A chromosome-level reference genome assembly and annotation of Serrodes campana

Abstract

Serrodes campana (Lepidoptera: Noctuidae) is an economically important moth distributed in subtropical and tropical Asia, but genomic resources for this species have been lacking. Here, we present a chromosome-level reference genome assembly for S. campana generated using PacBio HiFi long reads, Illumina short reads, Hi-C sequencing, and transcriptomic data from 12 RNA-seq samples. The final assembly spans 627.9 Mb, with a contig N50 of 21.4 Mb and a scaffold N50 of 21.6 Mb, and 99.29% of the assembled sequences were anchored to 31 pseudochromosomes. BUSCO assessment recovered 99.56% complete insect genes, supporting the high completeness of the assembly. We predicted 16,047 protein-coding genes, of which 96.65% were functionally annotated, and identified repetitive elements accounting for 33.79% of the genome. This genome assembly and associated annotation resources provide a foundation for gene annotation, comparative genomics, genome evolution studies, and investigations of host-associated biology in S. campana and related moth species.

Background & Summary

Noctuidae is one of the most diverse and ecologically important families within Lepidoptera and includes numerous species with substantial agricultural and forestry impacts^1,2. Many noctuid moths are recognized as pests because of their broad host range, strong feeding capacity, and high reproductive potential^2,3. As high-throughput sequencing technologies have advanced, genome-scale resources have become increasingly useful for studying the biology, evolutionary history, and ecological adaptation of noctuid insects^4–8. In particular, chromosome-level genome assemblies facilitate comparative genomic analyses and provide a framework for the annotation of genes and gene families associated with host use, chemosensory perception, detoxification, development, and other ecologically relevant traits. However, such assemblies remain available for only a limited number of noctuid species. More broadly, chromosome-scale genome resources reported for Lepidoptera, including Cydia pomonella⁵, Hermonassa cecilia⁹, Paracymoriza distinctalis¹⁰, and Hestina assimilis¹¹, illustrate the utility of highly contiguous assemblies for describing genome organization and supporting downstream analyses of genome structure and gene repertoire evolution.

Serrodes campana is a noctuid moth distributed in tropical and subtropical regions of Asia and extending to northern Australia, with records mainly from southern and southwestern China^12–14. This species is associated with distinct host-use patterns across life stages. Larvae mainly feed on Sapindus mukorossii, consuming young leaves, tender shoots, inflorescences, and young fruits, which can result in defoliation, fruit drop, reduced plant vigor, and yield loss¹⁴. Adults also act as fruit-piercing moths and feed on the juices of several economically important fruits^12,15. Previous work on S. campana has focused primarily on field occurrence, morphology, life history, interception records, and chemical control^12–18, whereas genomic resources for this species have remained unavailable. The absence of a reference genome has limited comparative analyses of gene content and genome structure in this species and constrained efforts to document genes related to host plant use, chemosensory perception, nutrient acquisition, and other life-history-associated functions. Basic morphology across major developmental stages and representative larval damage are shown in Fig. 1 to provide biological context for the dataset.

Fig. 1.

Morphology and hazard of Serrodes campana. (A) Egg. (B) 6th instar larva. (C) Pre-pupa and Pupa. (D) Female adult. (E) Male adult. (F) Field damage by larva.

Recent advances in long-read sequencing, particularly PacBio HiFi sequencing, have improved the continuity and base-level accuracy of genome assemblies, including those of insect genomes with abundant repetitive elements and substantial heterozygosity. In parallel, transcriptomic data from multiple developmental stages and tissues provide evidence that can improve gene prediction and structural annotation. Here, we describe a chromosome-level genome assembly and annotation dataset for S. campana. By integrating PacBio HiFi sequencing with RNA-seq data from multiple developmental stages and tissues, we generated a highly contiguous genome assembly and an annotated gene set for this species. Repetitive elements and non-coding RNAs were also annotated to document the genomic composition of S. campana. These data provide a resource for comparative genomics and for the identification of genes and gene families that may be relevant to host-associated biology, chemosensory systems, and stage-dependent feeding traits in S. campana and related noctuid moths.

