)38 using PASA (Program to Assemble Spliced Alignment, v2.4.1)39 terminal exon support and including masked transposable elements as input into gene prediction. And to optimize the genome annotation, a full-length transcriptome40 was used to supplement genes set. To predict the function of genes, we used Blastp (with a threshold of E-value ≤ 1e−5)41 to align the protein sequences to the Swiss-Prot in order to gain the best match. Then the motifs and domains were annotated using InterProScan (v5.31)42 by searching against publicly available databases, including ProDom, PRINTS, Pfam, SMRT, PANTHER and PROSITE. We predicted the proteins function by transferring annotations from the closest BLAST (E-value ≤ 1e−5) in the Swissprot43 database and DIAMOND (v0.8.22)44 in the NR database. We also mapped gene sets to a KEGG45 pathway and identified the best match for each gene. In total, we successfully predicted 21,172 protein-coding genes within the genome. These predicted genes displayed an average coding sequence length of 1,385.78 bp, an average gene length of 13,876.34 bp, and an average exon number of 6.48 (Table 6). Further, 17,615 genes, which account for 83.20% of the total number of predicted genes, were successfully assigned with at least one functional annotation (Table 7).Table 6Basic statistical results of gene prediction.Gene setNumberAverage gene length (bp)Average CDS length (bp)Average exon number per geneAverage exon length (bp)Average intron length (bp)De novoAugustus146,7905,203.14603.223.20188.502,090.81GlimmerHMM202,9327,140.47443.213.01147.423,337.75SNAP162,1614,186.77404.882.69150.492,237.13Geneid43,29624,918.08618.894.64133.326,671.51Genscan60,61216,628.631,023.084.69218.254,231.87HomologPtr21,5162,342.37629.952.44258.261,189.87Dme5,5617,609.321,070.355.09210.371,599.58Dpu9,5006,309.55893.044.27209.291,657.92Lva19,7866,532.661,149.734.80239.601,417.10RNAseqPASA22,89428,069.641,789.058.04222.403,730.83Transcripts20,23128,727.242,643.937.60348.113,955.00EVM125,0426,125.45629.693.31189.982,374.58Pasa-update124,7116,256.07634.033.32190.762,419.35Final set21,17213,876.341,385.786.48213.852,279.19Table 7Functional annotation statistics.Gene numberPercent (%)Total21,172—Swissprot13,06761.70NR17,51482.70KEGG13,34463.00InterPro14,26367.40GO9,49044.80Pfam12,27358.00Annotated17,61583.20Unannotated3,55716.80 Table 6Basic statistical results of gene prediction.Gene setNumberAverage gene length (bp)Average CDS length (bp)Average exon number per geneAverage exon length (bp)Average intron length (bp)De novoAugustus146,7905,203.14603.223.20188.502,090.81GlimmerHMM202,9327,140.47443.213.01147.423,337.75SNAP162,1614,186.77404.882.69150.492,237.13Geneid43,29624,918.08618.894.64133.326,671.51Genscan60,61216,628.631,023.084.69218.254,231.87HomologPtr21,5162,342.37629.952.44258.261,189.87Dme5,5617,609.321,070.355.09210.371,599.58Dpu9,5006,309.55893.044.27209.291,657.92Lva19,7866,532.661,149.734.80239.601,417.10RNAseqPASA22,89428,069.641,789.058.04222.403,730.83Transcripts20,23128,727.242,643.937.60348.113,955.00EVM125,0426,125.45629.693.31189.982,374.58Pasa-update124,7116,256.07634.033.32190.762,419.35Final set21,17213,876.341,385.786.48213.852,279.19 Basic statistical results of gene prediction. Table 7Functional annotation statistics.Gene numberPercent (%)Total21,172—Swissprot13,06761.70NR17,51482.70KEGG13,34463.00InterPro14,26367.40GO9,49044.80Pfam12,27358.00Annotated17,61583.20Unannotated3,55716.80 Functional annotation statistics. Annotation of non-coding RNA genes To predict the tRNAs, we used the program tRNAscan-SE (v1.3.1)46. For rRNAs are highly conserved, we choose relative species rRNA sequence as references, predict rRNA sequences using Blast. The miRNAs and snRNAs were identified by Infernal (v1.1.2)47 against the Rfam (v14.1) database with default parameters. As a result, we annotated 195 rRNAs, 2,350 tRNAs, 36,384 miRNAs, and 3,741 snRNAs (Table 8).Table 8Statistics of non-coding RNA annotation.TypeCopyAverage Length (bp)Total Length (bp)% of genomemiRNA36,384126.524,603,3180.2800tRNA2,35075.29176,9430.0108rRNArRNA195142.5927,8050.001818 S78205.8116,0530.001028 S1097.909790.00015.8 S139390.00005 S106101.2610,7340.0007snRNAsnRNA3,741177.35663,4550.0405CD-box3,116168.73525,7730.0321HACA-box383243.3493,1990.0057splicing122144.5417,6340.0011scaRNA120223.7426,8490.0016 