This Gadus chalcogrammus genome is another reference for molecular studies in the Gadus genus. It will be a valuable resource to conduct comparative analyses within the Gadus genus, and enhance the genomic selection process in molecular-assisted breeding.

A single female fish (93 g) was obtained from the East Sea Fisheries Research Institute in March 2018, and maintained at 8 ± 0.5°C in aerated seawater. The abdominal muscle tissues were sampled aseptically and stored in liquid nitrogen for genomic DNA extraction. The complete experimental procedure from DNA isolation to sequencing was conducted by DNA Link, South Korea (www.dnalink.com), in accordance with the product protocol.

The concentrated genomic DNA (gDNA) (24 μg) from the given samples was prepared using the DNeasy Animal Mini Kit (Qiagen, Hilden, Germany). The completely isolated gDNA was quantified using an ND-1000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA) and a Qubit fluorometer. The gDNA samples were then subjected to the following steps: fragmentation using the g-TUBE (Covaris, Woburn, MA, USA) to obtain >20-kb fragments; small-fragment filtration using 0.45X AMPure (Beckman Coulter, Brea, CA, USA); fragment end repair by ExoVII treatment; ligation of blunt adapters with double-stranded DNA fragments; attachment of primer and polymerase to SMRTbell templates (Template Prep Kit 1.0); and addition of magnetic beads. The impurities were washed out carefully using 1.0X AMPure, and only the double-stranded DNA fragments with blunt adapters were used for sequencing with P6-C4-chemistry (DNA sequencing Reagent 4.0) on the Pacific Biosciences (PacBio) sequencing platform, by capturing 1 × 240-minute-long videos of each SMRT cell. Similarly, the isolated gDNAs were also subjected to sequencing library preparation using stranded Illumina paired-end (PE) protocols (Illumina, San Diego, CA, USA). The fragmented libraries were subjected to size selection and sequencing on the Illumina HiSeq 2000 platform (Illumina).

The Illumina DNA sequences were subjected to preprocessing steps; namely, adapter trimming, quality trimming (Q20), and contamination removal. The adapter and quality trims were conducted using Trimmomatic-0.32 functions (Bolger et al., 2014), and the microbial contamination was removed using CLCMapper v4.2.0 (www.qiagenbioinformatics.com) with an in-house database. The in-house database was constructed from the bacterial, viral, and marine meta-genomes (ftp://ftp.ncbi.nlm.nih.gov/genomes/GENOME_REPORTS/prokaryotes.txt, ftp://ftp.ncbi.nlm.nih.gov/genomes/Viruses/, and https://www.ncbi.nlm.nih.gov/bioproject/PRJNA13694, respectively). All preprocessed sequences from the PE library were subjected to genome size estimation using a k-mer-based method (Shin et al., 2018). The k-mer frequencies (k-mer size = 17) were obtained using the Jellyfish v2.0 method (Marçais and Kingsford, 2011), and the genome size was calculated using the given formulae: Genome coverage depth = (k-mer coverage depth × Average read length)/(Average read length – k-mer size + 1); and Genome size = Total base number/Genome coverage depth.

Complete sequence reads were error-corrected using SMRT Analysis v2.3, and imported into a diploid-aware hierarchical genome assembler to construct the contigs from the long-sequence PacBio reads (FALCON) (Chin et al., 2016). The assembled contigs were further subjected to sequence polishing using the Quiver consensus method to reduce the base-calling errors (Chin et al., 2016). Finally, the assembled and polished contigs were assessed for completeness of the genome using BUSCO v5.0 (Simão et al., 2015). The reference BUSCO datasets are actinopterygii_odb10 and vertebrate_ odb10. The quality of the assembly was assessed by short-read mapping to the draft by BWA v0.7.15 (Li and Durbin, 2010) (Supplementary Figure 2).