Methods

Sample collection and sequencing

Mature larvae of S. campana were collected from Jianning Economic Development Zone, Jianning County, Fujian Province, China (26.87167 °N, 116.85710 °E), in late July 2025 and reared in the laboratory on fresh host-plant tissues until adult emergence. The emerged adults were allowed to mate and oviposit under laboratory conditions. Species identification was confirmed based on morphological characteristics. One F1 male pupa produced in September 2025 was selected for high-molecular-weight genomic DNA extraction, which was subsequently used for library construction and whole-genome sequencing. Three types of sequencing libraries were constructed, including PacBio HiFi long-read libraries, Illumina short-read libraries, and Hi-C libraries.

For transcriptome sequencing, 12 RNA-seq samples representing major developmental stages and adult tissues were collected from the subsequent laboratory-reared F1 generation derived from these eggs. The samples included eggs (Egg.RNA), one fifth-instar larva (LA.RNA), one female pupa (FP.RNA), one male pupa (MP.RNA), one female adult body (FA.RNA), one male adult body (MA.RNA), two biological replicates of female adult antennae (FA.ANT1 and FA.ANT2), two biological replicates of male adult antennae (MA.ANT1 and MA.ANT2), female terminal abdominal gland tissue (FA.GON), and male terminal abdominal coremata (MA.COR). Egg samples consisted of 20 pooled eggs. Female and male antenna samples each consisted of 10 pooled antennae per replicate. Female terminal abdominal gland tissue consisted of tissues pooled from 10 females, and male coremata consisted of terminal abdominal scent-brush structures pooled from 10 males. Pupal samples were collected at the middle pupal stage, corresponding to day 3 after the prepupal stage. Whole-body samples were used for eggs, larvae, pupae, and adult bodies, whereas antennae and terminal abdominal tissues were dissected from adults using sterilized fine scissors and forceps.

All dissected tissues and developmental-stage materials used for RNA sequencing were immediately frozen in liquid nitrogen after collection and stored at −80 °C until RNA extraction. Total RNA was extracted from each sample using TRNzol Universal Reagent (DP424, TIANGEN) according to the manufacturer’s instructions. Poly(A) + mRNA was enriched using oligo(dT) magnetic beads and fragmented prior to cDNA synthesis. RNA-seq libraries were prepared through first- and second-strand cDNA synthesis, end repair, A-tailing, adapter ligation, size selection, and PCR amplification, and then sequenced on the DNBSEQ-T7 platform (DNBSEQ-T7RS, BGI) in paired-end 150-bp mode.

PacBio HiFi sequencing was performed on the Sequel II platform. A total of 2.94 million HiFi reads were generated, yielding 61.45 Gb of clean sequence data, corresponding to approximately 98× coverage based on the final genome assembly size of 627.9 Mb. After filtering using the PacBio SMRT Link pipeline, the HiFi reads had an average read length of 20.9 kb and an N50 length of 21.1 kb, with an average quality value of Q37.8. The GC content of the HiFi reads was 36.23%.

Illumina short-read libraries were sequenced in paired-end mode (PE150) on the Illumina platform. Raw reads were quality-filtered using fastp (v0.23.4)¹⁹. A total of 107.65 Gb of clean Illumina short-read data were retained, corresponding to approximately 171× coverage based on the final genome assembly size, with no evident adapter contamination or low-quality sequences.

Hi-C libraries were also sequenced on the Illumina platform using paired-end sequencing (PE150). Approximately 8.22 × 10⁸ Hi-C reads were generated. After quality filtering, 122.83 Gb of clean Hi-C data were retained, corresponding to approximately 196× coverage based on the final genome assembly size, providing long-range chromatin interaction information for downstream scaffolding.

Genome assembly and chromosome-level scaffolding

Prior to genome assembly, genome characteristics were estimated using Illumina short-read data based on 21-mer frequency analysis with Jellyfish (v2.1.4)²⁰ and GenomeScope 2.0 (v2.0.1)²¹. The genome size of S. campana was estimated to be approximately 614.5 Mb, with a heterozygosity rate of 3.29% and repetitive sequences accounting for approximately 28.3% of the genome (Fig. 2A). These results indicated a relatively heterozygous genome and provided a reference for subsequent assembly evaluation.

Fig. 2.

Genome feature estimation and Hi-C-assisted chromosome-level scaffolding of Serrodes campana. (A) GenomeScope profile based on 21-mer analysis of Illumina short reads, showing the estimated genome size, heterozygosity, and repeat characteristics of the S. campana genome. (B) Hi-C contact map of the chromosome-level genome assembly. Strong interaction signals along the diagonal and clear boundaries between pseudochromosomes support accurate contig ordering and orientation at the chromosome scale.

