Background & Summary

The Mediterranean mussel, Mytilus galloprovincialis Lamarck 1819, is a gregarious species that attaches to rocks or other hard surfaces using byssal threads. The mussel plays a crucial ecological role as an ecosystem engineer by creating habitat and promoting environmental heterogeneity, thereby enhancing local biodiversity1. The mussel is also known to accumulate contaminants in its tissues from the surrounding environment2,3,4,5. As a result, it has been widely used as a reliable bioindicator in various monitoring programs, such as the Mussel Watch Programme6. In addition to its ecological role as an ecosystem engineer and bioindicator, M. galloprovincialis also holds considerable commercial value. M. galloprovincialis is a widely cultivated bivalve species globally. In 2015, the worldwide production of M. galloprovincialis for human consumption exceeded 1.1 million tonnes7.

The mussel M. galloprovincialis is a highly invasive species that originated in the Mediterranean Sea and the eastern Atlantic, extending north to the British Isles. The species has been introduced to different areas via ballast water over the past century, and is now found in temperate coastal regions of both the northern and southern hemispheres (see Fig. 1a for detail). The Global Invasive Species Programme has identified M. galloprovincialis as one of the top 100 worst invasive species globally due to its significant impact on biological diversity8. The invasive success of M. galloprovincialis in central and southern California is believed to be partly attributed to its physiological adaptations that allow it to outperform M. trossulus in high temperatures9. Fields and colleagues9 discovered that a slight alteration in the structure of cytosolic malate dehydrogenases in M. galloprovincialis allows them to function effectively at higher temperatures. Recent advancements in genomic screening have enabled the identification of a greater number of genetic variations associated with differences in environmental conditions, such as temperature.

Fig. 1
figure 1

The global distribution of the Mediterranean mussel, Mytilus galloprovincialis, is shown in the map (a). The green regions indicate the native range, while the red regions indicate the invasive range. The distribution data was compiled from various sources51,52,53,54,55,56,57,58. In the photo (b), gregarious M. galloprovincialis can be seen on a rocky shore in Dalian, China. Photo credit: Guo-dong Han.

Full size image

Chromosome-scale reference genomes are essential for applying genomics to biology, aquaculture, and biodiversity conservation. They offer improved contiguity and completeness compared to fragmented genome assemblies, enabling the testing of crucial ecological and evolutionary hypotheses. Although there is significant scientific and commercial interest in mussels for both biology and aquaculture purposes, the availability of a high-quality, chromosome-scale reference genome for M. galloprovincialis is currently limited10,11,12. The genome of M. galloprovincialis, similar to other bivalves, is relatively large and complex, and frequently exhibits high heterozygosity13. Particularly, the genome of M. galloprovincialis exhibits high levels of hemizygosity (only one of the two chromosomal pairs encodes a region or block of DNA) compared to other molluscs11,14. Previous attempts at assembly of this species were significantly hindered by these factors.

In this study, we used PacBio HiFi technology to sequence and assemble the genome of M. galloprovincialis. We also utilized high-throughput chromosome conformation capture (Hi-C) technology to achieve chromosome-scale scaffolding. As a result, we constructed a high-quality chromosome-scale genome assembly for M. galloprovincialis. The primary assembly is highly continuous, complete, and accurate. Key metrics include a scaffold N50 of 96.9 Mb (Fig. 2b), k-mer completeness of 68.8, and a k-mer-based quality value (QV) of 51.1 (Table 1). Gene annotation analysis using the metazoa_odb10 lineage dataset showed a completeness of 97.5%, indicating high annotation quality. Certain taxonomically restricted genes were not identified during the BUSCO assessment of genome completeness. For instance, myticin was not part of the gene set in metazoa_odb10, but it was found in the genome. Compared to the other available M. galloprovincialis assemblies in GenBank, our assembly significantly improves contiguity and functional completeness (as measured by the number of complete BUSCO notations). Previous karyometric analysis has shown that M. galloprovincialis has 14 chromosomes15,16. The application of Hi-C in this study resulted in 14 long scaffolds in the primary assembly, approaching chromosome-level assembly. The Hi-C contact map suggests that the primary assembly is highly contiguous (Fig. 3).

