Chromosome-level genome of the shining chafers Kibakoganea tamdaoensis (Coleoptera: Scarabaeidae: Rutelinae)

Abstract

The beetle subfamily Rutelinae (shining leaf chafers) in the family Scarabaeidae is famous for its colorful body and phytophagous habitat. Among these shining leaf chafers, Kibakoganea tamdaoensis Miyake et Muramoto, 1992 is one with ornamental value because of their enlarged and horn-like mandibles. Here we generated high-quality genome assembly of KTAMD, integrating PacBio HiFi and Hi-C, and annotated with RNA sequence data. The resulting assembly is 605.37 Mb, of which 96.09% is anchored into 10 chromosomes. The scaffold N50 is 54.00 Mb, and BUSCO completeness is 96.6%. Repetitive elements span 49.43% of the genome. Our annotation pipeline identified 984 noncoding RNAs and predicted 12,945 protein-coding genes in the genome (96.40% BUSCO completeness). As the chromosome-scale genome from the subfamily Rutelinae, the KTAMD genome provides important new insights into scarabae genome evolution and provides a new vantage point from which to study the evolution of specialized beetles.

Background & Summary

The subfamily Rutelinae (Coleoptera: Scarabaeoidea: Scarabaeidae) consists of 258 genera and over 4,900 species, distributed all over the world^1–4. Adults of Rutelinae beetles are phytophagous, feeding on flowers and leaves, while the larvae feed on roots and decaying plants⁴. Many species of Rutelinae are agricultural and forestry pests, such as the famous Popillia japonica Newman, 1841⁵. Rutelinae is divided into seven tribes, one of which is the tribe Rutelini¹. Most members in the tribe Rutelini shows wide range of morphological features, such as enlarged mandibles and hind legs, prominent thoracic horns and stunning metallic or iridescent colouration⁶. Among Rutelini, the genus Kibakoganea Nagai, 1984, includes species that have been reported as pests of economically important crops and forest trees, posing potential risks to agriculture, horticulture, and forestry. Additionally, this genus is among the few in Rutelinae with ornamental value, primarily due to its bright metallic coloration and, in some species, enlarged mandibles⁷. The genus Kibakoganea was established firstly as a subgenus of Fruhstoferia⁷, and later raised to the generic status by Hirasawa⁸. This genus could be separated with the other genera in Rutelini by the following characters: elongate oval, almost glabrous and sparsely punctate; without prosternal process; with three outer teeth in protibial; the male with elongated mandible, and small, flattened parameres⁸. Kibakoganea has recorded 16 species, mainly spread in Southeast Asia, including the oriental region of China. Although Kibakoganea consists of species with ecological importance, limited availability of genomic data makes it difficult to understand the genetic mechanisms underlying their development and adaptation. Kibakoganea tamdaoensis Miyake et Muramoto, 1992 (here after: KTAMD) which is only found in the Northern Vietnam (Mt. Tam, Dao, Sapa), Northeastern Laos⁷. This species is one with ornamental appeal due to its brightly coloured exoskeleton and enlarged and horn-like mandibles. To further understand the biology of Kibakoganea we have sequenced and assembled a reference quality genome of KTAMD.

Rutelinae is represented by three genomes in NCBI (assessed: Sep 8, 2024). Popillia japonica Newman, 1841 (assembly ID: GCA_040143775.1)⁹, Chrysina resplendens (Boucard, 1875) (assembly ID: GCA_010091895.2)¹⁰ and Chrysina gloriosa (LeConte), 1854 (assembly ID: GCA_038088585.1)¹¹. Among these, only C. gloriosa is assembled up to chromosome level. The genome of KTAMD, which we have scaffolded to the chromosome level, represents a high-quality addition to the limited genomic data available and will serve as a vital resource for conserving Rutelinae genetic diversity. Furthermore, from an evolutionary perspective, the gene annotation data presented here, particularly the functional gene set, will be invaluable for comparative genomic analyses.