PacBio HiFi reads were used for de novo genome assembly with hifiasm (v0.19.9-r616)²², which is optimized for high-accuracy long-read data. The initial assembly produced highly contiguous contigs. To improve assembly accuracy, Illumina short reads were aligned to the assembled contigs using BWA (v0.7.18-r1243)²³, and long-read alignments were generated using minimap2 (v2.28-r1209)²⁴. Consensus polishing was first performed using Racon (v1.5.0)²⁵ based on long-read alignments, followed by additional polishing with NextPolish2 (v0.2.1)²⁶ using Illumina short-read data.

Hi-C data were used to anchor contigs into chromosome-scale scaffolds and to validate long-range chromatin interaction consistency. Hi-C reads were mapped to the assembled contigs, and chromatin interaction information was used to order and orient contigs into 31 chromosomes using ALLHiC (v0.9.8)²⁷. Genome-wide Hi-C contact maps exhibited strong interaction signals along the diagonal within each pseudochromosome, with clear boundaries between chromosomes, supporting accurate contig ordering and orientation at the chromosome scale (Fig. 2B).

The final genome assembly had a total size of 627.9 Mb, comprising 98 contigs and 95 scaffolds, with a scaffold N50 of 21.6 Mb and a contig N50 of 21.4 Mb (Table 1). The longest scaffold was 35.1 Mb, and only three gaps were detected in the assembly. The overall GC content of the assembled genome was 36.42%. The final assembly size was broadly consistent with the genome size estimated from K-mer analysis. Summary statistics of the assembled contigs and Hi-C-anchored scaffolds are provided in Table 1.

Table 1.

Summary statistics of the Serrodes campana genome assembly.

Assembly metric	Value
Genome size	627.9 Mb
Number of scaffolds	95
Number of contigs	98
Scaffold N50	21.6 Mb
Contig N50	21.4 Mb
Longest scaffold	35.1 Mb
GC content	36.42%
Number of gaps	3

Core software tools, databases, versions, and their major uses in genome survey, assembly, and annotation are summarized in Table 6.

Table 6.

Core software and databases used for genome survey, assembly, annotation, and validation of Serrodes campana.

Category	Software	Version	Purpose
Read quality control	fastp	v0.23.4	Quality filtering of Ilumina short reads
Genome feature estimation	Jellyfish	v2.1.4	21-mer counting for genome survey
Genome feature estimation	GenomeScope 2.0	v2.0.1	Estimation of genome size, heterozygosity, and repeat content
Long-read assembly	hifiasm	v0.19.9-r616	De novo genome assembly using PacBio HiFi reads
Read alignment	minimap2	v2.28-r1209	Alignment of long reads and Hi-C reads for assembly evaluation and scaffolding
	BWA	v0.7.18-r1243	Alignment of llumina short reads to the assembly
Assembly polishing	Racon	v1.5.0	Consensus polishing using long- read alignments
	nextpolish2	v0.2.1	Polishing of assembled contigs using short-read data
Assembly assessment	BUSCO	v5.7.1	Assessment of genome and gene set completeness
	Reploc	2021/3/12	Repeated sequence annotation
	mcl	v22-282	Establishment of a consensus sequence library for repetitive sequences
	repbase	v26.11	Supplementation of the repetitive sequence library
Repeat masking	RepeatMasker	v4.1.8	Genome-wide annotation of repetitive sequences
LTR identification	LTR_Finder	v1.07	Detection of LTR retrotransposons
LTR refinement	LTR_retriever	v3.0.1	Refinement and classification of LTR elements
RNA-seq alignment	HISAT2	v2.2.1	Alignment of RNA-seq reads to the genome
Gene prediction	AUGUSTUS	v3.5.0	De novo prediction of gene structure
Evidence integration	EvidenceModeler	v2.1.0	Integration of ab initio,transcript, and protein evidence
Functional annotation	BLAST+	v2.15.0+	Sequence similarity search against public databases
	DIAMOND	v2.1.9.163	Comparison of the amino acid sequence of genes with various protein libraries
	HMMER	v3.4	Comparison of the amino acid sequence of the gene with the Pfam database
Orthology annotation	eggNOG-mapper	2.1.12	Functional annotation based on orthology
Pathway annotation	KEGG	v5.0	Assignment of KEGG pathway information based on KO annotation
Non-coding RNA prediction	tRNAscan	2.0.12	ldentification of tRNA genes
Non-coding RNA prediction	Infernal	v1.1.5	Annotation of snRNA and miRNA

