Comprehensive stage- and tissue-specific transcriptome of the global ecosystem service insect, marmalade hoverfly Episyrphus balteatus

Abstract

Hoverflies (Diptera: Syrphidae) are globally beneficial insects that provide key ecological services, such as pollination and biological pest control. Importantly, extensive research has revealed that numerous species within this group engage in long-distance migrations, thereby enabling these services to operate across broad spatial scales. Hence, the conservation and sustainable use of this functionally important group is crucial for maintaining ecosystem balance and stability, especially in the face of ongoing losses of ecosystem functioning. However, limited genomic and transcriptomic resources hinder the advancement of research on this significant group. To address this gap, we generated a comprehensive developmental- and tissue-resolved transcriptome for Episyrphus balteatus, the dominant and most extensively researched migratory hoverfly species. We sequenced 30 RNA-seq libraries across three life stages (egg, larva, and pupa) and seven adult tissues (antenna, proboscis, head, thorax, body, leg, and wing), yielding 133.87 Gb of clean reads. The subsequent de novo assembly yielded 85,676 unigenes (N50 = 1,028 bp) exhibiting high completeness (97.9% BUSCO), and 45,479 unigenes (53.08%) were functionally annotated. Notably, differential expression analysis identified key gene sets and enriched pathways linked to development, sensory perception, and environmental responsiveness, which are molecular characteristics that may facilitate long-distance migration and ecosystem service delivery. Together, this dataset serves as a valuable community resource for ecological, evolutionary, and comparative research on hoverflies and other beneficial insects.

Background & Summary

Insects, one of the most diverse and abundant groups on Earth, perform essential functions in maintaining ecosystem health and stability. Among them, migratory insects have attracted increasing scientific attention due to their remarkable behavioral flexibility and large-scale seasonal movements, which allow them to influence ecological processes across vast spatial scales^1–5. Notably, although many insect taxa are undergoing severe global declines^6–8, populations of migratory insects have shown relatively stable trends. This resilience suggests that migratory species may play increasingly indispensable ecological roles under ongoing environmental change⁵. These circumstances underscore the pressing need for more comprehensive biological and ecological research to enhance our understanding and management of migratory insect groups.

Within this broader migratory assemblage, hoverflies (Diptera: Syrphidae) represent a notably significant focal group. Their dual roles as key pollinators and effective biological control agents of aphids and other agricultural pests have drawn substantial scientific interest^9–11. Over the past fifty years, hoverfly migrations have been extensively documented worldwide, occurring on every continent except Antarctica^9,11–14. Moreover, accumulating evidence from Europe, China, and other regions demonstrates that migration greatly expands the spatial reach and functional effectiveness of both pollination and pest-suppression services^9,11. Consequently, conducting comprehensive research on these beneficial insects is crucial for a thorough understanding of their ecological potential, particularly in light of the global decline in insect biodiversity and ecosystem services. However, progress toward these goals has been significantly hindered by a lack of essential molecular resources, as existing genomic and transcriptomic datasets for hoverflies remain highly limited.

In this context, we employed transcriptomic approaches to systematically examine Episyrphus balteatus (De Geer, 1776) (Fig. 1), a widely distributed and dominant aphidophagous hoverfly, and it is recognized as the most extensively researched migratory species within this group^9,11,15–17. In detail, we produced a comprehensive, stage- and tissue-specific transcriptome dataset for E. balteatus, covering three developmental stages (egg, larva, and pupa) and seven adult tissues (antenna, proboscis, head, thorax, body, leg, and wing). This dataset not only offers a high-quality molecular resource but also facilitates the investigation of gene expression patterns linked to development, sensory perception, and environmental responsiveness, which are likely to support long-distance migration and ecosystem service delivery. Furthermore, these data establish a basis for comparative and functional genomic research in other beneficial insects and for understanding the genetic mechanisms that support the contributions of migratory insects to ecosystem functioning.

Fig. 1.

Larval (left) and adult (right) stages of the hoverfly Episyrphus balteatus. The 3rd larva preys on pest insects, while the adult feeds on floral nectar and pollen.

Methods

Ethics statement

This study did not involve human or animal subjects, and thus, no ethical approval was required.

