The genome of a specimen of Thunnus thynnus ( Figure 1) was sequenced using Pacific Biosciences single-molecule HiFi long reads, generating 30.14 Gb (gigabases) from 5.19 million reads. GenomeScope analysis of the PacBio HiFi data estimated the haploid genome size at 782.63 Mb, with a heterozygosity of 0.75% and repeat content of 10.14%. These values provide an initial assessment of genome complexity and the challenges anticipated during assembly. Based on this estimated genome size, the sequencing data provided approximately 37.0x coverage of the genome. Chromosome conformation Hi-C sequencing produced 145.99 Gb from 966.80 million reads. Table 1 summarises the specimen and sequencing information.

The primary haplotype was assembled, and contigs corresponding to an alternate haplotype were also deposited in INSDC databases. The assembly was improved by manual curation, which corrected 49 misjoins or missing joins and removed 7 haplotypic duplications. These interventions decreased the scaffold count by 9.77%. The final assembly has a total length of 799.05 Mb in 119 scaffolds, with 468 gaps, and a scaffold N50 of 34.32 Mb ( Table 2).

The snail plot in Figure 2 provides a summary of the assembly statistics, indicating the distribution of scaffold lengths and other assembly metrics. Figure 3 shows the distribution of scaffolds by GC proportion and coverage. Figure 4 presents a cumulative assembly plot, with separate curves representing different scaffold subsets assigned to various phyla, illustrating the completeness of the assembly.

Most of the assembly sequence (99.16%) was assigned to 24 chromosomal-level scaffolds. These chromosome-level scaffolds, confirmed by Hi-C data, are named according to size ( Figure 5; Table 3).

The mitochondrial genome was also assembled. This sequence is included as a contig in the multifasta file of the genome submission and as a standalone record.

The estimated Quality Value (QV) and k-mer completeness metrics, along with BUSCO completeness scores, were calculated for each haplotype and the combined assembly. The QV reflects the base-level accuracy of the assembly, while k-mer completeness indicates the proportion of expected k-mers identified in the assembly. BUSCO scores provide a measure of completeness based on benchmarking universal single-copy orthologues.

The combined primary and alternate assemblies achieve an estimated QV of 60.3. The k-mer recovery for the primary haplotype is 84.08%, and for the alternate haplotype 83.23%; the combined primary and alternate assemblies have a k-mer recovery of 99.87%. BUSCO v5.5.0 analysis using the actinopterygii_odb10 reference set ( n = 3,640) identified 99.2% of the expected gene set (single = 98.5%, duplicated = 0.7%).

Table 2 provides assembly metric benchmarks adapted from Rhie et al. (2021) and the Earth BioGenome Project (EBP) Report on Assembly Standards September 2024. The assembly achieves the EBP reference standard of 6.C.Q60.

Thunnus thynnus young-of-the-year individual sampling was carried out in Italian waters during scientific activities (tagging and fishing surveys carried out in the Tyrrenian and Ionian Sea during late summer 2013 and summer 2022) following accepted methodologies and under permit. Tissues were sampled and stored in 5 mL ethanol-filled tubes placed in dry ice while onboard. Samples were subsequently stored at –20°C after landing. After 24 hours, samples were shipped in dry ice to the University of Bologna facility where they were stored at -80°C before shipment to the Wellcome Sanger Institute. All specimens were collected and identified by experienced scientists as indicated in the sample manifest. One of the specimens (specimen ID ERGA AV MT 03, ToLID fThuThy2) was used for PacBio DNA sequencing. The specimen used for Illumina Hi-C sequencing was specimen ID ERGA AV MT 06 (ToLID fThuThy2).

The workflow for high molecular weight (HMW) DNA extraction at the Wellcome Sanger Institute (WSI) Tree of Life Core Laboratory includes a sequence of procedures: sample preparation and homogenisation, DNA extraction, fragmentation and purification. Detailed protocols are available on protocols.io ( Denton et al., 2023b). The fThuThy2 sample was prepared for DNA extraction by weighing and dissecting it on dry ice ( Jay et al., 2023). Tissue from the muscle was homogenised using a PowerMasher II tissue disruptor ( Denton et al., 2023a).

HMW DNA was extracted at the WSI Scientific Operations core using the Automated MagAttract v2 protocol ( Oatley et al., 2023). The DNA was sheared into an average fragment size of 12–20 kb in a Megaruptor 3 system ( Bates et al., 2023). Sheared DNA was purified by solid-phase reversible immobilisation, using AMPure PB beads to eliminate shorter fragments and concentrate the DNA ( Strickland et al., 2023). The concentration of the sheared and purified DNA was assessed using a Nanodrop spectrophotometer and Qubit Fluorometer using the Qubit dsDNA High Sensitivity Assay kit. Fragment size distribution was evaluated by running the sample on the FemtoPulse system.

