Predicting the dynamics of protein abundance

Area of Science:

• Molecular Biology
• Systems Biology
• Bioinformatics

Background:

• Protein synthesis regulation is crucial across all organisms.
• Protein abundance exhibits greater dynamic range than transcript levels.
• Measuring protein abundance is complex and expensive compared to mRNA sequencing.

Purpose of the Study:

• To develop a Bayesian network integrating transcriptomic and proteomic data for predicting protein abundance.
• To model the determinants of protein abundance and track molecular responses over time.
• To explore protein-level regulation in Saccharomyces cerevisiae and Schizosaccharomyces pombe.

Main Methods:

• Developed a Bayesian network integrating mRNA levels, mRNA-protein interactions, mRNA folding energy, half-life, and tRNA adaptation.
• Applied the model to Saccharomyces cerevisiae and Schizosaccharomyces pombe data.
• Validated predictions against experimental data from cell cycle studies and a human cell line.

Main Results:

• The predictor achieved robust, albeit minor, accuracy by combining key features to handle data uncertainty.
• Predicted protein abundance identified twice as many cell-cycle-associated proteins compared to experimental mRNA levels.
• Predicted protein abundance demonstrated greater dynamism than observed mRNA expression, consistent with human cell line data.

Conclusions:

• The Bayesian network effectively predicts protein abundance and improves the analysis of protein regulation.
• The model offers a valuable tool for understanding cellular adaptation and molecular responses.
• The approach supports the emerging view that mRNA folding influences translation efficiency.