Methods

Sample collection and sequencing

Adult specimens of KTAMD were collected on April 02, 2024, at Rongjiang County, China, 25.9356 N, 108.5175 E, Guizhou province, China (Fig. 1). The samples were initially placed in liquid nitrogen and stored at −80 °C before DNA extraction. One single female adult specimen was used for PacBio HiFi and Illumina whole genome sequencing and Illumina transcriptome sequencing. Another single female adult specimen was used for Hi-C sequencing. Genomic DNA and RNA were extracted using the TIANamp Genomic DNA Kit (Tiangen, China) and TRIzol Reagent, respectively. One PacBio single-molecule real-time (SMRT) library with a 15 kb insert size for genome sequencing. Illumina libraries were prepared with a 350 bp insert size using the TruSeq DNA PCR-Free LT Library Preparation Kit and TruSeq RNA v2 Kit (Illumina, San Diego, CA, USA). The restriction enzyme DpnII was used to digest DNA for the Hi-C assay. Short-read and long-read libraries were sequenced on Illumina NovaSeq. 6000 (150 bp paired-end reads) and PacBio Sequel II platforms, respectively. Finally, we obtained 34.31 Gb (57.17X) of PacBio HiFi reads, 29.86 Gb (49.77X) of Illumina PE150 reads, and 93.61 Gb (156.01X) of clean Hi-C reads for assembly. For RNA, one Illumina library was sequenced, from which we obtained 6 Gb Illumina RNA-seq reads (Table 1).

Fig. 1.

Live larva and adult of Kibakoganea tamdaoensis, female. (a) Larva. (b) Adult.

Table 1.

Library sequencing data and methods used in this study to assemble the Kibakoganea tamdaoensis genome.

Sequencing strategy	Clean data (Gb)	Sequencing coverage (x)	Insertion size	Platform	Usage
DNA PacBio HiFi (long-reads)	34.31	57.17X	15–18 Kb	PacBio Sequel II	Genome assembly
DNA Illumina (short-reads)	29.87	49.77X	350 bp	Illumina NovaSeq. 6000	Genome survey
DNA Hi-C	93.61	156.01X	350 bp	Illumina NovaSeq. 6000	Hi-C assembly
RNA Illumina	6	/	350 bp	Illumina NovaSeq. 6000	Anno-evidence

Genome assembly and scaffolding

The resulting DNA Illumina sequence data was subjected to quality control and trimming using Fastp version 0.23.0¹², with the following parameters: quality trimming (>Q20), remove repetitive sequences (-D), trim polyG/X tails (-g -x), proportion of unqualified bases does not exceed 10% (-u 10), reads shorter than 15 bp or with >5 Ns discarded, use overlap regions (overlapping reads), and correct bases (-c). Genome size was estimated using the Illumina short-read sequencing data in the program GenomeScope2 version 2.0¹³ with the k-mer size set as 33 and a maximum k-mer coverage cutoff of 10,000. The genome of KTAMD was predicted to be approximately 533.13 Mb with 47.87% repetitive sequence data and 1.66% heterozygosity.

To assemble the KTAMD nuclear genome, we generated a fastq file from the PacBio HiFi reads bam file using the bam2fq module in SAMtools version 1.10¹⁴, and HiFiasm version 0.19.9-r616^15,16 was used for preliminary genome assembly, using default parameters. The preliminary assembly was polished with two rounds of Illumina short reads using NextPolish version 1.4.1¹⁷. Purge_dups version 1.2.6¹⁸ was used to remove duplications caused by heterozygosity in the assembly. Minimap2 version 2.28-r1209¹⁹ was used for mapping reads during the redundancy removal and short-read polishing steps. Juicer version 1.6²⁰ and 3D-DNA²¹ were used to anchor primary contigs into chromosomes. Juicebox was used to correct errors manually. BUSCO version 5.2.2²² with the database Coleoptera_odb12 was used to evaluate the assembly quality. To detect potential contaminant sequences within the assembly, we used the UniVec and NCBI non-redundant preformatted nucleotide blast databases (accessed: 08-Dec-2022) to perform a blastn-like search using MMseqs. 2 version 11²³ employing the easy-search module with a target sensitivity of 7.5, a maximum e-value threshold of 1e-5, and five hits accepted per query sequence (–s 7.5 –alignment-mode 3 –num-iterations 1 –e 1e-5 –max-accept 5). The resulting chromosome-level genome was 605.37 Mb, comprising 92 contigs, with an N50 of 61.84 Mb and the largest contig size of 72.24 Mb, and the GC content of 34.41% (Table 2). The genome assembly is slightly smaller than Chrysina gloriosa (GCA_038088585.1) (642.3 Mb)¹¹. Upon Hi-C scaffolding, 96.08% of the assembly was anchored on 10 chromosomes (Table S1), which were well-distinguished from each other based on the chromatin interaction heatmap (Fig. 2). In total, the length of pseudochromosomes was 605.37 Mb, and ranged from 13.18 Mb to 72.24 Mb in length (Fig. 3).