Table 8Statistics of non-coding RNA annotation.TypeCopyAverage Length (bp)Total Length (bp)% of genomemiRNA36,384126.524,603,3180.2800tRNA2,35075.29176,9430.0108rRNArRNA195142.5927,8050.001818 S78205.8116,0530.001028 S1097.909790.00015.8 S139390.00005 S106101.2610,7340.0007snRNAsnRNA3,741177.35663,4550.0405CD-box3,116168.73525,7730.0321HACA-box383243.3493,1990.0057splicing122144.5417,6340.0011scaRNA120223.7426,8490.0016 Statistics of non-coding RNA annotation. Evolutionary analysis and divergence time prediction A total of 450 strictly single-copy orthologous genes was identified by OrthoFinder (v2.5.4)48 from M. japonicus, L. vannamei, P. monodon, P. trituberculatus, Armadillidium vulgare, Hyalella Azteca, and D. pulex. Multiple sequence alignment was performed on these single-copy orthologous gene families with MUSCLE49, and a phylogenetic tree was reconstructed under the Maximum Likelihood method using RaxML50 (Fig. 3a).Fig. 3Evolutionary analysis and divergence time prediction of M. japonicus. (a) Phylogenetic tree of M. japonicus and 6 other species. (b) Estimates of divergence time between M. japonicus and six other species are shown in blue. Gene family expansions and contractions are also presented, with green and red numbers indicating expanded and contracted gene families, respectively. Evolutionary analysis and divergence time prediction of M. japonicus. (a) Phylogenetic tree of M. japonicus and 6 other species. (b) Estimates of divergence time between M. japonicus and six other species are shown in blue. Gene family expansions and contractions are also presented, with green and red numbers indicating expanded and contracted gene families, respectively. Gene family expansion and contraction were analyzed using CAFE. The results indicated that 6,804 gene families expanded and 938 contracted in M. japonicus (Fig. 3b). In comparison, the closely related species L. vannamei and P. monodon exhibited greater expansions, with 9,511 and 9,142 expanded families, alongside 611 and 816 contracted families, respectively. Evolutionary analysis further revealed that M. japonicus was closely related to L. vannamei and P. monodon, with these lineages diverging from a common ancestor approximately 88.33 million years ago (Mya) (Fig. 3b). Data Records All the raw sequencing data used for genome assembly were deposited in the NCBI Sequence Read Archive (SRA) database under the accession number SRP58162351. The chromosome-level assembly of the M. japonicus genome was deposited in the European Nucleotide Archive (ENA) under the accession number PRJEB10264252. The genome annotation files were deposited at the figshare (10.6084/m9.figshare.28874273.v1)53. Technical Validation The BUSCO was used to evaluate the quality of the genome assembly. We employed BUSCO (v3.0.2)54 and CEGMA55 to evaluate genome assembly completeness. The final genome assembly showed a BUSCO completeness of 91.0%, consisting of 914 (90.2%) single-copy BUSCOs, 8 (0.8%) duplicated BUSCOs, 26 (2.6%) fragmented BUSCOs, and 65 (6.4%) missing BUSCOs (Table 9). The result of CEGMA showed that 91.13% (226) of 248 CEGs had been assembled.Table 9Statistical result of BUSCO evaluation results of genome assembly.NumberPercentage (%)Complete BUSCOs92291.0Complete and single-copy BUSCOs91490.2Complete and duplicated BUSCOs80.8Fragmented BUSCOs262.6Missing BUSCOs656.4Total BUSCO groups searched1,013100 Table 9Statistical result of BUSCO evaluation results of genome assembly.NumberPercentage (%)Complete BUSCOs92291.0Complete and single-copy BUSCOs91490.2Complete and duplicated BUSCOs80.8Fragmented BUSCOs262.6Missing BUSCOs656.4Total BUSCO groups searched1,013100 Statistical result of BUSCO evaluation results of genome assembly. Then we used the sequence alignment method to evaluate possible contamination and the completeness of the genome assembly. The BWA (0.7.8-r455)56 was used to map Illumina short reads to the assembled genome with default parameters. The mapping results indicate that 92.98% of reads were successfully mapped, and the coverage rate was approximately 82.92%. The relationship between sequencing depth and GC content distribution calculating with 10k windows revealed no deviations from the expected levels.