Nuclear and mitochondrial genomes of Polylopha cassiicola: the first assembly in Chlidanotinae (Tortricidae)

Abstract

Tortricidae is one of the largest families in Lepidoptera, including subfamilies of Tortricinae, Olethreutinae, and Chlidanotinae. Here, we assembled the gap-free genome for the subfamily Chlidanotinae using Illumina, Nanopore, and Hi-C sequencing from Polylopha cassiicola, a pest of camphor trees in southern China. The nuclear genome is 302.03 Mb in size, with 36.82% of repeats and 98.4% of BUCSO completeness. The karyotype is 2n = 44 for males. We identified 15412 protein-coding genes, 1052 tRNAs, and 67 rRNAs. We also determined the mitochondrial genome of this species and annotated 13 protein-coding genes, 22 tRNAs, and one rRNA. These high-quality genomes provide valuable information for studying phylogeny, karyotypic evolution, and adaptive evolution of tortricid moths.

Background & Summary

Tortricidae, the leafroller moths, is one of the largest families of Lepidoptera (butterflies and moths)¹, including numerous notorious economic pests such as the spruce budworm, Choristoneura fumiferana², oriental fruit moth Grapholita molesta³ and codling moth, Cydia pomonella⁴. The two main subfamilies are Tortricinae and Olethreutinae, which are relatively young⁵, comprising over 95% of tortricid species. Genomes of many species in these two subfamilies have been determined⁶, revealing an ancestral sex chromosome-autosome fusion and two subsequent autosome fusions relative to the ancestral karyotype of Lepidoptera⁷. Compared to the two successful subfamilies, the relict subfamily Chlidanotinae is much more limited in distribution range, host range, species richness, and population size. Species of this subfamily are mainly distributed in tropical regions, indicating varied climatic adaptability compared to species of the other subfamilies. Thus, this group can provide valuable insights into the phylogeny and pest adaptation and evolution of Tortricidae. However, no genome has been assembled for species of Chlidanotinae.

Here, we present the first chromosome-level genome assembly and annotation in the Olethreutinae using high-coverage long-read and Hi-C sequencing from Polylopha cassiicola⁸. This species is mainly distributed in the southern coastal regions of China and Southeast Asia. It is a pest of trees Cinnamomum cassia and C. camphora. We also assembled the mitochondrial genome of this species from the Illumina short sequencing reads. These genomes are expected to provide information for understanding the phylogeny, karyotypic, and adaptive evolution of Tortricidae.

Methods

Sample collection and sequencing

P. cassiicola larvae were collected from the tops of C. camphora in Guangxi, China. The larvae were reared in the laboratory to pupae and adults for genomic and transcriptome sequencing. Three individuals were used for three types of genome sequencing: one male pupae for Nanopore long-read sequencing, one male pupae for Illumina short-read sequencing, and one female adult for Hi-C sequencing. In addition, four larvae were used for RNA sequencing. Nucleic acid extraction and sequencing libraries was contracted by BerryGenomic (Beijing, China). Methods for nucleic acid extraction, platforms for sequencing, and sequencing outputs are provided in Table 1.

Table 1.

Methods and outputs for sequencing experiments.

Experiment	Method/Platform	Manufacturer	Insertion size	Output	Coverage
DNA extraction	Magnetic bead method	Invitrogen, Thermo Fisher Scientific, USA	NA	NA	NA
RNA extraction	TRIzol reagent	Thermo Fisher Scientific, USA	NA	NA	NA
Short-read seq	NovaSeq 6000; paired-end	Illumina, USA	350 bp	68.7 Gb	115×
Long-read seq	PromethION	Oxford Nanopore Technologies, UK	N50 = 16.7 Kb	117.6 Gb	196×
Hi-C seq	NovaSeq 6000; paired-end; digested by MboI	Illumina, USA	350 bp	178.1 Gb	297×
RNA seq	NovaSeq 6000; paired-end	Illumina, USA	350 bp	16.3 Gb	NA

NA, not available.

Genome assembly

The Nanopore long reads were assembled into 76 contigs using NextDenovo 2.5.2 (https://github.com/Nextomics/NextDenovo) with parameters: “read_cutoff = 4k, genome_ size = 400 m, nextgraph_options = -a 1”. Redundant sequences in contigs were removed using Purge_dups⁹. The cleaned contigs containing 65 sequences were then assembled to chromosome-level using Hi-C information. In this analysis, we mapped the Hi-C reads to cleaned contigs using BWA¹⁰ with options: “mem -SP5”, anchored contigs using YaHS 1.2a.1¹¹ with option: “-e GATC”, and manually adjusted using Juicerbox 1.22.01¹². We removed the contigs that did not have any contact information with the chromosomes, which could be from potential contamination, such as symbiotic microbes. At last, the chromosomal-level genomic sequences were subjected to two rounds of long-read polishing and two rounds of short-read polishing using Nextpolish 1.4.1¹³. The obtained P. cassiicola genome is 302.03 Mb in size and contains 21 autosomes and one Z sex-chromosome (Fig. 1a).

Fig. 1.

Genomic feature of nuclear genome of Polylopha cassiicola. (a) Hi-C contact matrix of 22 putative chromosomes. (b) Synteny among four tortricid species from four subfamilies and an outgroup. The labels at the bottom marked the ancestral linkage groups of Lepidoptera⁶.

We also assembled mitochondrial genome using MitoZ 3.6¹⁴ based on the short-reads. In the mitochondrial genome, we identified 13 protein-coding genes, 22 tRNAs, and 1 rRNA (Fig. 2).

Fig. 2.