Tissue from the fThuThy3 sample was processed for Hi-C sequencing at the WSI Scientific Operations core, using the Arima-HiC v2 kit. In brief, 20–50 mg of frozen tissue (stored at –80 °C) was fixed, and the DNA crosslinked using a TC buffer with 22% formaldehyde concentration. After crosslinking, the tissue was homogenised using the Diagnocine Power Masher-II and BioMasher-II tubes and pestles. Following the Arima-HiC v2 kit manufacturer's instructions, crosslinked DNA was digested using a restriction enzyme master mix. The 5’-overhangs were filled in and labelled with biotinylated nucleotides and proximally ligated. An overnight incubation was carried out for enzymes to digest remaining proteins and for crosslinks to reverse. A clean up was performed with SPRIselect beads prior to library preparation. Additionally, the biotinylation percentage was estimated using the Qubit Fluorometer v4.0 (Thermo Fisher Scientific) and Qubit HS Assay Kit and Arima-HiC v2 QC beads.

Library preparation and sequencing were performed at the WSI Scientific Operations core.

PacBio HiFi

At a minimum, samples were required to have an average fragment size exceeding 8 kb and a total mass over 400 ng to proceed to the low input SMRTbell Prep Kit 3.0 protocol (Pacific Biosciences, California, USA), depending on genome size and sequencing depth required. Libraries were prepared using the SMRTbell Prep Kit 3.0 (Pacific Biosciences, California, USA) as per the manufacturer's instructions. The kit includes the reagents required for end repair/A-tailing, adapter ligation, post-ligation SMRTbell bead cleanup, and nuclease treatment. Following the manufacturer’s instructions, size selection and clean up was carried out using diluted AMPure PB beads (Pacific Biosciences, California, USA). DNA concentration was quantified using the Qubit Fluorometer v4.0 (Thermo Fisher Scientific) with Qubit 1X dsDNA HS assay kit and the final library fragment size analysis was carried out using the Agilent Femto Pulse Automated Pulsed Field CE Instrument (Agilent Technologies) and gDNA 55kb BAC analysis kit.

Samples were sequenced using both the Sequel IIe system and the Revio instrument (Pacific Biosciences, California, USA). The concentration of the library loaded onto the Sequel IIe was in the range 40–135 pM. Samples were sequenced on a Revio instrument (Pacific Biosciences, California, USA). Prepared libraries were normalised to 2 nM, and 15 μL was used for making complexes. Primers were annealed and polymerases were hybridised to create circularised complexes according to manufacturer’s instructions. The complexes were purified with the 1.2X clean up with SMRTbell beads. The purified complexes were then diluted to the Revio loading concentration (in the range 200–300 pM), and spiked with a Revio sequencing internal control. Samples were sequenced on Revio 25M SMRT cells (Pacific Biosciences, California, USA). The SMRT link software, a PacBio web-based end-to-end workflow manager, was used to set-up and monitor the sequencing runs, as well as perform primary and secondary analysis of the data upon completion.

Hi-C

For Hi-C library preparation, biotinylated DNA constructs were fragmented using the Covaris E220 sonicator and size selected using SPRISelect beads to 400 to 600 bp using the INTEGRA MINI 96 (INTEGRA). The DNA was then enriched using the Arima-HiC v2 kit Enrichment beads on the INTEGRA MINI 96 (INTEGRA). The NEBNext Ultra II DNA Library Prep Kit (New England Biolabs) was used for end repair, A-tailing, and adapter ligation. This uses a custom protocol which resembles the standard NEBNext Ultra II DNA Library Prep protocol but where library preparation occurs while DNA is bound to the Enrichment beads. Library PCR amplification was carried out using KAPA HiFi Hot start mix and a custom IDT UDI (Unique Dual Index) 96 barcode plate (Integrated DNA Technologies). Depending on sample concentration and biotinylation percentage determined at the crosslinking stage, samples were run for 10-16 PCR cycles. Post-PCR, samples were cleaned-up using SPRISelect beads and the INTEGRA MINI 96 (INTEGRA). The Hi-C sequencing was performed using paired-end sequencing with a read length of 150 bp on an Illumina NovaSeq 6000 instrument.

Assembly

Prior to assembly of the PacBio HiFi reads, a database of k-mer counts ( k = 31) was generated from the filtered reads using FastK. GenomeScope2 ( Ranallo-Benavidez et al., 2020) was used to analyse the k-mer frequency distributions, providing estimates of genome size, heterozygosity, and repeat content.

