Physics-based machine learning for modeling stochastic IP3-dependent calcium dynamics

TL;DR

This paper introduces a physics-informed machine learning approach for reducing complex stochastic models of IP3-dependent calcium dynamics, improving efficiency and accuracy by integrating domain-specific physics into the learning process.

Contribution

It develops a novel model reduction method that incorporates physics-based functions, enabling effective approximation of stochastic calcium oscillation models with significantly fewer variables.

Findings

Enhanced model generalization and accuracy

Significant reduction in network size for calcium dynamics models

Effective integration of physics into machine learning for biological systems

Abstract

We present a machine learning method for model reduction which incorporates domain-specific physics through candidate functions. Our method estimates an effective probability distribution and differential equation model from stochastic simulations of a reaction network. The close connection between reduced and fine scale descriptions allows approximations derived from the master equation to be introduced into the learning problem. This representation is shown to improve generalization and allows a large reduction in network size for a classic model of inositol trisphosphate (IP3) dependent calcium oscillations in non-excitable cells.