The repeat regions were predicted using the de novo method, and classified into repeat subclasses. The de novo repeat prediction for G. chalcogrammus was conducted using RepeatModeler (www.repeatmasker.org/RepeatModeler/), which includes methods such as RECON (Bao and Eddy, 2002) (http://eddylab.org/software/recon/), RepeatScout (Price et al., 2005) (https://bix.ucsd.edu/repeatscout/), and TRF (Benson, 1999) (https://tandem.bu.edu/trf/trf.html). The modeled repeats were classified into subclasses by referencing the Repbase v20.08 database (www.girinst.org/repbase/) (Bao et al., 2015) and the repeats were masked using RepeatMasker v4.0.5 (www.repeatmasker.org) with RMBlastn v2.2.27⁺.

The genes from the G. chalcogrammus draft genome were predicted using an in-house gene prediction pipeline, which includes three modules: an evidence-based gene modeler, an ab initio gene modeler, and a consensus gene modeler. Finally, the functional annotation processing was performed for the consensus genes. Initially, sequenced transcriptomes obtained from the two methods (Illumina [156.9 Gb] and Iso-Seq [75.6 MB]) were assembled with Trinity(v2.2.0) (Grabherr et al., 2011) and transdecoder v5.5.0 and the proteins sequence mapped to masked G. chalcogrammus draft genome. To train the ab initio and evidence-based gene modelers [including Exonerate v.2.2.0 (Slater and Birney, 2005), AUGUSTUS v.3.1 (Stanke et al., 2006), and GENEID v.1.3 (Blanco et al., 2002)], several genomes (Supplementary Table 4) were used for prediction. Finally, the transcript models and predicted models from the evidence-based and ab initio gene modelers were subjected to the consensus gene modeler to produce the final gene and transcript models. The consensus transcripts were then subjected to functional annotation from biological databases (NCBI-NR, Swiss-Prot, Gene Ontology, and KEGG Pathway) using OmicsBox v1.2 (Götz et al., 2008).

Initially, the genome size of G. chalcogrammus was estimated to be 683.61 Mb (Figure 1A) with 42 Gb of short-read sequences (Table 1A, Supplementary Table 2) and 629.66 Mb of representative contigs from 97 Gb of error-corrected long-read sequences (Supplementary Tables 1, 3). The contigs were then assembled into 116 scaffolds in the reference draft genome (Table 1B). The N50 of the assembled genome was 27,035,343 bases, and 245 Mb (38.89%) of the assembled contigs were covered by repeats, in which the long terminal repeat (LTR) elements dominated (34%). In total, 23,353 genes were predicted from the genome, with an average size of 9261.51 bases, and 90.4% completeness on the BUSCO score (Table 1C). Homologous sequences were found for 19,760 (84.61%) genes in GenBank, and 17,259 (73.90%) genes had Gene Ontology descriptions (Table 1D). The first genome published for the Gadus genus was G. morhua (gadMor1) in 2011, as an 832-Mb genome with an N50 of 2.3 kb (scaffold N50; 0.14 Mb) (Star et al., 2011). An improved version of the same genome (gadMor2) was published in 2017 with 116 kb (scaffold N50; 1.15 Mb) (Tørresen et al., 2017), and the third NCBI version (gadMor3) was 669 Mb with a contig N50 of 1.01 Mb (scaffold N50; 28.7 Mb) and 23 chromosomes. The gadMor3 genome was used as a reference to scaffold the contigs (N50: 3.6 Mb) with the RaGOO method (Alonge et al., 2019), and 167 scaffolds were obtained with an N50 of 27.03 Mb and 23 chromosomes. The complete workflow used in this study is illustrated in Supplementary Figure 1. Overall, this genome assembly improved significantly in fragmented assembly (Figures 1B–F) and BUSCO completeness score (Table 1B). However, there is conflict in chromosome number i.e, G. morhua have 23 chromosome and G. chalcogrammus has 22 chromosomes (Supplementary Table 5). Since, the contigs scaffold well with all G. morhua 23 chromosomes, this will be improved in future version of this genome assembly (Ishii and Yabu, 1985).

The complete sequences generated in this study were deposited in the NCBI Sequence Read Archive under accession no. PRJNA736536. The assembled contigs and the annotation files (CDS, gff, repeats, and proteins) are available in the Figshare repository (https://figshare.com/s/2ff9e3a49a07c990a400) with all of the annotation details in a Readme file. The contig assembly of this draft genome was submitted to the NCBI Assembly database under accession no. JAHRIL000000000.