BMDExpress: a software tool for the benchmark dose analyses of genomic data

Area of Science:

• Toxicogenomics
• Computational Biology

Background:

• Dose-dependent biological processes are common, involving phenotypic changes from molecular exposures.
• Microarray technology is increasingly used to study these processes, but quantitative analysis tools are limited.

Purpose of the Study:

• To develop a software tool for quantitative analysis and interpretation of dose-response microarray data.
• To integrate benchmark dose (BMD) methods with gene ontology (GO) classification.

Main Methods:

• Developed BMDExpress, a Java application for analyzing dose-response microarray data.
• Employs a stepwise analysis: ANOVA for significant dose-response genes, followed by statistical model fitting (linear, polynomial, power) to estimate BMD.
• Summarizes results by matching genes to GO categories and calculating dose-dependent behavior for biological processes.

Main Results:

• BMDExpress identifies genes with significant dose-response behavior.
• Estimates benchmark doses (BMDs) where gene expression deviates from controls.
• Characterizes dose-dependent transcriptional changes for biological processes and molecular functions.

Conclusions:

• BMDExpress combined with microarray technology effectively characterizes dose-dependent transcriptional changes.
• The software enables efficient analysis of large dose-response microarray studies.
• Identifies reference doses altering specific cellular processes; freely available.