Table 2.

Genome assembly and scaffolding statistics of Kibakoganea tamdaoensis.

Characteristics
Genome assembly
Assembly size (Mb)	605.37
Number of scaffolds/contigs	92/207
Longest scaffold/contig (Mb)	72.24/38.50
N50 scaffold/contig length (Mb)	54.00/18.34
GC	36.44%
Anchored to chromosome (Mb, %)	581.68 (96.09%)
BUSCO (%)-contigs
Complete	96.6
Complete single copy	93.7
Complete duplicated	3.0
Fragmented	1.8
Missing	1.6
BUSCO (%)-scaffolds
Complete	96.6
Complete single copy	95.6
Complete duplicated	1.0
Fragmented	1.8
Missing	1.6

Fig. 2.

Heatmap of genome-wide Hi-C data showing 10 chromosomes boundaries of Kibakoganea tamdaoensis. Resolution: 600 kb.

Fig. 3.

Genomic landscape of Kibakoganea tamdaoensis. Circos tracks blocks on the outmost circle represent all 10 chromosomes of K. tamdaoensis. Peak plots from outer to inner circles represent the length of each chromosome, the GC content of each chromosome, protein-coding genes, DNA TEs, and the density of repeat sequences (LINE, long interspersed elements; LTR, long terminal repeat elements; simple repeats), respectively.

Genome annotation

Repetitive elements in the genome of KTAMD were identified by both homology-based and de novo prediction methods. RepeatModeler version 2.0.5²⁴ with the LTR discovery pipeline (-LTRStruct) was used to create a de novo repeat library for KTAMD. A custom repeat library was created by combining the de novo repeat library and the Dfam (release 3.8) (http://www.dfam.org/)²⁵ and RepBase-20181026 databases (https://www.girinst.org/repbase/)²⁶. Repetitive sequences were identified and masked using RepeatMasker version 4.1.2 (http://repeatmasker.org) and the custom library. A total of 299.22 Mb of repeat sequences were identified, constituting 49.43% of the KTAMD genome, including 18.25% interspersed repeats, 0.59% simple repeats, and 0.14% complex repeats. Among interspersed repeats, 5.28% of the genome was retroelements, and 10.81% was DNA transposons. Unclassified repeats accounted for 32.59% of the genome (Table S2). The proportion of repetitive elements was lower than that of C. gloriosa (52.94%).

Noncoding RNAs, including transfer RNAs (tRNAs), microRNAs (miRNAs), ribosomal RNAs (rRNAs), and small nuclear RNAs (snRNAs) were also identified in the KTAMD genome. rRNAs, snRNAs, and miRNAs were detected using the Rfam database (release 14.10, 2023-11-13, 4170 families)²⁷ and the program Infernal version 1.1.5²⁸. The tRNAs were predicted using tRNAscan–SE version 2.0²⁹ with “EukHighConfidenceFilter”. The rRNAs and subunits were predicted using RNAmmer version 1.2³⁰. We identified 984 noncoding RNAs (ncRNAs), which contained infrastructural (housekeeping) and regulatory ncRNAs. The numbers of ribosomal RNA (rRNA), transfer RNA (tRNA), small nuclear RNA (snRNA), microRNA (miRNA), long ncRNA (lncRNA), ribozyme, and other ncRNA were 227, 332, 67, 62, 2, 2, and 291, respectively. The rRNAs included 37 5_8S_rRNA, 93 5S rRNAs, 54 large subunit rRNAs, and 43 small subunit rRNA. The tRNAs had 22 isotypes. The snRNAs were classified into 18 snoRNAs (Small Cajal body specific RNA; 7 CD-box, 1 HACA-box) and nine spliceosomal RNAs containing three minor ones. The miRNAs were classified into 47 families, ribozymes into two families, and the other ncRNAs into 5 families (Table S3).