The HiFi reads were first assembled using Hifiasm ( Cheng et al., 2021) with the --primary option. Haplotypic duplications were identified and removed using purge_dups ( Guan et al., 2020). The Hi-C reads were mapped to the primary contigs using bwa-mem2 ( Vasimuddin et al., 2019). The contigs were further scaffolded using the provided Hi-C data ( Rao et al., 2014) in YaHS ( Zhou et al., 2023) using the --break option for handling potential misassemblies. The scaffolded assemblies were evaluated using Gfastats ( Formenti et al., 2022), BUSCO ( Manni et al., 2021) and MERQURY.FK ( Rhie et al., 2020).

The mitochondrial genome was assembled using MitoHiFi ( Uliano-Silva et al., 2023), which runs MitoFinder ( Allio et al., 2020) and uses these annotations to select the final mitochondrial contig and to ensure the general quality of the sequence.

Assembly curation

The assembly was decontaminated using the Assembly Screen for Cobionts and Contaminants (ASCC) pipeline. Flat files and maps used in curation were generated via the TreeVal pipeline ( Pointon et al., 2023). Manual curation was conducted primarily in PretextView ( Harry, 2022) and HiGlass ( Kerpedjiev et al., 2018), with additional insights provided by JBrowse2 ( Diesh et al., 2023). Scaffolds were visually inspected and corrected as described by Howe et al. (2021). Any identified contamination, missed joins, and mis-joins were amended, and duplicate sequences were tagged and removed. The curation process is documented at https://gitlab.com/wtsi-grit/rapid-curation.

Assembly quality assessment

The Merqury.FK tool ( Rhie et al., 2020), run in a Singularity container ( Kurtzer et al., 2017), was used to evaluate k-mer completeness and assembly quality for the primary and alternate haplotypes using the k-mer databases ( k = 31) that were computed prior to genome assembly. The analysis outputs included assembly QV scores and completeness statistics.

A Hi-C contact map was produced for the final version of the assembly. The Hi-C reads were aligned using bwa-mem2 ( Vasimuddin et al., 2019) and the alignment files were combined using SAMtools ( Danecek et al., 2021). The Hi-C alignments were converted into a contact map using BEDTools ( Quinlan & Hall, 2010) and the Cooler tool suite ( Abdennur & Mirny, 2020). The contact map was visualised in HiGlass ( Kerpedjiev et al., 2018).

The blobtoolkit pipeline is a Nextflow port of the previous Snakemake Blobtoolkit pipeline ( Challis et al., 2020). It aligns the PacBio reads in SAMtools and minimap2 ( Li, 2018) and generates coverage tracks for regions of fixed size. In parallel, it queries the GoaT database ( Challis et al., 2023) to identify all matching BUSCO lineages to run BUSCO ( Manni et al., 2021). For the three domain-level BUSCO lineages, the pipeline aligns the BUSCO genes to the UniProt Reference Proteomes database ( Bateman et al., 2023) with DIAMOND blastp ( Buchfink et al., 2021). The genome is also divided into chunks according to the density of the BUSCO genes from the closest taxonomic lineage, and each chunk is aligned to the UniProt Reference Proteomes database using DIAMOND blastx. Genome sequences without a hit are chunked using seqtk and aligned to the NT database with blastn ( Altschul et al., 1990). The blobtools suite combines all these outputs into a blobdir for visualisation.

The blobtoolkit pipeline was developed using nf-core tooling ( Ewels et al., 2020) and MultiQC ( Ewels et al., 2016), relying on the Conda package manager, the Bioconda initiative ( Grüning et al., 2018), the Biocontainers infrastructure ( da Veiga Leprevost et al., 2017), as well as the Docker ( Merkel, 2014) and Singularity ( Kurtzer et al., 2017) containerisation solutions.

Table 4 contains a list of relevant software tool versions and sources.

The Ensembl Genebuild annotation system ( Aken et al., 2016) was used to generate annotation for the Thunnus thynnus assembly (GCA_963924715.1) in Ensembl Rapid Release at the EBI. Annotation was created primarily through alignment of transcriptomic data to the genome, with gap filling via protein-to-genome alignments of a select set of proteins from UniProt ( UniProt Consortium, 2019).

The materials that have contributed to this genome note have been supplied by a Tree of Life collaborator. The Wellcome Sanger Institute employs a process whereby due diligence is carried out proportionate to the nature of the materials themselves, and the circumstances under which they have been/are to be collected and provided for use. The purpose of this is to address and mitigate any potential legal and/or ethical implications of receipt and use of the materials as part of the research project, and to ensure that in doing so we align with best practice wherever possible.

The overarching areas of consideration are:

•   Ethical review of provenance and sourcing of the material

•   Legality of collection, transfer and use (national and international)

Each transfer of samples is undertaken according to a Research Collaboration Agreement or Material Transfer Agreement entered into by the Tree of Life collaborator, Genome Research Limited (operating as the Wellcome Sanger Institute) and in some circumstances other Tree of Life collaborators.