Fig. 2
figure 2

Characteristics of the M. galloprovincialis genome. (a) The k-mer spectra output is generated by GenomeScope 2.0 using short reads obtained from whole-genome sequencing (WGS). (b) N(x) plots illustrate the contiguity profile of the primary assembly. (c) The cumulative length of the primary assembly is plotted for all scaffolds.

Full size image
Table 1 Statistics of sequencing and assembly.
Full size table
Fig. 3
figure 3

Hi-C contact maps for the primary genome assembly. Hi-C contact map represents the spatial proximity of genomic regions in a linear format. Each cell in the contact map contains sequencing data that confirms the connection between two specific regions. The scaffolds are arranged in order of length and are visually separated by white lines.

Full size image

The assembly exhibited widespread hemizygous regions, with 8.79% deletions of the total genome (Fig. 4). A total of 18,429 genes were identified that are located within hemizygous regions. The hemizygosity we observed in this study is higher than that of other molluscs (ranging from 0.17% to 6.69%14). However, it is significantly lower than the hemizygosity estimated in a previous study for the same species (36.78% with a genome size of 1.28Gb11). The observed differences may be due to the fact that the sample used in our study was collected from the invasive region, whereas the sample used in the previous study was collected from its native region11. Furthermore, the genome size of M. galloprovincialis could be notably impacted by hemizygous regions. Increased hemizygosity may lead to a reduction in genome size of this species. Therefore, it seems that structural variation, especially large insertion/deletion polymorphisms, play a crucial role in promoting successful invasiveness. Future studies should focus on confirming the significance of structural variation in invasive adaptation by resequencing individuals from a wider range of native and invasive locations.

Fig. 4
figure 4

Chromosomal map of hemizygous loci. Deletions at least 10 kb in length that were not associated with tandem repeats were shown in black.

Full size image

Our chromosome-level reference genome for M. galloprovincialis provides a more complete understanding of its resilience and invasiveness compared to incomplete and fragmented reference genomes. Previous studies using a fragmented reference genome (GCA_001676915.1) identified a significant number of single nucleotide polymorphisms (SNPs) in M. galloprovincialis, indicating the presence of standing genetic variation that enables rapid adaptation to ocean acidification17. Another study found a large number of SNPs and demonstrated that divergent climatic factors have driven adaptive genetic variation in M. galloprovincialis over the past century, contributing to its successful invasion of various thermal habitats18. Our complete reference genome will further enhance the identification and annotation of standing genetic variation and adaptive genetic variation in this species. Additionally, the whole genome data will be valuable for the aquaculture industry in developing genetic markers for economically important traits. Genotypes in quantitative trait loci (QTLs) are associated with specific phenotypes, such as shell size and body weight, and can be used as markers for selection in breeding programs13. In conclusion, our chromosome-level reference genome, combined with future population genomics studies, will greatly contribute to a deeper understanding of the resilience and invasiveness of M. galloprovincialis and facilitate its application in aquaculture.

Methods

Biological materials

One female mussel was collected from Yangmadao, Shandong Province, China (37.46342°N, 121.59541°E) on June 1, 2022. The mussel had a shell length of 76 mm. The specimen was subsequently taken to Yantai University, where the adductor muscle was dissected and promptly frozen using liquid nitrogen. The adductor muscle tissue was used to prepare Illumina paired-end library, HiFi SMRTbell library and Hi-C library.

Nucleic acid extraction, library preparation and sequencing

The genomic DNA was extracted using the QIAGEN Genomic-tip 100/G kit (QIAGEN, Germany) following the manufacturer’s instructions. DNA quality was assessed using pulsed field gel electrophoresis, and DNA concentration and purity were measured using the Qubit DNA Assay Kit (Invitrogen, USA) and Nanodrop 2000 (Thermo, USA), respectively.

