Background & Summary

Stag beetles (family Lucanidae) belong to the superfamily Scarabaeoidea within the order Coleoptera, comprising approximately 1,500 species distributed globally1. Male stag beetles are renowned for their enlarged mandibles, which they use in combative displays to secure preferred mating sites and food competition2. Owing to their striking morphology and complex behavior, many lucanid species have become model organisms for studies on behavioral ecology and functional morphology3. Their impressive mandibles also contribute to their popularity as exotic pets and valuable items in private collections4. Stag beetle larvae develop in and feed on decaying wood, playing a crucial role in forest ecosystems by promoting wood decomposition, nutrient recycling, and vegetation regeneration5,6. Adults of many species are nocturnal and primarily feed on tree sap and fermenting fruits4,7. Due to their ecological role and sensitivity to habitat changes, lucanid beetles are considered reliable bioindicators of forest matter cycling and ecosystem health8.

These beetles are distributed globally, occurring on all continents except Antarctica and inhabiting a diverse array of ecosystems, including forests, grasslands, and deserts9. The Lucanidae family is considered one of the most basal lineages within the superfamily Scarabaeoidea, underscoring its significant evolutionary importance10,11. Current research on stag beetles has primarily focused on taxonomy and phylogenetic relationships, drawing on nuclear gene fragments and mitochondrial multi-gene sequences12. High-quality genomic data are essential for gaining deeper insights into the evolutionary placement of Lucanidae within Scarabaeoidea. As of April 2025, only six Lucanidae genomes have been deposited in the NCBI database. In contrast to the rapidly growing number of genome assemblies for other beetle families, the availability of high-quality genomes for Lucanidae remains limited, highlighting the urgent need for additional genome sequencing and assembly efforts in this group.

To deepen our understanding of Lucanidae evolution and ecological adaptations, we assembled a chromosome-level genome of Odontolabis cuvera (Boisduval, 1835) by integrating PacBio HiFi long reads, Illumina short reads, and Hi-C data. Comprehensive genome annotation was performed, including identifying repetitive elements, non-coding RNAs, and protein-coding genes. This high-quality reference genome marks a significant advancement in Lucanidae research and provides a valuable genomic resource for exploring this beetle family’s evolutionary history and ecological adaptations.

Methods

Sample collection and sequencing

A single female specimen of O. cuvera was collected in Yunnan Province, China, on 24 October 2024 for concurrent DNA and RNA sequencing. Muscle tissue was carefully extracted from the pronotum and posterior abdominal segments. The tissue was washed in phosphate-buffered saline for five minutes to eliminate external contaminants. It was then flash-frozen in liquid nitrogen for 20 minutes and subsequently stored at −80 °C until sequencing procedures were initiated.

Genomic DNA was extracted using the DNeasy Blood & Tissue Kit (Qiagen), and total RNA was isolated with TRIzol Reagent (Thermo Fisher Scientific), following the manufacturers’ standard protocols. Illumina TruSeq DNA PCR-Free Kit was used to construct PCR-free libraries, yielding 150 bp paired-end reads. Hi-C libraries were generated by formaldehyde cross-linking, followed by MboI digestion, end-repair, and purification steps, following a standard protocol13. Short-read data were generated using the Illumina NovaSeq. 6000 platform. A 20 kb SMRTbell library was constructed (PacBio SMRTbell Express Template Prep Kit 2.0) and sequenced in HiFi mode on a PacBio Sequel II system. Berry Genomics (Beijing, China) conducted all library preparations and sequencing. In total, our sequencing efforts generated 160.95 Gb of data, including 36.70 Gb of PacBio HiFi long reads (61.02× coverage), 56.09 Gb of Illumina short reads (93.26×), and 58.56 Gb of Hi-C data (97.36×) (Table 1). PacBio HiFi sequencing generated reads with a scaffold N50 of 15.88 kb and an average read length of 15.93 kb.

Table 1 Statistics of the sequencing data used for genome assembly.
Full size table

Genome assembly

Raw Illumina reads were processed for quality control using BBTools v38.8214. Duplicate reads were first removed with “clumpify.sh”. Subsequently, bbduk.sh was applied to trim low-quality bases and adapter sequences according to strict quality criteria. This process involved discarding reads with Q < 20, removing reads with >5 Ns, trimming poly-A/G/C tails longer than 10 bp, and correcting overlapping paired reads. We conducted a k-mer-based genome survey analysis using GenomeScope v2.015 to estimate the genome size, heterozygosity, and repetitive sequence content of the O. cuvera genome. The estimated genome size ranged from 900.52 to 906.45 Mb, with repetitive elements comprising approximately 37.18–37.19% of the total genome. The analysis also revealed a heterozygosity rate of 1.13–1.39%, indicating a moderately high level of genetic diversity (Fig. 1).

