Chromosome-level genome assembly of the hamour (orange-spotted grouper), Epinephelus coioides

Sampling and nucleic acids extractions

One individual male hamour (43 cm) was caught from the wild (East Qatar, Arabian Gulf, September 2022), and sampled directly. For genomic DNA extraction, 50 mg of liver tissue were excised, submerged in isopentane bath (prepared over dry ice) until frozen (30 s), stored in a pre-chilled cryotube and preserved at -80 °C until further analyses. In parallel, 50 mg of tissue from liver, kidney, heart, spleen, gonad, muscle, skin, gill, and tail were collected from the same specimen and placed in cryotubes containing RNAlater^® (Sigma-Aldrich), stored at +4 °C for 24 hrs to ensure appropriate tissue preservation, and then moved to -80 °C until RNA extraction.

High molecular weight (HMW) DNA was extracted from the liver sample using QIAGEN^® Genomic tip Blood & Cell Culture DNA Midi Kit (QIAGEN) following the manufacturer’s protocol. Purified DNA was quantified using an Invitrogen™ Qubit™ DNA BR Assay kit (Thermo Fisher Scientific), the purity of the sample was analysed using a NanoDrop™ 2000 spectrophotometer (Thermo Fisher Scientific) UV/Vis, and the integrity of the DNA was assessed with Femto Pulse Genomic DNA 165 kb kit (Agilent). DNA was stored at +4 °C until library preparation.

Total RNA was extracted using a Promega Maxwell^® RSC 48 Instrument and Maxwell RSC simply RNA Tissue kit (Promega), quantified using an Invitrogen Qubit RNA BR Assay kit (Thermo Fisher Scientific), and analysed for purity using an RNA 6000 Nano Bioanalyzer 2100 Assay (Agilent). Purified RNA samples were stored at -80 °C until library preparation.

Library preparation and sequencing

A long-read library for Oxford Nanopore Technologies™ (ONT) sequencing was prepared from DNA extracted from the liver using the 1D Sequencing kit SQK-LSK110 (ONT). In brief, 3 μg of DNA underwent end-repair and adenylation using the NEBNext^® Ultra™ II End Repair/dA-Tailing Module (New England Biolabs), followed by ligation of sequencing adaptors. The ligation product was purified using Beckman Coulter™ AMPure XP Beads (Beckman Coulter Life Science) and eluted in Elution Buffer (ONT). The library was sequenced on a PromethION™ 24 instrument with a R9.4.1 flow cell, and data collected for 110 hours. The quality parameters of the sequencing run were monitored in real time using the MinKNOW™ platform v22.10.7 (ONT, https://nanoporetech.com/document/experiment-companion-minknow) and basecalling performed using Guppy v6.3.9 (ONT, https://nanoporetech.com/document/Guppy-protocol ).

For the proximity ligation library, the Dovetail^® Omni-C^® Kit (Dovetail Genomics, Cantata Bio) was used on the HMW DNA extracted from the liver, following the manufacturer’s protocol. After reversal crosslinking, the DNA was purified and biotinylated chimeric molecules isolated using streptavidin beads before PCR enrichment with 12 PCR cycles using KAPA HiFi HotStart Ready Mix (Roche). The short-insert paired-end library for whole genome sequencing was prepared using the PCR-free protocol and the KAPA HyperPrep kit (Roche). After end-repair and adenylation, Illumina™ platform-compatible adaptors with unique dual indexes and unique molecular identifiers (Integrated DNA Technologies) were ligated. The sequencing library was quality controlled on a 2100 Bioanalyzer using the DNA 7500 assay (Agilent), quantified with KAPA Library Quantification Kit (Roche), and sequenced on four lanes of a NovaSeq™ 6000 (Illumina) with a read length of 2×151 bp.

Total RNA from individual tissues (tail, gonad, heart, gill, skin, spleen, liver, kidney, and muscle) was used to prepare RNA-Seq tissue-specific libraries with a KAPA Stranded mRNA-Seq kit (Roche) following the manufacturer’s protocol. The transcriptomes were sequenced on 4 lanes of a NovaSeq 6000 (Illumina) with a read length of 2×151 bp.

Nuclear genome assembly, Omni-C scaffolding, curation, and assembly quality checks

Data generated with the PromethION and NovaSeq 6000 were assembled with the Centro Nacional de Análisis Genómico (CNAG) Snakemake pipeline v2.0 (https://github.com/cnag-aat/assembly_pipeline) to obtain an optimal base assembly for further Omni-C scaffolding. The list of programs, parameters and versions used to assemble and quality check the genome are presented in Table 1. In brief, Illumina reads were processed with Cutadapt (Martin, 2011), while ONT reads were filtered with FiltLong (https://github.com/rrwick/Filtlong). Filtered ONT reads were assembled with both Flye (Kolmogorov et al., 2019) and NextDenovo (Hu et al., 2024). GenomeScope2 (Ranallo-Benavidez et al., 2020; Vurture et al., 2017) was used to estimate genome size with the 20-mers present in the pre-processed Illumina reads. The NextDenovo (Hu et al., 2024) assembly was polished with both ONT and Illumina paired-end reads using Hypo (Kundu et al., 2019) and then the polished assembly was collapsed with purge_dups (Guan et al., 2020) to remove haplotypic duplications.