A library with an insert size of 500 bp was prepared using the TruSeq Nano DNA Library kit (Illumina, CA, USA) by randomly fragmenting and ligating adaptors to the DNA sequences. Paired-end sequencing with 150 bp was performed using the NovaSeq 6000 sequencing system (Illumina, CA, USA).

For the HiFi SMRTbell library, 8 μg of DNA was used to prepare library using the SMRTbell Express Template Preparation Kit 2.0 (Pacific Biosciences, USA) following the manufacturer’s recommendations. The library was sequenced using the HiFi SMRTbell sequencing method with an average fragment size of 15 to 20 kb. The DNA library was loaded onto the flow cell with sequencing buffer and loading beads, and sequencing was performed using the DNA Sequencing Reagent Kit (Pacific Biosciences, USA) according to the manual.

To construct Hi-C libraries, the adductor muscle was grinded with liquid nitrogen and cross-linked with 4% formaldehyde solution and then quenched with glycine. The cross-linked sample was lysed and the nuclei were isolated. The nuclei were then solubilized and digested with the restriction enzyme MboI. The DNA ends were marked with biotin-14-dCTP and the cross-linked fragments were ligated. The nuclear complexes were reverse cross-linked and the DNA was purified. Non-ligated fragment ends were treated with T4 DNA polymerase to remove biotin. The sheared fragments were repaired and the Hi-C samples were enriched using streptavidin C1 magnetic beads. The Hi-C libraries were prepared by adding A-tails to the fragment ends and ligating Illumina paired-end (PE) sequencing adapters. The libraries were amplified by PCR and sequenced on an NovaSeq 6000 sequencing system.

Genome assembly

We used Meryl v1.319 to generate k-mer counts (k = 21) from Illumina WGS reads. The resulting k-mer database was then utilized in GenomeScope2.020 to estimate various genome features, such as sequencing error, genome size, heterozygosity, and repeat content. We estimated the genome size to be 1.34 Gb and the nucleotide heterozygosity rate to be 3.27% (Fig. 2a).

For the assembly of the M. galloprovincialis genome, we employed the de novo assembler HiFiasm v0.19.5-r59221. The final output is a diploid assembly comprising two pseudohaplotypes: a primary assembly and an alternate assembly. We used purge_dups v1.2.522 to identify duplicated sequences and overlapping contigs in the assemblies. The assemblies were then scaffolded using Hi-C data with SALSA23. Gaps remaining after scaffolding were closed using PacBio HiFi reads and YAGCloser (https://github.com/merlyescalona/yagcloser). Contamination was checked using BlobToolKit v2.6.524.

The primary assembly was manually curated using Hi-C contact maps. The Hi-C data was aligned to the reference using bwa mem v0.7.17-r118825. Ligation junctions were identified and Hi-C pairs were generated using pairtools v1.0.226. A 50 kb Hi-C matrix was created using cooler v0.9.227 and balanced using hicExplorer v3.628. And PretextMap v0.1.9 (https://github.com/sanger-tol/PretextMap) were used for visualizing the contact maps (Fig. 3).

Structural variation detection

We employed the allelic structural variation detection pipeline29 to determine the hemizygous regions of the primary assembly. Tandem repeats in genome assembly were identify using trf v4.0930. PacBio HiFi reads were aligned to the primary assembly using minimap2 v2.17-r94131. Structural variations in the primary assembly were identified using pbsv v2.9.0 (https://github.com/PacificBiosciences/pbsv). Deletions at least 10 kb in length that were not associated with tandem repeats were visualized using chromoMap v4.1.132. The primary assembly exhibited widespread hemizygous regions, with 8.79% deletions of the total genome length (Fig. 4). When insertions were taken into account, the rate of hemizygosity of the primary assembly increased to 16.32%.