Fig. 1
figure 1

Genome size estimation of Odontolabis cuvera using GenomeScope.

Full size image

The primary genome assembly of O. cuvera was performed using PacBio HiFi long reads with Hifiasm v0.19.816, applying default parameters. To eliminate redundant heterozygous sequences, Purge_Dups v1.2.517 was employed with a haploid cutoff value of 70 to identify and remove haplotigs effectively. Following quality control, Hi-C reads were aligned to the draft assembly using Juicer v1.6.218. Chromosome-level scaffolding was carried out with 3D-DNA v18092219, anchoring the primary contigs into chromosome-scale assemblies. The resulting genome assembly was meticulously reviewed, and any potential misassemblies were manually corrected using Juicebox v1.11.0818. To detect potential contaminants, we employed MMseqs. 2 v11.120 to conduct BLASTN-like searches against both the NCBI nucleotide and UniVec databases. Additional screening for vector contamination was performed using blastn (BLAST + v2.11.0)21 against the UniVec database. Sequences with over 90% identity to entries in either database were flagged as potential contaminants, while those with 80–90% identity underwent further verification through online BLASTN searches against the NCBI nucleotide database. Suspected bacterial and fungal contaminants were subsequently removed from the assembled sequences. The final O. cuvera genome assembly achieved chromosome-level resolution, with a total size of 908.07 Mb, comprising 66 scaffolds and 147 contigs, and a GC content of 32.65% (Table 2). A total of 81 gaps were present in the assembly. The scaffold and contig N50 values were 65.36 Mb and 16.39 Mb, respectively. In total, 99.58% of the assembled sequence (904.22 Mb) was successfully anchored to 14 chromosomes, which were ordered by descending length and ranged from 49.09 Mb to 94.68 Mb (Table 3; Figs. 2, 3).

Table 2 Genome assembly statistics for Odontolabis cuvera.
Full size table
Table 3 Statistics for chromosomes sequence length.
Full size table
Fig. 2
figure 2

Genome-wide chromosomal heatmap of Odontolabis cuvera, with individual chromosomes outlined in blue and contigs outlined in green.

Full size image
Fig. 3
figure 3

Genome characteristics of Odontolabis cuvera. The circular genome plot displays, from the outermost to the innermost ring: (1) chromosome length, (2) GC content, (3) gene density, and (4) the distribution of major transposable elements, including DNA transposons, SINEs, LINEs, LTR retrotransposons, and simple repeats.

Full size image

Genome annotation

To characterize repetitive elements in the O. cuvera genome, we performed de novo repeat annotation using RepeatModeler v2.0.422, incorporating the “-LTRStruct” pipeline to enhance the identification of LTR retrotransposons. The resulting repeat library was merged with RepBase-2023090923 and Dfam v3.524 to construct a comprehensive custom repeat database. RepeatMasker v4.1.225 was then employed to identify and mask repetitive sequences by aligning the genome against this integrated library. The RepeatMasker analysis revealed that approximately 481.28 Mb, accounting for 53.00% of the genome, consists of repetitive sequences. These include 233.31 Mb (25.69%) of unclassified repeats, 119.93 Mb (13.19%) of DNA transposons, 86.65 Mb (9.55%) of LINEs, 32.94 Mb (3.63%) of LTRs, and 5.86 Mb (0.65%) of simple repeats, along with additional repeat categories (Table 4).

Table 4 Genome assembly and annotation statistics of Odontolabis cuvera.
Full size table

Non-coding RNAs (ncRNAs) in the O. cuvera genome were annotated using Infernal v1.1.226 against the Rfam v14.1027 database, while tRNAscan-SE v2.0.928 was employed to predict transfer RNAs (tRNAs). In total, 1,219 ncRNAs were identified, including 4 long non-coding RNAs (lncRNAs), 64 ribozymes, 93 small nuclear RNAs (snRNAs), 99 microRNAs (miRNAs), 507 tRNAs, and 222 ribosomal RNAs (rRNAs) (Table 4).