For structural gene annotation we used MAKER pipeline version 3.01.03³¹, incorporating ab initio prediction, transcriptome-based prediction, and protein homology-based prediction. Previously, RNA-Seq data were mapped to the KTAMD assembled genome sequence using from KTAMD transcriptome generated in this study. Raw reads were adapter trimmed using Hisat2 version 2.2.1³², and then assembled into transcripts using StringTie version 2.1.4³³. For ab initio gene prediction, we used Augustus version 3.4.0 and GeneMarkES/ET/EP version 4.68_lic³⁴. To accurately model the sequence properties, both gene finders were initially trained using the BRAKER version 2.1.6 pipeline³⁵, which uses the mapped transcriptome sequence data. For protein homology-based evidence, we downloaded the protein sequences of Tribolium castaneum (Coleoptera), Chrysina gloriosa (Coleoptera), Chrysina resplendens (Coleoptera), and Popillia japonica (Coleoptera) from NCBI. For the Maker pipeline, the transcripts were provided as input via the “est” option and protein homology-based evidence as input via the “protein” option. And then removed redundant isoforms, kept the longest isoforms, and checked the possible errors for “two mRNAs extracted for single redundant seq”, and deleted proteins of length smaller than 50. The above results were integrated with MAKER version 3.01.03, genes with a start codon and a stop codon were selected to generate the final gene models. The MAKER pipeline predicted 11,956 genes with the mean gene length of 12,292.2 bp. The numbers of exons and introns per gene were 7.41 and 6.41, respectively, and their corresponding mean lengths were 317.76 and 1868.73 bp.

Finally, we performed a functional annotation of the genome. The protein sequences of the genome were searched for homology-based function assignments against the KEGG, NR, TrEMBL, and SwissProt databases using BLASTP v2.2.26³⁶ with an e-value cut-off of 1e-5 (Table 2). Domains in the KTAMD genome using InterProScan version 5.60-92.0³⁷ against the databases (Pfam)³⁸, Smart (http://smart.embl-heidelberg.de/)³⁹, Gene3D version 21.0 (http://gene3d.biochem.ucl.ac.uk/)⁴⁰, Superfamily⁴¹ and Conserved Domains Database (CDD)⁴². And combined above results, 98.69% of the predicted genes were functionally annotated using five functional protein databases (Table 3).

Table 3.

Statistics of functional annotation of Kibakoganea tamdaoensis protein-coding genes.

Functional database	Number of genes annotated (11,956)
Annotated	11,799 (98.69%)
Unannotated	157 (1.31%)
InterPro	11,438 (95.67%)
KEGG	10,622 (88.84%)
SwissProt	8,805 (73.65%)
TrEMBL	11,120 (93.00%)
NR	11,244 (94.04%)

Data Records

The raw sequencing data and genome assembly of KTAMD have been deposited at the National Center for Biotechnology Information (NCBI). The PacBio, Illumina, Hi-C, and transcriptome data can be found under identification numbers SRR31479242⁴³, SRR31479241⁴⁴, SRR31479240⁴⁵, SRR31479243⁴⁶. The assembled genome has been deposited in the NCBI assembly with the accession number JBJNIT000000000⁴⁷. Additionally, the results of annotation for repeated sequences, gene structure, and functional prediction have been deposited in the Figshare database⁴⁸.

Technical Validation

Two methods were used to evaluate the quality of the genome assembly. Firstly, we assessed assembly completeness using BUSCO v5.2.2 with the reference Coleoptera_odb12 BUSCO dataset (n = 3,729). 96.6% of the complete BUSCOs were included in the assembled genome (Table 2). Secondly, we calculated the mapping rate as a measure of assembly accuracy. The mapping rates for PacBio, Illumina, RNA reads were 99.74%, 98.74% and 91.50%, respectively. These evaluations collectively reflect the high quality of the genome assembly produced in this study.

Author contributions

M.B. and R.S. contributed to the research design. C.W. Y.X. C.H. N.L. and S.W. and Y.T. collected the samples. X.Y. Q.D. and T.S. analyzed the data. R.S. performed data analysis and wrote the manuscript. All authors contributed to revising the manuscript. All authors have read and approved the final version.

Code availability

No specific codes or scripts were used in this study. All software used is in the public domain, with parameters clearly described in the Methods section.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41597-025-05657-7.