Gene co-expression network analysis of Trypanosoma brucei in tsetse fly vector

Abstract

Background: Trypanosoma brucei species are motile protozoan parasites that are cyclically transmitted by tsetse fly (genus Glossina) causing human sleeping sickness and nagana in livestock in sub-Saharan Africa. African trypanosomes display digenetic life cycle stages in the tsetse fly vector and in their mammalian host. Experimental work on insect-stage trypanosomes is challenging because of the difficulty in setting up successful in vitro cultures. Therefore, there is limited knowledge on the trypanosome biology during its development in the tsetse fly. Consequently, this limits the development of new strategies for blocking parasite transmission in the tsetse fly.

Methods: In this study, RNA-Seq data of insect-stage trypanosomes were used to construct a T. brucei gene co-expression network using the weighted gene co-expression analysis (WGCNA) method. The study identified significant enriched modules for genes that play key roles during the parasite's development in tsetse fly. Furthermore, potential 3' untranslated region (UTR) regulatory elements for genes that clustered in the same module were identified using the Finding Informative Regulatory Elements (FIRE) tool.

Results: A fraction of gene modules (12 out of 27 modules) in the constructed network were found to be enriched in functional roles associated with the cell division, protein biosynthesis, mitochondrion, and cell surface. Additionally, 12 hub genes encoding proteins such as RNA-binding protein 6 (RBP6), arginine kinase 1 (AK1), brucei alanine-rich protein (BARP), among others, were identified for the 12 significantly enriched gene modules. In addition, the potential regulatory elements located in the 3' untranslated regions of genes within the same module were predicted.

Conclusions: The constructed gene co-expression network provides a useful resource for network-based data mining to identify candidate genes for functional studies. This will enhance understanding of the molecular mechanisms that underlie important biological processes during parasite's development in tsetse fly. Ultimately, these findings will be key in the identification of potential molecular targets for disease control.

Keywords: Gene co-expression network; Trypanosoma brucei; Tsetse fly; Weighted gene co-expression network analysis.

Fig. 1

Global gene expression profiles of Trypanosoma brucei. a Principal component analysis (PCA) plot. Each point in the PCA plot represents a sample, and point color indicates a batch that consists of the biological replicates. b Sample correlation heatmap using hierarchical clustering. Color codes along the left side of the sample correlation heatmap indicate samples based on the batch they belong to. MG1 and MG2 are midgut samples, PV2 proventriculus samples, and SA2 salivary gland samples

Fig. 2

An illustration of the identified gene co-expression network modules in T. brucei. a Hierarchical cluster dendrogram. The x-axis represents the co-expression distance of the genes, while the y-axis represents the genes. A dynamic tree cutting algorithm identified the modules by splitting the tree at significant branching points. Modules are represented by different colors as shown by the dendrogram. b Co-expression network from weighted gene co-expression network analysis (WGCNA) based on topological overlap measures (TOMs) > 0.3 for visualization. Each point (or node) on the network represents a gene, and points of the same color form a gene module. Lines (edges) on the network connecting the nodes represent a relationship between the genes

Fig. 3

Number of genes identified in each module. In total, there were 28 modules. The gray module contains 59 genes that could not be assigned to any module and was excluded from downstream analysis

Fig. 4

Prediction of regulatory elements in the 3′ untranslated regions (UTR) based on gene co-expression modules. a Predicted motifs for the gene modules are shown. Columns represent gene modules, while rows represent the predicted motifs with consensus sequence on the right side. Over-representation of a motif for a given gene module is indicated by yellow color with significant over-representation highlighted by red frames. Blue color map and frames indicate under-representation. b Motif pairs co-occurring in the 3′ UTR are shown in the heatmap where each row and each column correspond to a predicted motif. Light colors indicate the presence of another motif within the same 3′ UTR while dark colors indicate that the motifs are absent in the same 3′ UTR. “+” indicates significant spatial co-localization between pairs of motifs