Genome assembly quality assessment

The completeness of the chromosome-level genome assembly was assessed using BUSCO (v5.7.1)²⁸ with the insecta_odb10 dataset. BUSCO analyses were performed on both the assembled genome and the predicted protein-coding gene set. The genome assembly recovered 99.56% complete BUSCOs, including 99.20% single-copy and 0.37% duplicated BUSCOs, whereas 0.22% were fragmented and 0.22% were missing (Table 2). For the predicted protein set, BUSCO analysis identified 98.24% complete BUSCOs, including 97.51% single-copy and 0.73% duplicated BUSCOs. Together, these results support the high completeness of both the assembly and the annotated gene set.

Table 2.

BUSCO assessment of the genome assembly and predicted protein set of Serrodes campana.

Dataset	Complete BUSCOs (C)	Single-copy (S)	Duplicated (D)	Fragmented (F)	Missing (M)	Total BUSCOs
Genome assembly	1361 (99.56%)	1356 (99.20%)	5 (0.37%)	3 (0.22%)	3 (0.22%)	1367
Protein set	1343 (98.24%)	1333 (97.51%)	10 (0.73%)	5 (0.37%)	19 (1.39%)	1367

A Circos plot was generated to visualize the chromosomal organization and genome-wide distribution of GC content, genes, and major repeat classes in the assembled genome (Fig. 3).

Fig. 3.

Circos representation of the chromosome-level genome assembly of Serrodes campana. Tracks from outer to inner circles indicate: (a) pseudochromosome coordinates; (b) GC content; (c) gene density; (d) density of all repetitive sequences; (e) density of tandem repeats (TRs); (f) density of long terminal repeat (LTR) elements; and (g) density of DNA transposons. GC content and density values were calculated in 500-kb windows. For density tracks, the color gradient from light to dark indicates increasing normalized density values from 0 to 1.

Repeat annotation and non-coding RNA prediction

Repetitive and non-coding components of the genome were annotated prior to protein-coding gene prediction. Repetitive elements in the S. campana genome were identified using a combination of de novo and homology-based approaches. For de novo repeat identification, candidate repetitive sequences were detected using RepLoc (2021-03-12)²⁹, followed by sequence similarity searches with BLAST + (v2.15.0) and clustering with MCL (v22-282) to construct a species-specific repeat library. For homology-based annotation, known repetitive elements were identified by aligning the genome against the Repbase database (v26.11). The de novo library and Repbase-based annotations were subsequently combined and used for genome-wide repeat masking with RepeatMasker (v4.1.8)³⁰ Long terminal repeat (LTR) retrotransposons were identified using LTR_Finder (v1.07)³¹, and candidate LTR elements were further refined and classified using LTR_retriever (v3.0.1)³².

Repetitive sequences accounted for 33.79% of the assembled genome, with DNA transposons representing the largest category (16.19%), followed by LTR elements (9.09%) and LINEs (5.23%) (Table 3). Genome-wide distributions of major repeat classes are shown in Fig. 3.

Table 3.

Statistics of genome repetitive sequences and non-coding RNA annotation of Serrodes campana.

A. Repetitive Sequence Composition
Category	Type	Length (bp)	Percentage (%)
Repeat	DNA transposon	101,674,364	16.19
Repeat	LTR	57,095,033	9.09
Repeat	LINE	32,831,745	5.23
Repeat	SINE	12,107	0.002
Repeat	Simple repeat	5,112,583	0.81
Repeat	Low complexity	663,404	0.11
Repeat	Unknown	59,162,237	9.42
Total (TEs)		206,838,524	32.94
Tandem repeats	TRF	8,652,536	1.38
Repeat combined		212,168,108	33.79
B. Non-coding RNA annotation results
Type	Number	Length (bp)	Percentage (%)
miRNA	1,107	111,521	0.0178
rRNA	1,029	1,546,292	0.2463
snRNA	259	34,655	0.0055
tRNA	61,342	2,236,725	0.3562
Total	63,737	3,929,193	0.6258

Non-coding RNAs were annotated using tRNAscan-SE (v2.0.12)³³ and Infernal (v1.1.5)³⁴. In total, 63,737 non-coding RNAs were identified.

