Computational Construction of Toxicant Signaling Networks

Authors:

Abstract
Humans and animals are regularly exposed to compoundsthat mayhave adverse effects on health. The Toxicity Forecaster (ToxCast)program was developed to use high throughput screening assays to quicklyscreen chemicals by measuring their effects on many biological endpoints. Many of these assays test for effects on cellular receptorsand transcription factors (TFs), under the assumption that a toxicantmay perturb normal signaling pathways in the cell. We hypothesizedthat we could reconstruct the intermediate proteins in these pathwaysthat may be directly or indirectly affected by the toxicant, potentiallyrevealing important physiological processes not yet tested for manychemicals. We integrate data from ToxCast with a human protein interactometo build toxicant signaling networks that contain physical and signalingprotein interactions that may be affected as a result of toxicantexposure. To build these networks, we developed the EdgeLinker algorithm,which efficiently finds short paths in the interactome that connectthe receptors to TFs for each toxicant. We performed multiple evaluationsand found evidence suggesting that these signaling networks capturebiologically relevant effects of toxicants. To aid in disseminationand interpretation, interactive visualizations of these networks areavailable at http://graphspace.org.
Funding Information
  • Division of Biological Infrastructure (DBI-1759858)
  • Division of Molecular and Cellular Biosciences (MCB-1817736)
  • Biological and Environmental Research (DE-SC0022090)
  • National Institute of Food and Agriculture (2018-07578)
  • Virginia Polytechnic Institute and State University
  • Division of Chemical, Bioengineering, Environmental and Transport Systems (CBET-1510920)

This publication has 57 references indexed in Scilit: