PKSPS

PKSPS predicts kinase–substrate relationships by integrating protein-protein interaction (PPI) network topology and local phosphorylation-site sequence information to assign kinases to phosphorylation sites.

Key Features:

• Integration of Network and Sequence Information: Combines PKSPS-Net and PKSPS-Seq to integrate PPI network topology (quantifying kinase–kinase and substrate–substrate similarities via topological features) with local phosphorylation-site sequence information.
• PKSPS-Net: Uses protein-protein interaction (PPI) network structural properties to quantify kinase–kinase and substrate–substrate similarities based on topological features and employs a maximum weighted bipartite matching algorithm to predict kinase–substrate relationships.
• PKSPS-Seq: Assesses local sequence similarity, sequence motifs, and phosphorylation-sequence enrichment patterns around phosphorylation sites to predict kinases associated with specific phosphorylation sites (KSP).
• Enhanced Predictive Performance: Comparative studies show PKSPS outperforms PKSPS-Net and PKSPS-Seq when used independently and exceeds performance of existing kinase prediction methods.

Scientific Applications:

• Kinase Annotation: Predicts kinases for unannotated phosphorylation sites to expand annotation of kinase–substrate relationships in human proteins.
• Protein Function Analysis: Identifies potential kinase interactions to elucidate protein functions and signaling pathways.
• Drug Discovery and Development: Informs drug-targeting strategies by analyzing kinase activities in diseases with dysregulated phosphorylation.

Methodology:

Network-Based Analysis (PKSPS-Net): Uses PPI network structural properties to infer kinase–substrate relationships via a maximum weighted bipartite matching algorithm.
Sequence-Based Analysis (PKSPS-Seq): Analyzes sequence motifs and enrichment patterns around phosphorylation sites to predict kinases associated with specific phosphorylation sites (KSP).
Effectiveness has been validated through case studies demonstrating accurate kinase predictions for specific phosphorylation sites.