Insect rearing and sample collection

The E. balteatus colony was initially collected from a wheat farm in Langfang, Hebei province, China, in 2017 and subsequently reared in the laboratory under controlled conditions of 23 ± 1 °C, 65 ± 5% relative humidity, and 14:10 (L:D) h photoperiod. The grown-ups were provided with pollen and 10% (w/v) honey solution. The larvae were fed with aphids (Aphis craccivora Koch) on pea plantlets (Fig. 1).

For trancriptome sequencing, different developmental stages of E. balteatus, including egg (1 d after oviposition, n = 100), larva (a mixture of 1st to 3rd instar, n = 10) and pupa (2 d after pupation, n = 10), and different tissues of adults, including antenna (n = 100), body (abdomen, n = 30), head (without antennae or proboscis) (n = 30), leg (n = 100), proboscis (n = 100), thorax (without legs, n = 30) and wing (n = 100) were collected and immediately frozen in liquid nitrogen. The whole body of the larvae was used for RNA extraction. Prior to sampling, larvae were subjected to a short starvation period to allow the gut contents to be largely cleared. The adults consist of female and male individuals, which were sampled within 12 h after eclosion. For transcriptome sequencing, tissues from both sexes were pooled when constructing the libraries. Three replicates were prepared for each sample. All samples were stored at –80 °C for subsequent RNA extraction.

Total RNA extraction

Total RNA was extracted from each sample using the TransZol kit (Transgen, Beijing) in accordance with the provided instructions. RNA concentration and purity were assessed using Nanodrop 2000, and the integrity of RNA was evaluated by 1% agarose gel electrophoresis. RNA quality number (RQN) was measured using the Agilent 5300. A total of 1 μg of high-quality RNA (OD260/280: 1.8~2.2; RQN > 6.5) was used for cDNA library construction by Shanghai Majorbio Biotech Corporation (Shanghai, China) (Fig. 2). Each developmental stage or tissue type was used as an independent sample to construct a cDNA library for transcriptome sequencing.

Fig. 2.

Experimental design and workflow.

Transcriptome sequencing

mRNA was isolated from total RNA using oligo(dT) magnetic beads to pair with its polyA tail. Subsequently, mRNA was randomly fragmented into small segments (~300 bp) by adding the fragmentation buffer. Double-stranded cDNA was synthesized utilizing random primers and reverse transcriptase with mRNA as template. Adapters were added to the ends of cDNA, and the resulting products were subjected to PCR amplification. Following purification, the final libraries were quantified using Qubit 4.0. The samples were mixed according to the required ratio, and bridge PCR amplification was conducted on the cBot Cluster Generation System to generate clusters. The final libraries were sequenced on a NovaSeq X Plus platform by Shanghai Majorbio Biotech Corporation (Shanghai, China) (Fig. 2).

Data analysis

The quality assessment of raw data was conducted. Following the removal of adapters and low-quality sequences, clean data were obtained for further analysis. High-quality sequences were assembled into contigs and singletons using Trinity v2.8.5¹⁸. Subsequently, the initial assembly was optimized and filtered by TransRate v1.0.3¹⁹. The transcriptome assembly integrity was assessed by BUSCO v3.0.2²⁰. The transcripts were blasted against six databases (NR, Swiss-Prot, Pfam, EggCOG, GO, and KEGG databases) to obtain annotation information (Fig. 2). Clean reads from each sample were aligned with the assembled reference sequences to generate the mapping results, which were fed to RSEM v1.3.1²¹ for calculating the expression levels of genes and transcripts. The expression levels were measured by transcripts per million²². The differential expression of genes among samples was assessed using DESeq2 v1.24.0²³ with default parameters (|log₂FC| >  = 1 and padjust <0.05) relying on the read counts of each sample. Significantly enriched KEGG pathways of differentially expressed genes (DEGs) among different developmental stages and different adult tissues were identified with an padjust < 0.05.

Data Records

The raw data for the 30 samples in this investigation have been deposited in the NCBI Sequence Read Archive database with accession number SRP640040²⁴. The accession number of each sample was as follows: egg: SRR35931322-SRR35931324; larva: SRR35946290-SRR35946292; pupa: SRR35977951-SRR35977953; antenna: SRR35931319-SRR35931321; head: SRR35946287-SRR35946289; proboscis: SRR35947499-SRR35947501; thorax: SRR35940144-SRR35940146; body: SRR35940141-SRR35940143; leg: SRR35947496-SRR35947498; wing: SRR35977948-SRR35977950²⁴. This Transcriptome Shotgun Assembly project has also been submitted to DDBJ/EMBL/GenBank under the accession GLKO00000000²⁵. The version described in this paper is the first version, GLKO01000000. The read counts of unigenes, transcriptome assembly and annotation files have been submitted to the Gene Expression Omnibus (GEO) database with the accession number GSE324442²⁶.