Distribution of annotated genes on mitochondrial genome. The inner ring shows the relative read coverage.

Genome synteny

We analysed the chromosomal synteny between P. cassiicola and three other species from Tortricidae and one from Sesiidae: Choristoneura fumiferana (Tortricidae: Tortricinae)², Grapholita molesta (Tortricidae: Olethreutinae)³, Tortricodes alternella (Tortricidae: Tortricinae; NCBI GenBank assembly: GCA_947859335.1¹⁵), and Sesia bembeciformis (Sesiidae: Sesiinae)¹⁶. Synteny analysis was conducted using MSCANX pipeline in JCVI utility libraries¹⁷. We assigned names of the ancestral linkage group in Lepidoptera⁶ (Merian elements, M1-31 and MZ) based on chromosomal homology. The results show different patterns of chromosomal fusion in species T. alternella and P. cassiicola (Fig. 1b).

Repeat element and non-coding RNA annotation

Repeat elements were detected using RepeatMasker 4.1.5¹⁸ with options “-no_is -norna -xsmall -q”. This analysis was conducted against three databases: Repbase (http://www.girinst.org), Dfam database1 specific to Arthropoda, and a species-specific repeat library constructed using RepeatModeler2¹⁹. Transfer RNA (tRNA) was predicted by tRNAscanSE 2.0.12²⁰ with default parameters, and ribosome RNA (rRNA) was predicted using Barrnap 0.9 (https://github.com/tseemann/barrnap). In the P. cassiicola genome, 36.82% of bases were annotated as repeat elements (Table 2). We identified 67 rRNAs, and 1052 tRNAs (Table 2).

Table 2.

Statistics of repeat elements and non-coding RNAs in Polylopha cassiicola genome.

Item	Number	Length (bp)	Content (%)
SINEs	5219	354577	0.12
LINEs	143490	16636355	5.51
LTR elements	35392	11777710	3.9
DNA transposons	23636	4498331	1.49
Rolling-circles	315217	39695809	13.14
Unclassified repeats	235177	34162427	11.31
Satellites	17	2222	0
Simple repeats	72367	3737369	1.24
Low complexity repeats	7520	355233	0.12
rRNAs	67	46500	0.015
tRNAs	1052	78802	0.026

SINEs, short interspersed nuclear elements; LINEs, long interspersed nuclear elements; LTR, long terminal repeat.

Gene prediction and functional annotation

Gene structure was predicted using an ab initio method, Helixer²¹, with options: “–subsequence-length 320760–batch-size 6”, and with a pre-trained model for invertebrate “invertebrate_v0.3_m_0200”. Gene function, Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) items for predicted genes were annotated using eggNOG-Mapper²² web tools, against the eggNOG Database 5. A total of 15412 protein-coding genes were predicted, in which 12671 genes were functionally annotated.

Data Records

The Nanopore reads, Illumina reads, Hi-C reads, and RNA reads for P. cassiicola genome assembly were deposited at NCBI under Sequence Read Archive under accession number SRP479759²³. The nuclear and mitochondrial genome assemblies were deposited in Genbank under accession number GCA_038024825.1²⁴. The genome annotation files are available in Figshare²⁵ at 10.6084/m9.figshare.24902046.

Technical Validation

To validate the accuracy of the final genome assembly, we mapped the Illumina short reads and Nanopore long reads to the P. cassiicola genome using Minimap2²⁶ with option “-ax sr” for short reads and option “-ax map-ont” for long reads. The mapping rates for the short reads and long reads were calculated using Samtools²⁷. Analysis revealed 96.38% and 98.73% mapping rates for the short and long reads, respectively. We examined the coverage of short reads along the mitochondrial genome and showed 100% coverage (Fig. 1b).

Completeness of the assembly and gene prediction were evaluated using BUSCO 5.4.7²⁸ with lepidoptera_odb10 database. In this analysis, BUSCO examined the states and proportions of 5,286 single-copy orthologous of Lepidoptera in our genome assembly: single-copy (S), duplication (D), fragment (F), and missing (M). The analyses showed completeness ranging 95.1%–98.4% for each assembly stage (Table 3), and 97.8% for predicted gene set: “C: 97.8% [S: 97.2%, D: 0.6%], F: 0.9%, M: 1.3%”. Quality of gene prediction was manually evaluated using RNA-seq data. Specifically, RNA-seq reads were mapped to the genome using Hisat²⁹ and Samtools²⁷. We imported the obtained BAM file and annotation file into the IGV browser³⁰. Based on manual examination, we found that the machine learning-based annotation method has predicted a near-complete gene structure. These results indicate that we have obtained a high-quality assembly and annotation for P. cassiicola genome.

Table 3.

Statistics of Polylopha cassiicola assemblies.

Item	Contig	Purged contig	Hi-C raised scaffold	Polished scaffold
No. of contigs	76	65	22	22
Size (Mb)	297.20	294.45	294.46	302.03
N50 (Mb)	8.54	8.54	12.96	13.19
GC content	35.16%	35.12%	35.12%	35.14%
Single-copy BUSCOs	94.7%	94.8%	94.9%	98.1%
Duplicated BUSCOs	0.5%	0.3%	0.3%	0.3%
Fragmented BUSCOs	2.2%	2.2%	2.2%	0.3%
Missing BUSCOs	2.6%	2.7%	2.6%	1.3%

Author contributions

S.W. designed the study. L.C., Y.Y. and J.C. contribute to the materials of this study. F.Y. and W.S. analysed the data. F.Y. wrote the manuscript. S.W. revised the manuscript.

Code availability

No custom scripts or code were used in this study.

Competing interests

The authors declare no competing interests.