Data Sources for Signature Discovery in Toxicology

Abstract

From a systems biology point of view, signature of a chemical can be defined as a collection of data or a measure of cellular response to a certain chemical, where biomarkers are often the best characteristics objectively measured and evaluated as an indicator of the biological processes. Chemical profiles (signatures) related to chemical-induced gene expression and gene changes coupled to particular pathologies require additional interpretation and processing to identify true biomarker candidates and together with the mechanistic hypothesis underlying biological effects of that chemical enable comprehensive knowledge about chemical toxic effect. Knowledge bases such as ToxWiz capture a broad spectrum of mechanistic hypothesis and pathways for toxic effects derived from precise expert analysis of millions of scientific articles, establishing connections between mechanism of disease, pathology, and toxic endpoints, and representing them in a form of biological pathways for underlying toxic endpoints. A unique module connected to ToxWiz knowledge base represents a hand-curated database of gene expression signatures. This module contains a good size collection of 1000 unique toxicity signatures related to chemical-induced toxicities in human, rat, mouse, and several nonmammalian species covering studies with 297 compounds, known to induce major toxicities in liver, kidney, and most other major organ systems. Implemented software tools based on systems biology principles allow analysis of novel compounds for toxic effects and allows the analysis of your own data in order to identify new biomarker candidates by interrogating your -omics data with gene signatures module. By describing biological pathways underlying the toxic effects, and discovering and exploring related biomarkers, these tools promise to help design safer chemicals. Furthermore, we describe here an example exercise on annotated public data, using ToxWiz knowledge base and tools, which confirm and expand the science on the hazards of tobacco smoke exposure via this approach and we demonstrate its respectable power to make accurate predictions of possible toxicity and generation of mechanistic hypothesis of its effect, as an indication of the cellular response in vascular endothelium upon exposure to tobacco smoke components.