FpClass

FpClass predicts proteome-wide protein-protein interactions to expand the human interactome and support studies of protein function, disease associations, and drug mechanisms.

Key Features:

• Data mining-based methodology: Uses a data mining-based approach to generate proteome-wide PPI predictions.
• Predicted interactions: Predicted 250,498 potential interactions among 10,531 human proteins.
• Novel interactions for uncharacterized proteins: Identified 10,647 PPIs involving 1,089 proteins that previously had no known interactions.
• False discovery rate estimate: Estimates a false discovery rate of approximately 60% for its predictions.
• Experimental validation: Experimentally tested 233 high- and medium-confidence predictions and validated 137 interactions.
• p53 interactors: Validated seven novel putative interactors of the tumor suppressor p53.
• Agreement with experimental data: Demonstrated superior agreement with experimentally detected PPIs compared to previous prediction methods.

Scientific Applications:

• Interactome expansion: Expands the mapped human interactome by predicting proteome-wide PPIs including interactions for previously uncharacterized proteins.
• Protein function prediction: Supports inference of protein function through predicted interaction partners.
• Disease association: Aids in linking proteins to diseases via predicted interaction networks.
• Drug mechanism elucidation: Facilitates investigation of drug mechanisms of action through predicted PPIs.
• Experimental prioritization: Provides candidate interactions for targeted experimental validation.

Methodology:

Employs a data mining-based approach to predict PPIs proteome-wide and reports an estimated false discovery rate of ~60%.