PROFisis

PROFisis predicts protein-protein interaction sites from amino acid sequence data to identify interface residues in proteins lacking experimentally determined 3D structures.

Key Features:

• Machine Learning-Based Approach: Employs a machine learning algorithm to predict interacting residues using amino acid sequence data as the primary input.
• Integration of Structural and Evolutionary Information: Combines predicted structural characteristics with evolutionary conservation data to enhance prediction accuracy.
• High Prediction Accuracy: Demonstrated over 90% accuracy in identifying interacting residues in cross-validation experiments.
• Applicability to Transient Interfaces: Developed using transient protein-protein interfaces derived from experimentally determined 3D structures and intended to generalize across interaction types.

Scientific Applications:

• Drug Target Identification: Pinpoints interface residues that can serve as potential sites for therapeutic intervention.
• Understanding Interaction Mechanisms: Reveals residues involved in protein-protein interfaces to inform mechanistic and biochemical studies.
• Protein Engineering and Design: Guides modifications to interaction sites for protein engineering and synthetic biology applications.

Methodology:

Uses a machine learning algorithm trained on transient protein-protein interfaces from experimentally determined 3D structures, integrating predicted structural features and evolutionary conservation from sequence data, with performance evaluated by cross-validation (>90% accuracy).