Data Overview

A total of 133.87 Gb of clean data was obtained from the transcriptome sequencing of 30 samples. The mapping rates of clean reads to the assembled sequences of the 30 samples ranged from 71.29% to 78.58% (Table 1). The variation in mapping rates differed among different tissue types. For example, the average mapping rate for thorax libraries was approximately 77.71%, whereas the wing libraries showed a slightly lower average mapping rate (~71.79%). The assembly predicted a total of 85,676 unigenes. BUSCO analysis indicated that the assembly contains 97.9% complete and single-copy genes (Table 2).

Table 1.

Summary of sequencing read statistics.

Sample_ID	Total_Reads	Total_Bases	Error%	Q20%	Q30%	aligned_rate (%)
Antenna_1	31,439,620	4,241,957,328	0.01	100	100	73.49
Antenna_2	28,191,634	3,812,046,656	0.01	100	100	73.33
Antenna_3	31,040,694	4,190,827,753	0.01	100	100	73.09
Body1	26,041,054	3,590,768,109	0.01	100	100	75.91
Body2	52,373,914	7,111,136,203	0.01	100	100	77.23
Body3	52,115,652	7,098,021,619	0.01	100	100	76.71
Egg1	37,272,214	4,917,690,118	0.01	100	100	76.91
Egg2	37,522,996	4,991,840,654	0.01	100	100	76.67
Egg3	34,930,658	4,676,528,459	0.01	100	100	75.13
Head1	29,259,056	3,955,901,063	0.01	100	100	72.07
Head2	34,155,548	4,606,872,143	0.01	100	100	72.16
Head3	30,540,536	4,135,401,124	0.01	100	100	72.57
Larva1	31,268,786	4,236,646,137	0.01	100	100	75.21
Larva2	35,453,358	4,787,443,331	0.01	100	100	75.16
Larva3	33,647,890	4,540,808,148	0.01	100	100	75.05
Leg1	29,371,090	3,949,100,140	0.01	100	100	75.14
Leg2	30,840,014	4,149,722,482	0.01	100	100	75.26
Leg3	36,208,488	4,861,466,853	0.01	100	100	75.26
Proboscis1	28,787,470	3,897,770,704	0.01	100	100	76.26
Proboscis2	29,278,760	3,955,348,058	0.01	100	100	75.98
Proboscis3	31,604,708	4,263,464,818	0.01	100	100	75.88
Pupa1	36,843,702	4,964,876,880	0.01	100	100	73.99
Pupa2	32,698,678	4,424,855,532	0.01	100	100	73.72
Pupa3	28,140,476	3,821,635,532	0.01	100	100	74.08
Thorax1	27,484,790	3,692,096,279	0.01	100	100	77.24
Thorax2	32,682,098	4,394,848,790	0.01	100	100	77.30
Thorax3	29,505,610	3,952,216,548	0.01	100	100	78.58
Wing1	28,823,932	3,898,000,457	0.01	100	100	71.29
Wing2	34,079,346	4,601,416,469	0.01	100	100	71.71
Wing3	30,715,188	4,145,837,073	0.01	100	100	72.36

Table 2.

Statistics of Episyrphus balteatus transcriptome assembly.

Type	Unigene	Transcript
Total number	85,676	143,098
Total base	61,720,081	115,248,176
Largest length (bp)	24,494	24,494
Smallest length (bp)	201	201
Average length (bp)	720.39	805.38
N50 length (bp)	1,028	1,183
E90N50 length (bp)	2,980	2,440
Mean mapped percent (%)	58.497	69.689
GC percent (%)	35.92	35.89
TransRate score	0.22239	0.27134
BUSCO score	C: 97.9% [S: 81.3%, D: 16.6%], F: 1.2%, M: 0.9%, n: 978	C: 98.2% [S: 30.0%, D: 68.2%], F: 1.0%, M: 0.8%, n: 978

Technical Validation

Quality validation

The quality of clean reads was evaluated using fastp (https://github.com/OpenGene/fastp). The mean error rates of sequencing bases were below 0.01%. Q20 and Q30 represent the percentages of bases exhibiting sequencing quality above 99% and 99.9%, respectively, relative to the total bases. In this investigation, Q20 and Q30 values were 100% (Table 1).

