SeRenDIP

SeRenDIP predicts protein-protein interaction (PPI) interface positions from protein sequences to enable interpretation of sequence data lacking structural or experimental annotation.

Key Features:

• Sequence-based Random Forest Methodology: SeRenDIP applies a sequence-based random forest model to predict residue-level PPI interface propensity with improved performance for both homomeric and heteromeric interactions.
• Sequence Conservation Profiling via PSI-BLAST re-mastering: SeRenDIP re-masters alignments of homologous sequences identified by PSI-BLAST to derive conservation profiles.
• Increased Computational Speed: The re-mastering approach yields more than a 10-fold increase in computational speed while maintaining comparable prediction accuracy.
• Residue-level Probability Scores: SeRenDIP outputs per-residue scores indicating the likelihood of participation in an interaction interface.
• Broad Interaction-Type Applicability: As a sequence-based method, SeRenDIP can be applied across diverse protein interaction types without requiring structural input.

Scientific Applications:

• Structural Biology: Assisting identification of potential interaction interfaces for subsequent structural studies.
• Drug Discovery: Aiding target validation and the design of molecules that modulate protein-protein interactions.
• Systems Biology: Mapping protein-protein interactions to enhance understanding of complex biological networks.

Methodology:

SeRenDIP derives sequence conservation profiles by re-mastering PSI-BLAST alignments of homologous sequences and applies a sequence-based random forest model to predict PPI interface positions.