Learning robust cell signalling models from high throughput proteomic data.

Area of Science:

• Systems Biology
• Computational Biology
• Immunology

Background:

• Cell signaling networks are crucial for understanding biological processes.
• High-throughput proteomic data offers insights into these complex networks.
• Existing methods for learning signaling networks may lack robustness.

Purpose of the Study:

• To propose a robust framework for learning Bayesian network models of cell signaling.
• To improve the accuracy and reliability of network structure learning from proteomic data.
• To identify novel molecular interactions within signaling pathways.

Main Methods:

• Developed a novel framework for Bayesian network model learning.
• Utilized Bayesian bootstrap resampling for model averaging to enhance robustness.
• Created an algorithm for ranking network feature importance using bootstrap data.
• Applied the framework to T-cell signaling data from Sachs et al. (2005).

Main Results:

• Bayesian bootstrap resampling yielded more robust network structures compared to using all data.
• The developed algorithm successfully ranked the importance of network features.
• Identified several novel crosstalk mechanisms in the T-cell signaling network with high confidence.
• Six new crosstalk mechanisms await experimental validation.

Conclusions:

• The proposed framework provides a robust approach for learning cell signaling networks.
• The identified crosstalk mechanisms offer new avenues for T-cell signaling research.
• The findings have implications for understanding immune cell function and disease.