Transcriptome assembly and functional annotation

The average length of unigenes was 720.39 bp, and the length of N50 was 1,028 bp (Table 2). No unigenes were less than 200 bp, and 14,108 unigenes exceeded 1,000 bp (Fig. 3a). A total of 45,479 unigenes (53.08%) were annotated in at least one database (Fig. 3b).

Fig. 3.

(a) Gene length distribution and (b) functional annotation.

Differential expression

The log₁₀(TPM + 1) values illustrated the gene expression distributions throughout the 30 samples, which ranged from 0.30 to 1.13 (Fig. 4a). The principal component analysis (PCA) demonstrated that three replicates of each sample clustered together (Fig. 4b). The correlation analysis among tissues indicated that the correlation coefficients are relatively high, approaching 1 (Fig. 4c).

Fig. 4.

Gene expression distribution (a), PCA (b), and gene expression profiles among tissues (c).

In comparison to the larval stage, the number of down-regulated genes was greater than that of up-regulated genes during the egg and pupal stages of E. balteatus (Fig. 5a,c). Among the top 20 significantly regulated genes, many are cuticular proteins (Fig. 5a,c). These DEGs were enriched in protein digestion and absorption, galactose metabolism, glutathione metabolism, starch and sucrose metabolism, cholesterol metabolism, drug metabolism, and steroid hormone biosynthesis (Fig. 5b,d).

Fig. 5.

DEGs among different development stages of E. balteatus. Volcanic maps of DEGs in egg (a) and pupa (c) compared with larva of E. balteatus and the top 20 significantly regulated genes arranged by padjust were labeled. The top 20 significantly enriched KEGG pathways of DEGs in egg (b) and pupa (d) in comparison with larva of E. balteatus.

The number of DEGs among different tissues of adults varied (Fig. 6a). A total of 31,565, 21,312 and 20,742 DEGs were significantly expressed in tissues, compared to the antenna, leg and proboscis respectively (Fig. 6b). A total of 17,222 DEGs were significantly expressed across all three tissues (Fig. 6b). Additionally, KEGG enrichment analysis revealed that these 17,222 DEGs were enriched in drug metabolism, insect hormone biosynthesis, cardiac muscle contraction, circadian entrainment, phototransduction, taste transduction, salivary secretion, and olfactory transduction (Fig. 6c).

Fig. 6.

DEGs among different tissues of E. balteatus adults. The number of DEGs among different tissues of adults (a,b). The top 50 significantly enriched KEGG pathways of DEGs in antenna, leg, and proboscis (c).

Author contributions

H.Y., H.R.J. and K.M.W conceived the research project. H.Y., H.R.J. and X.Y.Z. participated in the data analysis. H.R.J. and H.L. collected the samples. H.Y., H.R.J. and X.Y.Z. wrote the manuscript. H.Y., H.R.J., X.Y.Z. and K.M.W. revised the manuscript. All authors have read and approved the final manuscript for submission.

Data availability

The raw data for the 30 samples of E. balteatus in this study are available at the SRA database, with accession number SRP640040 (https://identifiers.org/ncbi/insdc.sra:SRP640040)²⁴. The assembled E. balteatus transcriptome assembly has been submitted to DDBJ/EMBL/GenBank under the accession GLKO00000000 (http://identifiers.org/ncbi/insdc:GLKO01000000)²⁵. The read counts of unigenes, transcriptome assembly and annotation files have been submitted to the GEO database with accession number GSE324442 (https://identifiers.org/geo/GSE324442)²⁶.

Code availability

No custom code was used in this study. Fastp (Version 0.19.5): https://github.com/OpenGene/fastp. Trinity (Version v2.8.5): https://github.com/trinityrnaseq/trinityrnaseq. Transrate (Version v1.0.3): http://hibberdlab.com/transrate/index.html. BUSCO (Version 3.0.2): https://busco.ezlab.org/. RSEM (Version 1.3.1): http://deweylab.biostat.wisc.edu/rsem/. DESeq2 (Version 1.24.0): http://bioconductor.org/packages/stats/bioc/DESeq2/.

Competing interests

The authors declare no competing interests.