Emerging pattern mining to aid toxicological knowledge discovery.

Richard Sherhod,Valerie J Gillet,Samuel J Webb,Philip N Judson,Thierry Hanser,Jonathan D Vessey

doi:10.1021/ci5001828

Richard Sherhod, Valerie J Gillet + Show 4 more

Open Access

https://doi.org/10.1021/ci5001828

Copy DOI

Abstract

Knowledge-based systems for toxicity prediction are typically based on rules, known as structural alerts, that describe relationships between structural features and different toxic effects. The identification of structural features associated with toxicological activity can be a time-consuming process and often requires significant input from domain experts. Here, we describe an emerging pattern mining method for the automated identification of activating structural features in toxicity data sets that is designed to help expedite the process of alert development. We apply the contrast pattern tree mining algorithm to generate a set of emerging patterns of structural fragment descriptors. Using the emerging patterns it is possible to form hierarchical clusters of compounds that are defined by the presence of common structural features and represent distinct chemical classes. The method has been tested on a large public in vitro mutagenicity data set and a public hERG channel inhibition data set and is shown to be effective at identifying common toxic features and recognizable classes of toxicants. We also describe how knowledge developers can use emerging patterns to improve the specificity and sensitivity of an existing expert system.

Highlights

The development of accurate methods for the prediction of toxic hazard and environmental effects of chemical compounds is a topic of great importance to all chemical industries[1,2] both from the economic stand point of reducing the need for expensive in vivo or in vitro experiments and from the ethical stand point of reducing testing in animals
We have developed an emerging pattern mining method that can assist knowledge base developers in compiling structural alerts to improve the performance of expert systems for toxicity prediction
Emerging patterns have the significant advantage over the jumping emerging patterns described in our earlier work of being tolerant to noise in the data

Summary

Introduction

The development of accurate methods for the prediction of toxic hazard and environmental effects of chemical compounds is a topic of great importance to all chemical industries[1,2] both from the economic stand point of reducing the need for expensive in vivo or in vitro experiments and from the ethical stand point of reducing testing in animals. Expert systems are a class of computational tools that have shown success in predicting toxic hazard by applying established knowledge of toxicology. The Derek Nexus,[3] HazardExpert,[4] and CASEUltra[5] systems encode structural features that have been associated with particular toxicological effects, known as structural alerts, alongside other parameters such as physicochemical properties. These systems apply different rule-based or reasoning-based decision making algorithms to the stored knowledge in order to make predictions of toxicity. Derek Nexus uses a reasoning model to construct and weigh up arguments for and against toxicity. A disadvantage of expert systems is that developing new structural alerts to expand the knowledge bases requires considerable time and effort from domain experts and involves detailed analysis of relevant literature

Objectives

Methods

Results

Conclusion