Dataset for genome sequencing and de novo assembly of the Vietnamese bighead catfish (Clarias macrocephalus Günther, 1864)

Freshwater catfish of the genus Clarias, known as the airbreathing catfish, are widespread and important for food security through small scale inland fisheries and aquaculture. Limited genomic data are available for this important group of fishes. The bighead catfish (Clarias macrocephalus) is a commercial aquaculture species in southeast Asia used for aquaculture and threatened in its natural environment through habitat destruction, over-exploitation and competition from other introduced species of Clarias. Despite its commercial importance and threats to natural populations, public databases do not include any genomic data for C. macrocephalus. We present the first genomic data for the bighead catfish from Illumina sequencing. A total of 128 Gb of sequence data in paired-end 150 bp reads were assembled de novo, generating a final assembly of 883 Mbp contained in 27,833 scaffolds (N50 length: 80.8 kbp) with BUSCO completeness assessments of 96.3% and 87.6% based on metazoan and Actinopterygii ortholog datasets, respectively. Annotation of the genome predicted 21,124 gene sequences, which were assigned putative functions based on homology to existing protein sequences in public databases. Raw fastq reads and the final version of the genome assembly have been deposited in the NCBI (BioProject: PRJNA604477, WGS: JAAGKR000000000, SRA: SRR11188453). The complete C. macrocephalus mitochondrial genome was also recovered from the same sequence read dataset and is available on NCBI (accession: MT109097), representing the first mitogenome for this species. Lastly, we find an expansion of the mb and ora1 genes thought to be associated with adaptations to air-breathing and a semi-terrestrial life style in this genus of catfish.


a b s t r a c t
Freshwater catfish of the genus Clarias , known as the airbreathing catfish, are widespread and important for food security through small scale inland fisheries and aquaculture. Limited genomic data are available for this important group of fishes. The bighead catfish ( Clarias macrocephalus ) is a commercial aquaculture species in southeast Asia used for aquaculture and threatened in its natural environment through habitat destruction, over-exploitation and competition from other introduced species of Clarias . Despite its commercial importance and threats to natural populations, public databases do not include any genomic data for C. macrocephalus . We present the first genomic data for the bighead catfish from Illumina sequencing. A total of 128 Gb of sequence data in paired-end 150 bp reads were assembled de novo , generating a final assembly of 883 Mbp contained in 27,833 scaffolds (N 50 length: 80.8 kbp) with BUSCO completeness assessments of 96.3% and 87.6% based on metazoan and Actinopterygii ortholog datasets, respectively. Annotation of the genome predicted 21,124 gene sequences, which were assigned putative functions based on homology to existing protein sequences in public databases. Raw fastq reads and the final version of the genome as-sembly have been deposited in the NCBI (BioProject: PR-JNA604477, WGS: JAAGKR0 0 0 0 0 0 0 0 0, SRA: SRR11188453). The complete C. macrocephalus mitochondrial genome was also recovered from the same sequence read dataset and is available on NCBI (accession: MT109097), representing the first mitogenome for this species. Lastly, we find an expansion of the mb and ora1 genes thought to be associated with adaptations to air-breathing and a semi-terrestrial life style in this genus of catfish.
© 2020 Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license. ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ) Table   Subject Biology Specific subject area Genomics Type of data Sequencing raw reads, Assembly,

Value of the Data
First genomic dataset for the wild bighead catfish High BUSCO completeness and an assembly close to the estimated genome size indicate it will enable its use in selective breeding and population and conservation genetic studies of this native Vietnamese and commercial species. The data will facilitate genetic management for the genetic improvement and the conservation of bighead catfish populations including competition with introduced non-native species of Clarias . The data adds to the limited genomic available for the highly diverse catfish lineage [1][2][3][4] 1. Data description DNA sequencing using the Illumina platform was performed to generate the first de novo genome assembly for the commercial bighead catfish ( Clarias macrocephalus ). Approximately 853 million paired-end 150-bp reads were generated (128 GB of total bases) on a NovaSeq at the Deakin Genomics Centre. After poly-G, adapter and quality trimming, 98.75% of reads were re- tained and subsequently used for de novo assembly. The final 883 Mbp assembly consists of 27,833 scaffolds and contains 96.3% and 87.6% of complete metazoan and Actinopterygii BUSCOs, respectively ( Table 1 ). A kmer-based method estimates a heterozygosity rate of 0.5% and approximately 70.7 to 81% unique content in the genome. Subsequently, a total of 21,124 gene sequences were predicted from the genome, based on alignment evidence from protein sequences in Swis-sProt and C. macrocephalus transcript sequences assembled from short reads available on NCBI's SRA database ( Table 1 ). Significant expansions of each of the mb and ora1 genes were observed ( Fig. 1 ). The mitochondrial genome was extracted from the data set, which is 16,511 bp and is fully annotated with no gaps and contains 37 genes ( Fig. 2 ).