Gene prediction and structural annotation

Protein-coding genes were predicted using an evidence-based annotation strategy integrating ab initio prediction, transcriptomic evidence, and homologous protein information. RNA-seq reads from the 12 samples described above were aligned to the genome assembly using HISAT2 (v2.2.1)³⁵ to provide transcriptional evidence for gene model construction. Ab initio gene prediction was performed using AUGUSTUS (v3.5.0)³⁶. The different sources of evidence were then integrated using EVidenceModeler (v2.1.0)³⁷ to generate a final, non-redundant gene set.

A total of 16,047 protein-coding genes were predicted, with an average gene length of 21.2 kb and an average of 7.69 exons per gene (Table 4).

Table 4.

Basic features of predicted protein-coding genes in Serrodes campana.

Feature	Value
Total number of protein-coding genes	16,047
Mean gene length	21,218.15 bp
Median gene length	7,105bp
Mean number of exons per gene	7.69
Median exons number	6
Mean exon length	3,113.22 bp
Mean intron length	22,000.53 bp
Mean CDS length	1,662.92 bp
Genes with annotated UTRs	>80%

The average intron length was 22.0 kb. In the gene-structure comparison performed using representative lepidopteran annotation files, this value was comparable to those of Helicoverpa armigera (22.9 kb) and Catocala sponsa (20.9 kb), but greater than those of Lymantria dispar (11.5 kb) and Tineola bisselliella (4.7 kb).

Transcriptome-supported validation of gene models

Transcriptome data from the 12 RNA-seq samples were used to assess expression support for the predicted gene models. Across the 12 RNA-seq samples, most mapped reads were assigned to exonic regions, with exonic proportions ranging from 81.89% to 91.63%. In contrast, intronic and intergenic proportions ranged from 3.82% to 9.55% and from 3.52% to 8.56%, respectively (Fig. 4). In total, 96.7% of predicted genes showed detectable transcriptional signals in at least one sample, providing transcriptomic support for the completeness and reliability of the structural annotation.

Fig. 4.

Distribution of RNA-seq reads mapped to genomic features across different samples. Stacked bar plot showing the proportions of reads mapped to exonic, intronic, and intergenic regions in eggs (Egg.RNA), larvae (LA.RNA), female pupae (FP.RNA), male pupae (MP.RNA), female adults (FA.RNA), male adults (MA.RNA), female antennae replicate 1 (FA.ANT1), female antennae replicate 2 (FA.ANT2), male antennae replicate 1 (MA.ANT1), male antennae replicate 2 (MA.ANT2), female terminal abdominal gland tissue (FA.GON), and male terminal abdominal coremata (MA.COR).

Functional annotation of protein-coding genes

Functional annotation was performed using multiple public databases. Protein sequences were aligned against the NCBI NR and Swiss-Prot databases using DIAMOND (v2.1.9.163)³⁸. Nucleotide-level similarity searches involved in repeat identification were conducted using BLAST + (v2.15.0). Protein domain annotation was performed using HMMER (v3.4)³⁹ against the Pfam database. GO annotations were assigned based on eggNOG (v5.0.2), Pfam2GO, and Gene2GO mappings. Orthology-based annotation was conducted using eggNOG-mapper (v2.1.12), and KEGG pathway annotation was performed using KEGG Mapper (v5.0).

In total, 15,509 genes (96.65%) were annotated in at least one database (Table 5). The overlap of gene annotations among different databases is shown in Fig. 5.

Table 5.

Summary of functional annotation of protein-coding genes in Serrodes campana.

Database	Annotated genes	Percentage (%)
GO	14,549	90.66
KEGG_KO	7,919	49.35
KEGG_all	15,195	94.69
KOG	9,792	61.02
NR	15,390	95.91
PANTHER	11,318	70.53
Pfam	12,207	76.07
Swissprot	11,416	71.14
eggNOG	14,364	89.51
All annotated	15,509	96.65

Fig. 5.

Overlap relationship of annotation results of databases with different functions.

Data Records

All raw sequencing data generated in this study are available under NCBI BioProject PRJNA1419610⁴⁰ and NCBI Sequence Read Archive (SRA) accession SRP675733⁴¹. The SRA dataset includes genomic Illumina short-read data (SRR37461397), Hi-C sequencing data (SRR37461396), PacBio HiFi long-read data (SRR37461395), and RNA-seq datasets (SRR37461387–SRR37461394 and SRR37461398–SRR37461401). The whole genome shotgun project has been deposited at DDBJ/ENA/GenBank under accession JBVVXX000000000⁴². The version described in this paper is JBVVXX010000000. Genome annotation files, including the GFF3 annotation, CDS, and predicted protein sequences, are available at Zenodo under https://doi.org/10.5281/zenodo.19230235⁴³.

