Bottom-up modeling approach for the quantitative estimation of parameters in pathogen-host interactions

Area of Science:

• Computational Biology
• Immunology
• Mathematical Modeling

Background:

• Opportunistic fungal pathogens like Candida albicans can cause life-threatening bloodstream infections and sepsis.
• The early immune response involves antimicrobial peptides and innate immune cells (neutrophils, monocytes).
• Mathematical modeling offers a predictive approach to study complex pathogen-host interactions, but parameter estimation can be computationally intensive.

Purpose of the Study:

• To develop a reliable and computationally efficient strategy for estimating parameters in mathematical models of pathogen-host interactions.
• To quantify the relative impact of different immune response routes against Candida albicans in a human whole-blood assay.
• To enable future spatio-temporal simulations for timely sepsis patient stratification.

Main Methods:

• A bottom-up modeling approach combining a non-spatial state-based model (SBM) with an agent-based model (ABM).
• Utilized simulated annealing (global optimization) for parameter estimation in the SBM.
• Employed least-squares error estimation with adaptive grid search (local optimization) in the ABM, leveraging SBM parameters to reduce computational cost.

Main Results:

• Successfully applied a multi-model approach for reliable parameter estimation in a Candida albicans whole-blood infection model.
• Reduced the dimensionality of the parameter space for the ABM by utilizing parameters from the SBM.
• Enabled prediction of cell migration parameters not directly accessible through experiments.

Conclusions:

• The developed bottom-up modeling strategy offers an efficient method for parameter estimation in complex biological systems.
• This approach facilitates a deeper understanding of immune responses to fungal infections.
• Future applications include spatio-temporal simulations for improved sepsis patient management and treatment stratification.