A general approach to non-Markovian compartmental models

Area of Science:

• Mathematical Biology
• Biophysics
• Pharmacokinetics

Background:

• Stochastic compartmental models typically use continuous-time Markov processes, assuming exponential retention times.
• This assumption is often violated in real-world applications, such as calcium clearance from bone, necessitating non-Markovian models.

Purpose of the Study:

• To present a general and tractable procedure for estimating retention time distributions in non-Markovian phenomenological models using data.
• To address the limitations of Markovian models in applications with non-exponential retention times.

Main Methods:

• The proposed procedure utilizes phase-type distributions to model non-exponential retention times.
• Phase-type distributions are capable of describing long-tailed data, common in biological processes like calcium clearance.
• The method handles complex eigenvalues often missed in standard analyses.

Main Results:

• A practical procedure for estimating retention time distributions from data for non-Markovian models was developed.
• The effectiveness of phase-type distributions in capturing complex retention behaviors was demonstrated.
• The procedure was successfully illustrated using various models applied to a specific dataset.

Conclusions:

• The developed procedure offers a robust method for analyzing systems with non-Markovian dynamics.
• Phase-type distributions provide a flexible and powerful tool for modeling biological clearance processes.
• This work enhances the accuracy of stochastic compartmental modeling for applications with complex retention times.