Biomarkers of nanomaterials hazard from multi-layer data

Area of Science:

• Nanomaterial safety and toxicology
• Computational toxicology
• Predictive modeling in environmental health

Background:

• Engineered nanomaterials (ENMs) present unique toxicological challenges.
• Reliable prediction of ENM toxicity is crucial for risk assessment.
• Current predictive methods often lack comprehensive data integration.

Purpose of the Study:

• To develop and validate a predictive computational framework for ENM toxicity.
• To identify reliable molecular and phenotypic markers of ENM toxicity.
• To improve the accuracy of ENM toxicity predictions using multi-omics data.

Main Methods:

• Utilized a large panel of ENMs across in vitro and in vivo models.
• Employed multi-omics approaches (genomics, transcriptomics) for ENM grouping.
• Integrated omics-derived features with material intrinsic properties for predictive modeling.
• Validated toxicity markers across multiple independent datasets.

Main Results:

• Successfully grouped ENMs based on multi-omics profiles and toxicity data.
• Identified and replicated key mRNA-based toxicity markers.
• Achieved significantly improved predictive accuracy using omics features combined with material properties.
• Demonstrated superior performance compared to models based solely on physicochemical properties.

Conclusions:

• The developed computational framework offers a robust approach for predicting ENM toxicity.
• Multi-omics data combined with material properties enhance predictive accuracy.
• Identified mRNA markers serve as reliable indicators for ENM toxicity assessment.