Developing a Practical Toxicogenomics Data Analysis System Utilizing Open-Source Software

Abstract

Comprehensive gene expression analysis has been applied to investigate the molecular mechanism of toxicity, which is generally known as toxicogenomics (TGx). When analyzing large-scale gene expression data obtained by microarray analysis, typical multivariate data analysis methods performed with commercial software such as hierarchical clustering or principal component analysis usually do not provide conclusive outputs by themselves. To best utilize the TGx data for toxicity evaluation in the drug development process, fit-for-purpose customization of the analytical algorithm with user-friendly interface and intuitive outputs are required to practically address the toxicologists’ demands. However, commercial software is usually not very flexible in the customization of their functions or outputs. Owing to the recent advancement and accumulation of open-source software contributed by bioinformaticians all over the world, it becomes easier for us to develop practical and fit-for-purpose analytical software by ourselves with fairly low cost and efforts. The aim of this article is to present an example of developing an automated TGx data processing system (ATP system), which implements gene set-level analysis toxicogenomic profiling by D-score method and generates straightforward output that makes it easy to interpret the biological and toxicological significance of the TGx data. Our example will provide basic clues for readers to develop and customize their own TGx data analysis system which complements the function of existing commercial software.