Inferring phenomenological models of first passage processes

Area of Science:

• Computational Biology
• Biophysics
• Systems Biology

Background:

• Cellular biochemical networks are complex and stochastic.
• Microscopically accurate models are often infeasible due to complexity and diverse timescales.
• Phenomenological models are needed to approximate system behavior from data.

Purpose of the Study:

• To develop a systematic framework for building phenomenological models from experimental data.
• To accurately approximate first passage (FP) times in cellular processes.
• To enable prediction of FP system behavior under varying conditions.

Main Methods:

• Utilized mixtures of Gamma distributions for phenomenological models.
• Developed an adaptive complexity approach based on data availability and temporal resolution.
• Applied the framework to model inter-spike intervals of Purkinje cells using experimental and simulated data.

Main Results:

• The proposed models accurately fit experimental and simulated data for Purkinje cell inter-spike intervals.
• The models demonstrated the ability to make nontrivial predictions.
• Coarse-grained models provided valuable constraints for more mechanistically detailed models.

Conclusions:

• The framework offers a robust method for creating data-driven phenomenological models of complex biological processes.
• Gamma distribution mixtures provide a biophysically interpretable approach to modeling first passage times.
• This approach bridges the gap between data and mechanistic understanding in systems biology.