Background & Summary

The striped fruit fly, Zeugodacus scutellata (Diptera: Tephritidae), is one of the most destructive pests of Cucurbitaceae plants1. The species can be morphologically distinguished from other Zeugodacus spp. by the presence of three postsutural yellow vittae with a pair of medium-sized circular black spots across the face, two pairs of scutellar bristles, and the costal band slightly enlarged at the apex2. This fly species shows two annual peaks in adult abundance and survives the winter in the adult stage under leaf litter near its habitat3,4,5. Its larvae infest flowers of pumpkin and cause >70% damage during an outbreak periods, posing a serious threat to crop production6. In addition, Z. scutellata can also damage flowers, fruits and stems of other host plants, such as squash, Chinese cucumber, cucurbit, wild gourds, eggplant and pear, thus causing agricultural losses7,8. This pest is also an important quarantine species, mainly distributed in Asian countries including China, Japan, Korea, India, Thailand, Bhutan and Malaysia9,10. In 1923, Z. scutellata was first reported in Jiangsu Province of China, and currently it has been found in many regions of Southern, Central and Southwest China11. Field monitoring indicated that the abundance of this species increased in Southern of China12.

In recent years, whole-genome sequencing of fruit fly species has greatly advanced our understanding of their evolution, ecology, and pest biology. High-quality reference genomes are now available for several economically important tephritid species, such as Bactrocera dorsalis13 and Bactrocera tau14, as well as for multiple model Drosophila species. These genomic resources have enabled detailed investigations into gene family evolution, host adaptation, chemosensory perception, detoxification pathways, insecticide resistance, and population structure, and have provided essential foundations for comparative genomics and pest management strategies. However, despite its economic importance and expanding geographic distribution, genomic resources for Z. scutellata remain lacking, which has limited systematic studies of its genetic basis of host use, environmental adaptation, and population dynamics. Whole-genome sequencing of Z. scutellata is therefore necessary to fill this critical gap. A chromosome-scale reference genome provides a comprehensive framework for accurate gene prediction and functional annotation, resolves repetitive and heterozygous genomic regions, and enables robust comparative analyses with other tephritid fruit flies. Such a resource is essential for identifying genes and pathways associated with key biological traits, including chemosensation, development, reproduction, detoxification, and stress response, and for supporting population genomic studies relevant to quarantine monitoring and pest control. The genome presented in this study thus represents an important addition to fruit fly genomic resources and provides a valuable foundation for future evolutionary, ecological, and applied research on Z. scutellata.

Methods

Sample preparation

The striped fruit fly samples of Z. scutellata were collected from luffa fields in Shuangjin Village, Jiangnan Subdistrict, Yongkang City, Zhejiang Province (28.884365°N, 120.008902°E) on July 12 and August 27, 2024, respectively. After collection, the flies were reared in a controlled climate chamber at a temperature of 27 ± 1 °C, relative humidity of 65 ± 5%, and a photoperiod of 14 L:10D. During rearing, adult Z. scutellata were fed an artificial diet consisting of a 1:1 mixture of beer yeast powder and sucrose. Once approximately 200 adult flies had been collected, individuals were frozen in liquid nitrogen and stored at −80 °C for subsequent analyses.

Genome survey and estimation

Genomic DNA (gDNA) was extracted from whole bodies of approximately ten adult female Z. scutellata individuals using a cetyltrimethylammonium bromide (CTAB) protocol. DNA integrity and purity were assessed via three complementary methods: (1) 1% agarose gel electrophoresis to evaluate fragment size distribution and check for RNA contamination; (2) spectrophotometric measurement with a NanoDrop™ instrument to determine A_260/A_280 and A_260/A_230 ratios; and (3) fluorometric quantification with a Qubit™ 2.0 system to obtain accurate concentration values. Samples that met DNA quality and quantity criteria, including sufficient concentration, high purity (A260/280 ≈ 1.8), and minimal degradation as assessed by agarose gel electrophoresis, were used for library preparation.

For genome survey sequencing, 1 µg of high-quality genomic DNA was fragmented to an average size of approximately 350 bp by ultrasonication using a Covaris system, in which controlled acoustic energy generates shear forces that randomly break DNA molecules, following the manufacturer’s recommended protocol. Libraries were constructed with the TruSeq Nano DNA HT Sample Preparation Kit (Illumina, USA) according to the manufacturer’s instructions. Briefly, fragmented DNA underwent end‐repair, 3′ adenylation, and ligation to indexed adapters, followed by PCR amplification and AMPure XP bead purification. Library size distribution was verified on an Agilent Bioanalyzer, and concentration was first estimated by Qubit™ before precise quantification via qPCR. Only libraries with an effective concentration >2 nM and an insert size centered at ~350 bp were sequenced. Paired‐end (2 × 150 bp) runs were performed on an Illumina NovaSeq 6000 platform (SmartGenomics Technology Institute, Tianjin, China), yielding 42.82 Gb of raw data (Table 1).

