Interactive online application for the prediction, ranking and prioritisation of drug targets in Schistosoma haematobium

Abstract

Background: Human schistosomiasis is a neglected tropical disease caused by parasitic worms of the genus Schistosoma that still affects some 200 million people. The mainstay of schistosomiasis control is a single drug, praziquantel. The reliance on this drug carries a risk of resistance emerging to this anthelmintic, such that research towards alternative anti-schistosomal drugs is warranted. In this context, a number of studies have employed computational approaches to prioritise proteins for investigation as drug targets, based on extensive genomic, transcriptomic and small-molecule data now available.

Methods: Here, we established a customisable, online application for the prioritisation of drug targets and applied it, for the first time, to the entire inferred proteome of S. haematobium. This application enables selection of weighted and ranked proteins representing potential drug targets, and integrates transcriptional data, orthology and gene essentiality information as well as drug-drug target associations and chemical properties of predicted ligands.

Results: Using this application, we defined 25 potential drug targets in S. haematobium that associated with approved drugs, and 3402 targets that (although they could not be linked to any compounds) are conserved among a range of socioeconomically important flatworm species and might represent targets for new trematocides.

Conclusions: The online application developed here represents an interactive, customisable, expandable and reproducible drug target ranking and prioritisation approach that should be useful for the prediction of drug targets in schistosomes and other species of parasitic worms in the future. We have demonstrated the utility of this online application by predicting potential drug targets in S. haematobium that can now be evaluated using functional genomics tools and/or small molecules, to establish whether they are indeed essential for parasite survival, and to assist in the discovery of novel anti-schistosomal compounds.

Keywords: Computational drug discovery; Drug targets; Prioritisation systems; Schistosoma.

Fig. 1

Distribution of all feature scores (unweighted) for Schistosoma haematobium genes/proteins (Table 1). Feature description ending with a “?” have an unweighted score of either “0” (“no”) or “1” (“yes”), whereas all other features are represented by percentage values of similarity or coverage, respectively. Proteins were clustered using the Ward clustering method based on the Euclidian distances between feature profiles of individual genes/proteins

Fig. 2

User interface of the online application. The weighting of features can be set via a slider (a) or features can be excluded, ignored or required (b). Additionally, feature restrictions/filters for associated drugs can be defined using a range slider (c) or check-boxes. Of the five panels that represent different steps in the ranking/prioritisation process (d) and the two panels that visually summarise and display the resulting proteins and drugs (e), the “Drugs” panel is shown here as an example

Fig. 3

Score distributions for inferred Schistosoma haematobium drug targets. The distributions of scores for targets with associated drugs (n = 25; a) and those without associated drugs (n = 3402; b) are shown