Catfish, tissue sampling and DNA extraction
A white muscle tissue sample was obtained from an adult catfish collected from a natural population which was used for aquaculture trials at Can Tho University, Viet Nam. The fish was anaesthetized in AQUI-S (20 ppm) before sampling. The tissue was preserved in ethanol and approximately 50 mg was used for genomic DNA (gDNA) extractions using a modified SDSchloroform method (Sokolov, 20 0 0). In brief, the tissue sample was cut into smaller pieces before an overnight digestion in 380μL lysis buffer (50 mM Tris-HCl, 10 mM EDTA, 20% SDS) and 20 μL Proteinase K ( > 600 mAU/mL). Protein precipitation was performed by adding 100 μL saturated > 5 M KCl and incubated on ice. 1x volume of chloroform was added and the aqueous layer containing the DNA was transferred into a new tube. DNA was precipitated using 1x volume of isopropanol and washed using 1 mL 80% ethanol before being eluted in 100 μL of elution buffer (10 mM Tris-HCl, 1 mM EDTA).

DNA library construction and sequencing
Approximately 2 μg of total DNA as measured by the Qubit 3.0 fluorometer (Invitrogen, USA) was used as the input for TruSeq PCR-Free kit (Illumina, San Diego, CA). Library was quantified using KAPA Library Quantification kit (KAPA Biosystems, CapeTown, South Africa) and Tapestation (Agilent, USA) for molarity and fragment length estimation, respectively. Sequencing of the PCR-free library was performed on an Illumina NovaSeq 60 0 0 sequencing platform using the run configuration of 2 × 150 bp.

Read pre-processing
A total of 128 Gbp of sequence reads was generated from sequencing. Reads were preprocessed with the removal of poly-Gs at the 3 ends ( -poly_g_ min _len 1 )

Whole genome assembly and annotation
The haploid genome size for Clarias macrocephalus was first estimated using a k-mer based approach. Pre-processed reads of mitochondrial origin were removed through the alignment of reads to the mitochondrial genome (assembly described in Section 2.6 , accession number: MT109097) with bowtie2 v2.3.3.1 (-I 100 -X 500 and using the -un-conc parameter for pairs of reads that did not align concordantly). Jellyfish was used to obtain a frequency distribution of kmers ( k = 19, 21 and 25). These k-mer distributions were processed with GenomeScope [7] (max kmer coverage disabled), which estimated an average haploid genome size of 881 Mbp, a 0.5% heterozygosity level and 70.7 to 81% of unique genome content (Supplementary Data 1).
MaSuRCA [8] was used to assemble the sequence reads in a de novo manner. The resulting assembly was subsequently run through Purge Haplotigs pipeline [9] to eliminate artefactual scaffolds caused by possible haplotigs. This produced a final assembly of 883 Mbp, a size that is consistent with the expected genome size for this species. The assembly is presented in 27,833 scaffolds with a scaffold N 50 length of 80.8 kbp and contains 96.3% and 87.6% of complete BUS-COs (v3.0.2) based on a set of 978 metazoan and 4584 Actinopterygii orthologs, respectively ( Table 1 ) (Supplementary Data 2), which exceeds other recently sequenced catfish genome assemblies [4] .
To obtain the repeat profile of the bighead catfish genome, RepeatModeler was used to build a de novo repeat library, the output of which was used by RepeatMasker to mask repetitive regions in the assembly. This process masked 38.28% of the assembly and identified the DNA/TcMar-Tc1 transposon as the most abundant repeat type (Supplementary Data 3). This is consistent with the observation of elevated levels of this transposon in the Ictalurus punctatus genome [10] and freshwater fish in general [ 10 , 11 ].
The MAKER pipeline [12] was used to predict gene boundaries and coding sequences within the assembly. For this purpose, two types of sequence information were provided to the pipeline to increase the accuracy and sensitivity of predictions. Firstly, publicly-available C. macrocephalus RNA reads were downloaded from NCBI's SRA database (accession: SRR3161876), pre-processed with Trimmomatic and assembled with Trinity v2.8.5 ( -SS_lib_type FR ) [13] . This produced a set of 124,748 transcripts. Secondly, all protein sequences from SwissProt (accessed on January 31st, 2020) was provided as protein homology evidence. Both sets of information were used by MAKER to produce initial gene models in its first iteration ( est2genome = 1, protein2genome = 1 ). These gene models were then used to train two ab initio gene predictors used in a second iteration of prediction. MAKER was run for a third iteration with gene models retrained with the output from the previous iteration. This process predicted 21,124 gene sequences with Annotation Edit Distances (AED) of ≤ 0.5 (Supplementary Data 4). A small AED indicates agreement of the annotation to aligned evidences while a AED = 1 shows no evidence support for the annotation [14] . At least 98% of these protein sequences possess a functional domain (98% from InterProScan analysis [15] ) or show homology (evalue 1e −10 , 98.4% from BLASTP analysis [16] ) to an existing sequence on NCBI's non-redundant database.