Table 1 Statistics of the sequencing data.
Full size table

Raw reads were filtered with fastp15 (v0.23.2) to remove adapter sequences, reads containing >5 bp of adapter contamination, reads with ≥15% of bases below Q19, reads with >5% ambiguous bases (N), and orphaned mates. Quality‐filtered reads were then used for k–mer–based genome characterization. Jellyfish16 (v1.0.0) was employed to count 17-mers, and GCE17 (v1.0.2) was applied to estimate genome parameters. The 17-mer distribution exhibited a peak depth of 92, corresponding to a raw genome size of 406.84 Mb. After correction for sequencing errors and duplications, the estimated genome size was 404.37 Mb, with a heterozygosity rate of 2.31% and a repeat duplication rate of 38.70% (Table S1; Figure S1).

Pacbio HiFi sequencing and contig assembly

High–molecular-weight genomic DNA was extracted from approximately twenty adult female Z. scutellata individuals and further purified using AMPure PB beads (PacBio, Cat. No. 100-265-900) to eliminate contaminants. Library construction followed the SMRTbell Express Template Prep Kit 3.0 protocol (Pacific Biosciences). Briefly, genomic DNA was sheared to ~15 kb fragments using the Megaruptor system (Diagenode, Cat. No. B06010001), followed by size selection with AMPure PB beads, damage repair, and end polishing. SMRTbell adapters were then ligated to the processed fragments, which were further size-selected using a BluePippin system (Sage Science) to enrich the desired insert range. Library quality and concentration were validated using an Agilent 2100 Bioanalyzer (Agilent Technologies).

Sequencing was carried out on the PacBio Sequel II/IIe platform using SMRT Cells with a 30-hour run time, generating 19.72 Gb of raw long reads (Table 1). Raw polymerase reads were filtered using the SMRT Analysis suite to remove low-quality sequences (<50 bp, read quality <0.8, and self-ligated adapter artifacts). Circular Consensus Sequencing (CCS) reads were produced using SMRT Link v9.0 with parameters --min-passes=3 and --min-rq = 0.99, yielding high-fidelity (HiFi) reads.

Genome assembly was conducted using hifiasm18 (v0.16.1), which generated a set of primary contigs. These contigs served as the foundation for downstream scaffolding and polishing, ultimately resulting in a high-quality, chromosome-scale genome assembly.

Hi-C sequencing and chromosome-scale scaffold assembly

To generate chromosome‐scale scaffolds, we constructed a Hi‐C library from approximately twenty adult female individuals following a proximity‐ligation protocol. Intact nuclei were crosslinked with 1% formaldehyde, quenched with glycine, and then digested overnight with HindIII restriction enzyme. After end repair and biotin labeling of DNA termini, spatially proximate fragments were ligated using T4 DNA ligase to capture long‐range interactions. Crosslinks were reversed by proteinase K digestion, and the resulting DNA was purified. The purified DNA was sheared to 300–500 bp, and biotinylated junctions were enriched on Dynabeads® M‐280 Streptavidin (Life Technologies) prior to library construction. Paired‐end 150 bp sequencing on the Illumina NovaSeq 6000 platform yielded 55.8 Gb of raw Hi‐C data (Table 1).

Hi‐C reads were trimmed and filtered as described for Illumina short‐read data and aligned to the preliminary contig assembly using HiCUP19 (v0.8.0). Uniquely mapping read pairs proximal to predicted HindIII sites were retained to generate a contact matrix. Contigs were polished twice with Pilon20 (v1.23) using the Illumina short reads, producing a draft assembly of 701.8 Mb in 685 contigs (N50 = 1.73 Mb; Table 1). We then applied ALLHiC21 (v0.9.8) to cluster, order, and orient contigs into linkage groups, using the Hi‐C contact map to resolve miss joins and confirm scaffold structure. Manual curation with Juicebox22 refined scaffold orientations and improved contact map consistency. The final assembly anchored 105 contigs (93.52% of the genome) into 6 chromosome‐scale scaffolds, achieving a scaffold N50 of 75.5 Mb (Fig. 1; Table 2 and S2).

