A multilayer network approach for guiding drug repositioning in neglected diseases

ARIEL BERENSTEIN; MARIA PAULA MAGARIÑOS; ARIEL CHERNOMORETZ; FERNAN AGUERO

Belo Horizonte

Congreso; ISCB-LA and Xmeeting in Bioinformatics; 2014

Int Society of Computational Biology

Historically, lack of interest from the pharmaceutical industry, resulted in the lack of drugsfor the majority of the pathogens that cause neglected diseases. With the advent of openchemical resources and the release of data from high throughput screenings theavailability of chemical information has increased greatly. However most of the informationon targets and the activity of drugs is for humans or model organisms. This creates a hugeopportunity for computational exercises that use comparative genomics andchemogenomic approaches. In this work, using chemical datasets containing bioactivitydata for pathogen and non-pathogen targets we built a three-layer network containingthree disjoint sets of vertexes with 1) 1.5 million drugs; 2) 167,000 proteins across 221species; and 3) three affiliation-type protein features (orthology, Pfam domains,participation in metabolic pathways). Only statistically significant features (in the context ofdrug-target predictions) were taken into account. A bipartite projection of the protein layerwas obtained capturing the protein affiliations to different annotation classes. Using thisnetwork model, we tackled the problems of i) prioritizing targets for drug discovery forpathogens; and ii) suggesting candidate targets for orphan compounds (those that areactive in whole-cell screenings but whose target is currently unknown). We used a tenfoldcross validation procedure to assess the performance of the network to prioritize relevanttargets and show that he method overcomes traditional similarity-based searches againstdruggable targets. In the presentation we will discuss some of the top ranked proteins. Wealso obtained candidate target proteins for orphan molecules that were active in whole-cellassays against P. falciparum and will discuss the proposed targets that mediate theantiplasmodial activity for some of these orphan compounds.