Enriched genes in Clarias macrocephalus
A distinctive feature of the genome of the related air-breathing walking catfish Clarias batrachus was an expansion of myoglobin ( mb ), olfactory receptor related to class A G protein-coupled receptor 1 ( ora1 ) and sulfotransferase 6b1 ( sult6b1 ) genes compared to non-air-breathing fishes [3] . We examined the generality of this finding by downloading the protein sequences of 17 other fish species, mostly from the Otophysi clade (https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id = 186626), from Ensembl. A list of species and Ensembl identifier is available in Table 2 which excludes protein sequences for C. batrachu s which were not made publicly available by the authors [3] and therefore could not be included in this analysis. Groups of orthologous proteins were identified by OrthoFinder v2.3.3 [17] , which uses DIAMOND blastp ( -more-sensitive, -e 0.001 ) for all pairwise alignments, resulting in the identification of orthogroups for mb, ora1 and sult6b1 genes. Sequences for each gene were aligned with MAFFT, trimmed with Gblocks (allowing for gaps) and finally used by IQ-TREE [18] to build Maximum-likelihood phylogenetic trees. Orthogroups and phylogenetic trees are available in Supplementary Data 5. This analysis recovered substantially higher numbers of mb (14 copies) and ora1 (10 copies) genes in the C. macrocephalus genome, clustered together in a phylogenetic clade ( Fig. 2 ), relative to the genomes of other fish species (mostly 1 or 2 copies). These results are similar to the finding for C. batrachus [3] , however an expansion was not observed for the sult6b1 gene, which was recovered in copy numbers similar to that in other fish genomes such as the channel catfish ( Ictalurus punctatus ) and the stickleback ( Gasterosteus aculeatus ) ( Table 2 ). The mb, ora1 and sult6b1 genes were found on 12, 3 and 4 scaffolds, respectively, with some evidence of tandem duplications similar to that reported for C. batrachus [3] .

Mitochondrial genome assembly and annotation
A subset of 20 million pre-processed reads were assembled with the IDBA-UD assembler [19] , which is used for the assembly of datasets with uneven read depths, to extract the C. macrocephalus mitochondrial genome using the genome skimming approach [20] , recovering a circular mitochondrial genome of 16,511 bp. This was annotated with MITOS, followed by manual adjustments of gene boundaries based on homology to protein sequences in NCBI's non-redundant database, generating a complete repertoire of 37 genes typically found in an animal mitochondrial genome. Fig. 2 shows the arrangement of 13 protein-coding genes, 22 transfer RNAs and 2 ribosomal RNAs visualized with OrganellarGenomeDRAW. The mitochondrial COI gene has 100% identity sequences for the species on NCBI and the whole mitogenome has the highest similarity with other Clarias species on NCBI (91.6-89.4%). The mitochondrial genome is available on NCBI under the accession number MT109097.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships which have, or could be perceived to have, influenced the work reported in this article.