The design and analysis of parallel experiments to produce structurally identifiable models

Area of Science:

• Pharmacokinetics
• Systems Biology
• Mathematical Modeling

Background:

• Compartmental models are crucial for pharmacokinetic analysis.
• Parameter estimation in these models is often restricted due to limited experimental inputs and observations.
• This limitation impacts the structural identifiability and mechanistic relevance of pharmacokinetic models.

Purpose of the Study:

• To address the identifiability challenges in pharmacokinetic compartmental models.
• To present strategies for enhancing global identifiability through parallel experimental designs.
• To demonstrate the application of these strategies using real-world pharmacokinetic data.

Main Methods:

• Development of strategies for rendering compartmental models globally identifiable.
• Implementation of a parallel experiment methodology.
• Analysis of pharmacokinetic literature examples using the proposed methodology.

Main Results:

• The parallel experiment approach can overcome limitations in parameter estimation.
• Sharing a subset of parameters across parallel experiments improves model identifiability.
• Case studies from pharmacokinetic literature illustrate the effectiveness of the methodology.

Conclusions:

• Parallel experimental designs offer a viable solution to enhance the identifiability of pharmacokinetic compartmental models.
• Strategic sharing of parameters across experiments is key to improving model complexity and mechanistic relevance.
• This methodology expands the applicability of detailed pharmacokinetic modeling in human studies.