For further proximity ligation-based scaffolding, a total of 206.96 million Omni-C read pairs were mapped to the assembled genome using the Dovetail Genomics recommended protocol (https://omni-c.readthedocs.io/en/latest/fastq_to_bam.html). After excluding PCR duplicates, 106.91 million valid Omni-C read pairs were used to scaffold the assembly with YaHS (Zhou et al., 2023) using the default initial contig error correction step.

To guide manual curation of the assembly, the ONT read coverage was computed for all positions in the assembly using minimap2 (Li, 2018), SAMtools (Danecek et al., 2021; Li et al., 2009), and BEDtools (Quinlan and Hall, 2010), as well as the location of gaps with fasta-stats.py (https://github.com/cnag-aat/scripts/blob/main/fasta-stats.py) and telomeres with telomeric-identifier (Brown et al., 2023). These extensions were added to the contact map using PretextGraph (https://github.com/sanger-tol/PretextGraph). Manual curation was performed according to the rapid curation protocol from The Sanger Institute (https://gitlab.com/wtsi-grit/rapid-curation ) using PretextView (https://github.com/sanger-tol/PretextView). The genome was assessed for completeness with BUSCO using the odb10 Actinopterygii database (Manni et al., 2021), with Merqury (Rhie et al., 2020) for consensus accuracy (QV) and k-mer statistics, for contiguity statistics with fasta-stats.py (https://github.com/cnag-aat/scripts/blob/main/fasta-stats.py), and for contamination with BlobToolKit (with NCBI nt database, August 2023 update) (Challis et al., 2020) and BUSCO using the odb10 databases for Fungi and Bacteria (Manni et al., 2021). For comparison with the genome assembled in this study, the genome previously described by Li and colleagues (Li et al., 2023) and available in the European Nucleotide Archive (accession ID: PRJEB28248) was also assessed for completeness with BUSCO’s odb10 Actinopterygii database (Manni et al., 2021).

Genome annotation

The hamour genome assembly annotation was obtained by combining transcript alignments, protein alignments and ab initio gene predictions. The list of programs, parameters, and versions used for genome annotation is provided in Table 1. In brief, repeats present in the genome assembly were annotated with RepeatMasker (Smith et al., 2007; http://www.repeatmasker.org) using the custom repeat library available for Danio rerio and a new repeat library specific for this study made with RepeatModeler (https://github.com/Dfam-consortium/RepeatModeler). After excluding repeats that were part of repetitive protein families from the resulting library, RepeatMasker (Smith et al., 2007; http://www.repeatmasker.org) was run again with this new library performing a BLAST (Altschul et al., 1990) search against UniProt (May 2023, https://www.uniprot.org/) to annotate the specific repeats. RNA-seq reads were aligned to the previously assembled genome using STAR (Dobin et al., 2013). Transcript models were subsequently generated using StringTie (Pertea et al., 2015) and merged using TACO (Niknafs et al., 2017). High-quality junctions to be used during the annotation process were obtained by running ESPRESSO (Gao et al., 2024) after mapping with STAR (Dobin et al., 2013). Finally, assembled spliced alignments were produced with PASA (Haas et al., 2008). TransDecoder (https://github.com/TransDecoder/TransDecoder) was run on the spliced alignments in PASA (Haas et al., 2008) to detect coding regions in the transcripts. The complete proteomes of Gymnodraco acuticeps, Sander lucioperca, Cottoperca gobio, and Perca fluviatilis were downloaded from UniProt (May 2023, https://www.uniprot.org/) and aligned to the genome using miniprot (Li, 2023). Ab initio gene predictions were performed on the repeat-masked assembly with GeneID (Alioto et al., 2018) and Augustus (with human parameters) (Stanke et al., 2006), and Genemark-ET in self-trained mode (Lomsadze et al., 2014) with and without incorporating evidence from the RNA-seq data. Finally, all the data were combined into consensus coding sequence models using EVidenceModeler (Haas et al., 2008). Additionally, untranslated regions (UTRs) and alternative splicing forms were annotated via two rounds of PASA (Haas et al., 2008) annotation updates. Functional annotation was performed on the annotated proteins with PANNZER’s online server (Törönen and Holm, 2022; http://ekhidna2.biocenter.helsinki.fi/sanspanz/).