Map and model-moving from observation to prediction in toxicogenomics

Area of Science:

• Toxicogenomics
• Environmental Health
• Computational Biology

Background:

• Chemicals induce unique transcriptomic changes (toxicogenomic fingerprints) offering insights into cellular responses and adverse effects.
• Assessing chemical hazards and environmental health requires comparing toxicogenomic experiments, but comparability is limited by a lack of response dynamics.
• Existing methods struggle to capture the dynamic nature of toxicogenomic responses, hindering cross-experiment interpretation.

Purpose of the Study:

• To develop an experimental design and bioinformatic strategy for inferring time- and concentration-resolved toxicogenomic fingerprints.
• To create a universal coordinate system for comparing toxicogenomic data across experiments.
• To enable quantitative description and extrapolation of gene expression responses to chemical exposure.

Main Methods:

• Developed a self-organizing map to project toxicogenomic fingerprints into a universal coordinate system.
• Created a time- and concentration-dependent regression model for quantitative analysis of gene expression.
• Applied the strategy to a zebrafish embryo microarray study with model compounds, including cyclooxygenase inhibitors.

Main Results:

• Successfully inferred time- and concentration-resolved toxicogenomic fingerprints.
• Identified gene clusters with functional relationships based on co-expression patterns.
• Characterized associations between transcriptomic responses and developmental, toxicokinetic, and toxicodynamic processes using model parameters.
• Discussed the link between toxicogenomic effects and measured internal chemical concentrations.

Conclusions:

• The developed pipeline provides a blueprint for generating comparable toxicogenomic fingerprints.
• The approach integrates, aggregates, and models time- and concentration-resolved toxicogenomic data.
• This facilitates a more robust assessment of chemical-related hazards and environmental health impacts.