Technical Validation

Assembly quality was evaluated using multiple complementary approaches, including read mapping, k-mer-based consensus quality assessment, BUSCO analysis, Hi-C contact consistency, and transcriptome-supported annotation. PacBio HiFi reads and Illumina short reads were independently aligned to the final assembly to assess mapping consistency and coverage uniformity. A total of 99.59% of PacBio HiFi reads and 98.54% of Illumina short reads were mapped to the assembled genome. In addition, 90.28% of Illumina read pairs were properly mapped, further supporting the accuracy and consistency of the assembly.

K-mer-based consensus quality assessment further supported the base-level accuracy of the assembly. The genome-wide QV was 58.69, corresponding to an estimated base error rate of 1.35 × 10⁻⁶. Per-pseudochromosome QV values ranged from 50.46 to 63.87, with estimated base error rates ranging from 4.10 × 10⁻⁷ to 8.99 × 10⁻⁶, indicating consistently high consensus accuracy across the assembled pseudochromosomes. Genome completeness was assessed using BUSCO (v5.7.1)²⁸ with the insecta_odb10 dataset, recovering 99.56% complete BUSCOs in the genome assembly and 98.24% in the predicted protein-coding gene set. Compared with available chromosome-level genome assemblies of noctuid and other lepidopteran insects, the S. campana assembly shows high continuity and completeness. The assembly has a contig N50 of 21.4 Mb and a scaffold N50 of 21.6 Mb, with 99.56% complete BUSCOs recovered using the insecta_odb10 dataset. These values are comparable to, or higher than, those reported for several representative lepidopteran genome assemblies^5,9–11. However, direct comparisons should be interpreted cautiously because sequencing platforms, assembly strategies, polishing procedures, annotation pipelines, and BUSCO datasets may differ among studies. Together with the high read-mapping rates, genome-wide QV of 58.69, and Hi-C anchoring of 99.29% of the assembly to 31 pseudochromosomes, these metrics indicate that the S. campana genome represents a high-quality chromosome-level reference assembly.

Hi-C data were used to assess long-range genome organization and scaffold correctness. The Hi-C contact map showed strong interaction signals along the diagonal within each chromosome-scale scaffold and clear boundaries between scaffolds, supporting accurate contig ordering and orientation. Using Hi-C-based scaffolding, 99.29% of the assembled genome sequence was anchored to 31 pseudochromosomes.

RNA-seq data from 12 samples provided additional support for gene model quality. Across all 12 RNA-seq samples, reads were primarily assigned to exonic regions, with exonic proportions ranging from 81.89% to 91.63%, consistent with the distribution pattern shown in Fig. 4. In addition, 96.7% of the predicted genes showed detectable transcriptional signals in at least one sample. Together, these results support the continuity, consensus accuracy, completeness, and annotation quality of the S. campana genome assembly .

Author contributions

L.L. collected the samples, reared the insects, prepared biological materials, curated the data, and drafted the manuscript. J.W. supervised the study and provided guidance on manuscript preparation and revision. L.J. contributed to study design, provided experimental resources and project support. All authors reviewed and approved the final manuscript.

Data availability

All raw sequencing data generated in this study are available under NCBI BioProject PRJNA1419610 36. The genomic Illumina short-read data (SRR37461397), Hi-C sequencing data (SRR37461396), PacBio HiFi long-read data (SRR37461395), and RNA-seq datasets (37461387-SRR37461394, SRR37461398-SRR37461401) are available through the NCBI Sequence Read Archive (SRA) The whole genome shotgun project has been deposited at DDBJ/ENA/GenBank under accession JBVVXX000000000. The version described in this paper is JBVVXX010000000. Genome annotation files, including the GFF3 annotation, CDS, and predicted protein sequences, are available at Zenodo under https://doi.org/10.5281/zenodo.19230235.

Code availability

All analyses were performed using publicly available bioinformatic software following the developers’ recommended workflows. Software versions and relevant parameters are provided in the Methods and Table 6. Custom R scripts were used for data visualization and figure preparation only, and no novel algorithm or proprietary software was developed in this study.

Competing interests

The authors declare no competing interests.