Comparative transcriptomics of early dipteran development

Abstract

Background: Modern sequencing technologies have massively increased the amount of data available for comparative genomics. Whole-transcriptome shotgun sequencing (RNA-seq) provides a powerful basis for comparative studies. In particular, this approach holds great promise for emerging model species in fields such as evolutionary developmental biology (evo-devo).

Results: We have sequenced early embryonic transcriptomes of two non-drosophilid dipteran species: the moth midge Clogmia albipunctata, and the scuttle fly Megaselia abdita. Our analysis includes a third, published, transcriptome for the hoverfly Episyrphus balteatus. These emerging models for comparative developmental studies close an important phylogenetic gap between Drosophila melanogaster and other insect model systems. In this paper, we provide a comparative analysis of early embryonic transcriptomes across species, and use our data for a phylogenomic re-evaluation of dipteran phylogenetic relationships.

Conclusions: We show how comparative transcriptomics can be used to create useful resources for evo-devo, and to investigate phylogenetic relationships. Our results demonstrate that de novo assembly of short (Illumina) reads yields high-quality, high-coverage transcriptomic data sets. We use these data to investigate deep dipteran phylogenetic relationships. Our results, based on a concatenation of 160 orthologous genes, provide support for the traditional view of Clogmia being the sister group of Brachycera (Megaselia, Episyrphus, Drosophila), rather than that of Culicomorpha (which includes mosquitoes and blackflies).

Figure 1

Phylogeny and development of Diptera. (A) Simplified phylogenetic tree displaying the relationships among the species used in this study, with respect to drosophilids and mosquitoes. The position of Clogmia albipunctata is controversial (see our Results). (B) Schematic representation of the dipteran life cycle, expanded view showing stages of embryo development. Transcriptomes were obtained from embryos at the following developmental stages: cleavage, blastoderm, gastrulation, and early germband extension. Image sources: Culiseta longiareolata and Episyrphus balteatus pictures by Joaquim Alves Gaspar; Clogmia albipunctata by Sanjay Acharya; Drosophila melanogaster picture by André Karwath (images publicly available through Wikimedia commons); Megaselia abdita picture taken by Karl R. Wotton; embryo, larvae, and fly drawings by Victor Jiménez-Guri.

Figure 2

Verification of transcriptome data by in situ hybridization. We tested several selected candidate genes involved in pattern formation for spatial gene expression during the blastoderm and later stages up to the extended germband. Examples of such patterns in both M. abdita and C. albipunctata are shown. Embryos are aligned anterior to the left, dorsal up. See text for details.

Figure 3

Comparative analysis of genes detected in different species. (A) Venn-diagram of annotated genes from all three species (C. albipunctata, M. abdita, and E. balteatus) compared to genes detected in the early embryonic transcriptome of D. melanogaster (developmental stages: 0–4 hrs, from [50]; see also Table  1). (B) Pie charts showing the number of genes per species which are conserved in all four, or only a subset of species shown in (A). The right-most pie chart shows numbers of conserved genes averaged across all three species.

Figure 4

Dipteran Phylogeny. This figure shows two alternative topologies obtained from the phylogenetic analysis of 21 species, differing only on the position of C. albipunctata. Numbers above branches (in black) indicate the percentage of individual gene trees from the four reconstructed phylomes supporting each bipartition. (A) Hypothesis supported by most of the phylogenetic methods tested, including maximum likelihood analysis of 160 concatenated genes and one supertree approach (DupTree-tree). Branch lengths and bootstrap values (in red) correspond to the RAxML-tree (see Methods). Bootstrap values were calculated for all branch points. For clarity, we only show the one for the branch leading to C. albipunctata. (B) Alternative topology supported by one supertree approach (Clann-SFIT-tree). Branch information and bootstrap supports are not available with this methodology (C) Summary of all phylogenetic results: four methods supported topology A, while only one supported topology B. We also show results from CONSEL-based hypothesis testing (yellow background; Approximately Unbiased (AU) Test, see Methods). These results indicate significant p-values for the rejection of topology B both in the case of raw and BMGE-corrected alignment (p = 0.071 and 0.027 respectively).

Figure 5

Rate of duplications per gene in dipteran lineages. Numbers above branches correspond to computed duplication rates (per gene) in the corresponding lineage. Superimposed bubbles are proportional to these numbers. All computations are based on the D. melanogaster phylome, except those specific to C. albipunctata, M. abdita, and E. balteatus, which are based on their corresponding phylomes (see main text for details).