Computer vision reveals hidden variables underlying NF-κB activation in single cells.

Area of Science:

• Cellular and Molecular Biology
• Immunology
• Computational Biology

Background:

• Cellular responses to external signals exhibit significant heterogeneity.
• Studying the molecular basis of this heterogeneity is challenging due to experimental limitations.
• Understanding single-cell immune pathway activation is crucial for therapeutic development.

Purpose of the Study:

• To develop a method for predicting single-cell responses to immune stimuli.
• To identify the molecular determinants of cellular heterogeneity in immune activation.
• To investigate the role of nuclear factor κB (NF-κB) in predetermining cell activation.

Main Methods:

• Development of an image-based support vector machine learning model.
• Application of computer vision for cell analysis.
• Mechanistic modeling to understand regulatory dynamics.

Main Results:

• Computer vision accurately predicts which cells will respond to cytokine stimulation.
• Pre-existing "leaky" nuclear localization of NF-κB (p65:p50) determines cell activation.
• The ratio of NF-κB to its inhibitor dictates leakiness and activation probability.
• Cells maintain stable activation probabilities despite dynamic molecular state transitions.

Conclusions:

• Image-based machine learning can uncover mechanisms of single-cell heterogeneity.
• NF-κB dynamics and its inhibitor ratio are key regulators of immune cell activation probability.
• This approach provides a powerful tool for studying cellular responses to proinflammatory stimuli.