QSAR Modeling of Tox21 Challenge Stress Response and Nuclear Receptor Signaling Toxicity Assays

Regina Politi,Olexandr Isayev,Stephen J Capuzzi,Sherif Farag,Alexander Tropsha

doi:10.3389/fenvs.2016.00003

Regina Politi, Olexandr Isayev + Show 3 more

Open Access

https://doi.org/10.3389/fenvs.2016.00003

Copy DOI

Abstract

The ability to determine which environmental chemicals pose the greatest potential threats to human health remains one of the major concerns in regulatory toxicology. Computation methods that can accurately predict the chemicals’ toxic potential in silico are increasingly sought-after to replace in vitro high-throughput screening (HTS) as well as controversial and costly in vivo animal studies. To this end, we have built Quantitative Structure-Activity Relationship (QSAR) models of twelve (12) stress response and nuclear receptor signaling pathways toxicity assays as part of the 2014 Tox21 Challenge. Our models were built using the Random Forest, Deep Neural Networks and various combinations of descriptors and balancing protocols. All of our models were statistically significant for each of the 12 assays with the balanced accuracy in the range between 0.58 and 0.82. Our results also show that models built with Deep Neural Networks had high accuracy than those developed with simple machine learning algorithms and that dataset balancing led to a significant accuracy decrease.

Highlights

The ability to determine which environmental chemicals pose the greatest potential threats to human health remains one of the major concerns in regulatory toxicology
We investigated the use of different Quantitative Structure-Activity Relationship (QSAR) approaches for toxicity assays prediction in the 2014 Tox21 challenge
The model performance was evaluated by the area under the receiver operating characteristic curve (AUCROC) and by the balanced accuracy (BA)

Summary

Introduction

The ability to determine which environmental chemicals pose the greatest potential threats to human health remains one of the major concerns in regulatory toxicology. The inability to recognize potentially toxic substances during the initial steps of drug development contributes to the failure of promising pharmaceutical leads in more than 30% of human clinical trials (Kola and Landis, 2004). The estimated human health impact of these chemicals has been assessed through in vivo animal studies. Animal studies are costly, laborious, impractical for evaluating large numbers of chemicals, and are being progressively eliminated due to their controversial nature (Anastas et al, 2010). Over the past several years, the focus has switched to high-throughput in vitro screening (HTS) in order to identify chemical hazards and prioritize chemicals for additional in vivo testing (O’Brien et al, 2006)

Methods

Discussion

Conclusion