Fig. 1
Fig. 1
Full size image

Chromosome-scale Hi-C contact map and interspecific synteny of Z. scutellata. (a) Hi‐C interaction heatmap for the six assembled chromosomes of Z. scutellata. Strong intra‐chromosomal contacts appear as prominent diagonal blocks, whereas inter‐chromosomal interactions are comparatively weaker. Color intensity reflects normalized contact frequency. (b) Genome‐wide collinearity between Z. scutellata and Z. cucurbitae as determined by Satsuma. Syntenic blocks illustrate conserved genomic segments, highlighting structural conservation between the two Tephritid species.

Table 2 The genome assembly informations of Z. scutellatus.
Full size table

Genome assembly quality assessment

Assembly completeness was assessed with BUSCO23 (v5.4.3) against the insecta_odb10 lineage dataset, which comprises 1,367 universal single‐copy orthologs. The analysis recovered 99.3% of BUSCO genes as complete, including 98.2% complete single‐copy and 1.1% complete duplicated orthologs, with only 0.5% fragmented and 0.2% missing (Table S3). These metrics demonstrate that our chromosome‐scale assembly is exceptionally complete and provides a robust foundation for downstream functional annotation, comparative genomics, and population‐level investigations.

Sexual chromosome identifies

To determine the sex chromosome of Z. scutellata, we performed a genome‐wide synteny analysis using Satsuma24 (v2.0), comparing our assembly to that of Z. cucurbitae (GenBank accession GCF_028554725.1; karyotype 5 + XY). Syntenic blocks were visualized with Circos25 (v0.69), which revealed extensive collinearity between chromosome 1 of Z. scutellata and the X chromosome of Z. cucurbitae (Fig. 2). This conserved syntenic relationship provides strong evidence that chromosome 1 in our assembly corresponds to the X chromosome in Z. scutellata.

Fig. 2
Fig. 2
Full size image

Circos plot summarizing genome features of Z. scutellate. A genome-wide overview was generated using a 100 kb sliding window. Concentric tracks from the outermost (IV) to the innermost (I) circle denote: IV. Chromosome ideograms: each colored segment corresponds to one of the nine assembled chromosomes, with scale bars indicating genomic position in megabases. III. Protein-coding gene density: normalized counts of annotated genes per window, plotted as a histogram to highlight gene-rich and gene-poor regions. II. GC content density: proportion of guanine + cytosine bases calculated per window, displayed as a line plot to reveal regional GC variation. I. Repeat element density: percentage of bases masked as repetitive sequences in each window, shown as a heat-map track to illustrate the distribution of transposable elements and other repeats.

Genome annotation

High–depth transcriptomic data were generated from total RNA extracted from a single adult female Z. scutellata using the Qiagen RNeasy Plus Mini Kit. RNA integrity was confirmed on an Agilent 2100 Bioanalyzer (RIN > 7.0), and stranded mRNA libraries were constructed with the Illumina TruSeq Stranded mRNA Library Prep Kit. Paired-end (2 × 150 bp) sequencing on the NovaSeq 6000 platform yielded 7.38 Gb of clean transcriptome reads (Table 1), which provided empirical evidence for gene model refinement.

Repetitive elements were annotated with the HiTE26 pipeline, which integrates homology-based searches against known repeat libraries and de novo discovery algorithms. Approximately 34.38% of the Z. scutellata genome was classified as repetitive, comprising 3.24% long terminal repeats (LTRs), 5.88% long interspersed nuclear elements (LINEs), 0.07% short interspersed nuclear elements (SINEs), 15.50% DNA transposons, and 0.42% unclassified repeats (Table 3).

Table 3 Classification of repetitive sequences in Z. scutellatus genome.
Full size table

Protein-coding genes were predicted using eGAPX (v0.3.2, https://github.com/ncbi/egapx), combining ab initio algorithms with RNA-Seq alignments to accurately delineate exon–intron structures and untranslated regions. This produced 13,327 gene models encoding 14,117 transcripts. Predicted proteins were functionally annotated via BLASTP searches against NCBI NR and UniProt Swiss-Prot databases (E-value < 1 × 10−⁵), and assigned Gene Ontology (GO) terms and KEGG pathway identifiers to infer biological roles and metabolic functions (Table S4).

Non–coding RNAs were identified following repeat masking and removal of protein-coding loci. Transfer RNAs were detected with tRNAscan-SE27 (v2.0) under eukaryotic parameters. Ribosomal RNA genes were annotated by BLASTN alignment to curated invertebrate rRNA references (E-value < 1 × 10−5). Small nuclear and nucleolar RNAs were predicted using INFERNAL28 (v1.1.4) against Rfam29 14.9 profiles. In total, we annotated 432 tRNAs, 2,466rRNAs, 83 snoRNAs, and 65 miRNAs (Table 4). Together, these comprehensive annotations establish a robust genomic framework for functional and comparative studies in Z. scutellata.

Table 4 Classification of non-coding RNAs in Z. scutellatus genome.
Full size table

Data Records

The genome sequencing and annotation data generated in this study have been deposited in multiple public repositories. At the National Genomic Data Center (NGDC), the genome assembly is available under the accession number of GWHGEAZ00000000.130 in the Genome Warehouse database31,32 (GWH, https://ngdc.cncb.ac.cn/gwh), with raw sequencing reads deposited in the Genome Sequence Archive33 (GSA, https://ngdc.cncb.ac.cn/gsa) under BioProject PRJCA041109. Specifically, Pacbio Hifi (CRR188411134), Illumina (CRR188411035), Hi-C (CRR188411236) and RNA-seq (CRR188411337) datasets are accessible through the GSA. In addition, all sequencing data have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject PRJNA122541838. The Illumina, PacBio, and Hi-C datasets are available under accession numbers SRR3522705439, SRR3522705540, and SRR3522705641, respectively, while the RNA-seq data are deposited under accession number SRR3522705342. The chromosome-scale genome assembly is available in GenBank under accession JBLXXD000000000, with the corresponding assembly accession GCA_051201265.143. Furthermore, the assembly genome and annotated gene sets have also been deposited in the FigShare repository44 to facilitate open access and reuse by the research community.

Technical Validation

Sequencing data quality

The quality of raw Illumina paired-end reads was assessed using FastQC, confirming high base-calling accuracy and the absence of technical artifacts. Over 93% of bases achieved a Phred quality score of Q30 or above, and the GC content was uniformly distributed (~38.4%), consistent with characteristics of Tephritid genomes. After adapter trimming and quality filtering with fastp, a total of 41.44 Gb of high-quality short-read data was retained. PacBio HiFi sequencing produced long reads with a minimum of three polymerase passes (min-RQ ≥ 0.99), yielding a median base accuracy exceeding 99.5% (Table 1). This high-fidelity dataset provided a solid foundation for accurate and contiguous genome assembly.

Assembly consistency and completeness

K-mer analysis of the Illumina data using 17-mers revealed a distinct peak at depth 92. Genome size and heterozygosity were estimated using GCE, resulting in a predicted genome size of 404.37 Mb and a heterozygosity rate of 2.31%. These estimates closely matched the final Hi-C–scaffolded assembly size of 408.2 Mb, indicating minimal loss due to collapsed heterozygous regions or overrepresentation of repetitive sequences (Figure S1, Table S1). Hi-C contact matrices generated during scaffolding with ALLHiC revealed strong intra-chromosomal interaction patterns with minimal off-diagonal noise, confirming correct contig orientation and placement across six pseudochromosomes (Fig. 1). Additionally, synteny analysis demonstrated high chromosomal collinearity with Z. cucurbitae, further validating the structural integrity of the assembly (Fig. 2). BUSCO analysis using the insecta_odb10 dataset recovered 99.3% of expected conserved genes, including 98.2% single-copy and 1.1% duplicated BUSCOs, with only 0.7% missing or fragmented. This result indicates near-complete coverage of both coding and non-coding genomic regions (Table S3).

Annotation validation

Transcriptomic validation of gene models was performed by mapping RNA-Seq reads to the genome. The mapping achieved an overall alignment rate exceeding 95.4%, with more than 86% of reads spanning annotated exon–intron boundaries, confirming strong transcriptomic support. Annotation Edit Distance (AED) scores further evaluated the quality of gene models, with 92% of predicted loci exhibiting AED values ≤ 0.25, indicating high consistency between predicted and empirically supported gene structures. A total of 13,327 protein-coding genes were predicted in the Z. scutellatus genome, of which 91.58% (12,205 genes) were successfully annotated in the NCBI non-redundant (NR) database, indicating a high degree of sequence homology with known proteins. Swiss-Prot and Pfam annotations covered 53.67% and 82.91% of the genes, respectively, reflecting both curated functional matches and conserved protein domains. Additionally, 84.62% of genes were assigned to KOG categories, and 72.67% were associated with Gene Ontology (GO) terms, supporting the structural and functional integrity of the gene models. KEGG pathway annotation was assigned to 59.31% of the genes, enabling reconstruction of key metabolic and signaling pathways. These results collectively confirm the robustness of gene prediction and highlight the biological relevance of the Z. scutellatus gene repertoire (Table S4). Non-coding RNAs and repetitive elements were annotated using Rfam and Repbase, respectively, ensuring comprehensive identification of small RNAs and transposable elements (Table 4).