IKAP

IKAP infers kinase activities from phosphoproteomics data to estimate kinase activity profiles and identify kinase-phosphosite interactions for cellular signaling analysis.

Key Features:

• Heuristic machine learning algorithm: Employs a heuristic machine learning approach to infer kinase activities from quantitative phosphoproteomics data.
• Integration with PhosphoSitePlus: Leverages kinase–target relationships from the PhosphoSitePlus database to map kinases to specific phosphosites.
• Comparative analysis of profiles: Compares estimated kinase activity profiles with measured phosphosite profiles to identify kinases likely responsible for observed phosphorylation changes.
• Application to experimental datasets: Applied to human cell cycle data in HeLaS3 cells and to insulin signaling dynamics in mouse hepatocytes.
• Discovery of novel kinase-phosphosite links: Enabled derivation of 94 novel kinase-phosphosite connections in cell cycle studies and 49 in insulin signaling analyses.
• Implementation: Implemented in MATLAB using the Optimization and Statistics toolboxes.

Scientific Applications:

• Cell cycle signaling (HeLaS3): Infers kinase activity dynamics across the human cell cycle in HeLaS3 datasets.
• Insulin signaling (mouse hepatocytes): Analyzes temporal kinase activity during insulin signaling in mouse hepatocytes.
• Translational research: Supports investigations into signaling network organization, disease mechanism elucidation, drug target identification, and therapeutic intervention studies.

Methodology:

Uses a heuristic machine learning algorithm that integrates kinase–target information from PhosphoSitePlus, performs comparative analysis between estimated kinase activity profiles and measured phosphosite profiles, and is implemented in MATLAB using the Optimization and Statistics toolboxes.