A Method to Calibrate Chemical-Agnostic Quantitative Adverse Outcome Pathways on Multiple Chemical Data.

Area of Science:

• Toxicology
• Computational Chemistry

Background:

• Quantitative Adverse Outcome Pathways (qAOPs) are crucial for next-generation risk assessment using New Approach Methods (NAMs).
• Existing qAOPs often pool data from multiple chemicals, leading to issues with cross-chemical heterogeneity that obscure pathway relationships and reduce model reliability.

Purpose of the Study:

• To develop and validate a novel calibration approach for qAOPs to address and mitigate cross-chemical heterogeneity.
• To ensure qAOPs are truly chemical-agnostic, enhancing their reliability and generalizability in risk assessment.

Main Methods:

• Developed a hierarchical calibration approach to systematically separate chemical-specific heterogeneity from underlying pathway effects.
• Modeled chemical-specific deviations as random effects to extract core mechanistic, chemical-independent pathway parameters.
• Utilized a simulation study to assess model performance differences with and without hierarchical calibration.

Main Results:

• Hierarchical calibration effectively separates chemical-specific heterogeneity from pathway effects.
• Simulation results demonstrated that model performance differences highlight the magnitude of data heterogeneity.
• Uncalibrated qAOPs with substantial heterogeneity confound pathway effects with chemical-specific variations, failing to be truly chemical-agnostic.

Conclusions:

• The developed hierarchical calibration approach enhances the chemical-agnostic nature of qAOPs.
• This method improves the reliability and generalizability of qAOPs for risk assessment.
• Applied the approach to a case study on nonmutagenic liver tumor qAOPs for deriving points of departure.