Genome annotation

Prior to gene structure annotation, repetitive sequences in the primary assembly were identified and masked. A de novo canonical database of repetitive elements was constructed using RepeatModeler v2.0.333 with the “-LTRStruct” option. Repeat families specific to Bivalvia were extracted from Repbase (RepeatMaskerEdition-20181026) and Dfam (v3.8) database to create a homology canonical database. These two databases were combined and used with RepeatMasker v4.1.234 to identify and classify repeats in the primary assembly. Approximately 60.42% (854 Mb) of the assembled sequences of M. galloprovincialis were identified as repetitive sequences (Table 2).

Table 2 Statistics of repetitive sequence of M. galloprovincialis genome.
Full size table

Protein-coding genes were identified using a combination of ab initio prediction and transcriptome-assisted methods. The BRAKER2 v2.1.635 gene prediction pipeline was used to predict genes from repeat-masked genome sequences. Transcriptome-assisted annotation was performed using downloaded Illumina RNA-Seq data from seven tissues (PRJNA230138). The RNA-Seq data were aligned to the primary assembly using HISAT2 v2.2.136. The resulting BAM files were merged and used for genome-guided assembly of transcripts using Trinity v2.1.137. The genome-guided RNA-Seq assemblies were then inputted to PASA v2.5.238 for transcriptome database generation. Finally, ab initio gene predictions and transcript alignments were combined into weighted consensus gene structures using EVidenceModeler v1.1.139. Finally, we identifying 58,480 genes in the primary assembly (Table 3). We identified protein-coding regions located within hemizygous regions using bedtools intersect v2.30.040. Genes located within hemizygous regions were determined by identifying at least one exon falling within these regions. A total of 38,301 exons were found to be located within hemizygous regions, and these exons are associated with 18,429 genes.

Table 3 Statistics of functionally annotated genes of M. galloprovincialis genome.
Full size table

To perform functional annotation, we compared the predicted protein sequences to the Nr and UniProt database using blastp v2.13.041 with an e-value threshold of 1e-5. Additionally, we used InterProScan v5.5742 and kofam_scan v1.3.0 (https://github.com/takaram/kofam_scan) for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotation, respectively. Among the predicted proteins, 54,374, 28,043, 23,011, and 9,547 proteins were matched to the Nr, Uniprot, Go, and KEGG databases, respectively (Table 3).

Data Records

The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive43 in National Genomics Data Center (NGDC)44, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA01559745) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa. The sequence accessions for the primary and alternate assemblies have been deposited in the NCBI, under accession number JAWDJN00000000046 and JAZKRD00000000047, respectively. The genome annotation file, functional annotations for predicted genes, bed files of hemizygous loci and genes located within hemizygous regions are available in the figshare repository, respectively48,49.

Technical Validation

We used the QV pipeline of Merqury19 to estimate the assembly QV based on k-mer analysis. The script “best_k.sh” in Merqury was employed to determine the optimal k-mer length. Meryl was then utilized to calculate the number of k-mers in the Illumina WGS reads using default settings. The QV evaluation was performed in Merqury using the output from Meryl and the assembly. The findings demonstrated a k-mer completeness of 68.8 and a k-mer-based QV of 51.1, suggesting high levels of completeness and accuracy. The spectra-asm plot illustrates a well-assembled diploid genome (Fig. 5a), with 1-copy k-mers unique to the primary assembly (red) and alternate assembly (blue), and 2-copy k-mers shared by both assemblies (green).

Fig. 5
figure 5

Quality assessment of assembled genome of M. galloprovincialis. (a) Merqury assembly spectrum plots for evaluating k-mer completeness. (b) Comparison of BUSCO scores (metazoa_odb10) for the assemblies of M. galloprovincialis with four other assemblies accessible in the NCBI database.

Full size image

To evaluate the accuracy of the assemblies and the predicted gene set, we employed BUSCO v5.2.050 with metazoa_odb10 to determine standard metrics of assembly completeness. Gene annotation analysis revealed a completeness rate of 97.5%, signifying high-quality annotation. The ultimate predicted gene set comprised 58,480 genes with a BUSCO value of 97.2%. Our assembly demonstrated superior contiguity and functional completeness compared to other M. galloprovincialis assemblies available in GenBank (Fig. 5b).