The annotation of protein-coding genes in O. cuvera was conducted using MAKER v3.01.0329, an annotation pipeline that integrates multiple sources of evidence to produce high-confidence gene models. Three primary lines of evidence were incorporated: (1) transcriptomic evidence derived from RNA-seq reads aligned with HISAT2 v2.2.130 and assembled using StringTie v2.1.631; (2) ab initio predictions from BRAKER v2.1.632, incorporating both GeneMark-ES/ET/EP v4.68_lic33 and AUGUSTUS v3.4.034 pipelines trained on RNA-seq alignments and OrthoDB v1135 reference proteins; and (3) homology-based predictions generated by GeMoMa v1.936, leveraging protein sequences from five reference species: Drosophila melanogaster37 (GCF_000001215.4), Apis mellifera38 (GCA_003254395.2), Coccinella septempunctata39 (GCA_907165205.1), Prosopocoilus inquinatus40 (GCA_036172665.1), and Tribolium castaneum41 (GCA_031307605.1) (Table 5). The outputs from BRAKER and GeMoMa were merged and provided as ab initio input to the MAKER pipeline. A total of 21,798 predicted protein sequences were identified, reflecting that many genes produce multiple transcript variants. When considering only the longest transcript for each gene, the O. cuvera genome contained 18,332 predicted protein-coding genes, with an average gene length of 10,552.3 bp. Genes exhibited a mean structure of 5.4 exons, 4.4 introns, and 5.2 coding sequences (CDSs). Average exon length was 314.6 bp, while introns and CDSs measured 2,101.4 bp and 262.9 bp, respectively (Table 4). Gene set completeness was evaluated using BUSCO with the insecta_odb10 dataset (n = 1,367). The annotated protein-coding gene set exhibited 98.8% completeness, including 1,350 (97.7%) single-copy orthologs, 15 (1.1%) duplicated genes, 4 (0.3%) fragmented genes, and 13 (0.9%) missing genes. These results demonstrate that the gene annotations for O. cuvera are both comprehensive and of high quality.

Table 5 Species taxonomic information and accession code of all samples used in this study.
Full size table

Gene functional annotation was conducted using DIAMOND v2.0.11.142 in sensitive mode (–more-sensitive -e 1e-5) to align predicted protein sequences against the UniProtKB database. To further assign Gene Ontology (GO) terms, identify metabolic pathways (KEGG and Reactome), and annotate protein domains, we employed eggNOG-mapper v2.0.143 and InterProScan v5.53-87.044. The InterProScan analysis incorporated five databases: Pfam45, SMART46, SUPERFAMILY47, Gene3D48, and CDD49. Outputs from all tools were integrated to generate comprehensive functional annotations. In total, 16,972 genes were annotated with UniProt entries, 11,405 were assigned GO terms, 5,467 were mapped to KEGG pathways, 3,096 were associated with Enzyme Commission numbers, and 15,008 were classified into Clusters of Orthologous Groups (COG). Additionally, genome-wide distributions of repeat elements, gene density, and GC content across individual pseudochromosomes were visualized using TBtools50.

Data Records

The raw sequencing data and genome assembly of Odontolabis cuvera are publicly available through the National Center for Biotechnology Information (NCBI). The sequencing datasets, including Hi-C (SRR3279340551), transcriptome (SRR3183488052), Illumina short reads (SRR3183488153), and PacBio HiFi long reads (SRR3183488254), are publicly available under their respective accession numbers. The final genome assembly is available under NCBI accession GCA_049462965.155. Genome annotation files, including repeat element profiles, gene structure predictions, and functional annotations, are available via Figshare56.

Technical Validation

To evaluate the quality of the Odontolabis cuvera genome assembly, two complementary approaches were employed. First, genome assembly completeness was assessed using BUSCO v5.0.457 with the Insecta gene set (n = 1,367), revealing a high completeness score of 99.1%, with 98.3% single-copy, 0.8% duplicated, 0.3% fragmented, and 0.6% missing BUSCOs. Second, assembly accuracy was verified by mapping PacBio, Illumina, and RNA-seq reads to the final assembly using Minimap2 v2.2358 and SAMtools v1.959, achieving mapping rates of 99.99%, 88.28%, and 97.86%, respectively. These results demonstrate the high completeness and accuracy of the O. cuvera genome assembly.