A Kolmogorov metric embedding for live cell microscopy signaling patterns

Area of Science:

• Cellular and Molecular Biology
• Biophysics
• Computational Biology

Background:

• Live cell microscopy generates complex spatiotemporal data.
• Analyzing cell signaling dynamics requires robust quantitative methods.
• Existing methods may require prior knowledge or training data.

Purpose of the Study:

• To present a novel metric embedding for analyzing 5D live cell microscopy movies.
• To capture spatiotemporal patterns of cell signaling dynamics.
• To provide a theoretically grounded, data-driven analysis approach.

Main Methods:

• Utilized normalized information distance (NID) based on Kolmogorov complexity.
• Developed the cell signaling structure function (SSF) using metric 3D image filters.
• Integrated SSF with lossless compression for metric embedding of 5D movies.

Main Results:

• The NID provides an absolute measure of information content between movies.
• The metric embedding optimally approximates pattern differences.
• Demonstrated utility on synthetic data and various live cell imaging experiments (ERK, AKT signaling).

Conclusions:

• The developed metric embedding effectively captures cell signaling dynamics.
• This method offers a powerful, parameter-light tool for analyzing complex biological data.
• The open-source software facilitates broader application in cell biology research.