Building and analyzing protein interactome networks by cross-species comparisons

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Protein-protein interactions (PPIs) are crucial for understanding cellular functions.
• Existing PPI databases are incomplete for any single species, necessitating new methods for network expansion.
• Comparative genomics offers a powerful approach to infer novel interactions.

Purpose of the Study:

• To predict and catalogue protein-protein interactions across five species (human, mouse, fly, worm, yeast).
• To develop a user-friendly tool (InterologFinder) for accessing and navigating predicted PPI data.
• To establish a confidence scoring system (InteroScore) for predicted interactions.

Main Methods:

• Comparative genomics approach utilizing orthology to predict PPIs.
• Development of the InterologFinder web tool for data accessibility.
• Creation of the InteroScore, integrating homology, orthologue evidence, and interaction observations.

Main Results:

• Predicted protein-protein interaction networks exhibit scale-free and small-world properties.
• The addition of predicted interactions did not disrupt established network topology.
• Validation via co-immunoprecipitation confirmed a high rate of accuracy for predicted interactions.
• Demonstrated utility of enhanced interactomes for analyzing large-scale datasets and assigning protein functions.

Conclusions:

• Successfully predicted protein-protein interactions in five species using orthology-based methods.
• The InteroScore provides a robust measure of confidence for known and predicted interactions.
• The InterologFinder website offers intuitive access to a valuable resource for biological research.