PhosphoPICK

PhosphoPICK predicts kinase substrates by integrating cellular context information with sequence-based methods to improve specificity in protein phosphorylation site identification.

Key Features:

• Integration of cellular context: Incorporates protein-protein interaction data and protein abundance information across phases of the cell cycle, accounting for cell cycle progression, interactions with mediating proteins, and prior phosphorylation events.
• Sequence-context combination: Combines sequence-based phosphorylation site predictions with contextual data to reduce false positives from sequence-only approaches.
• Comprehensive kinase coverage: Makes predictions for 59 human kinases.
• High prediction accuracy: Achieves an average Area Under the Curve (AUC) of 0.86 across the tested kinases.
• Performance for CMGC family: Shows particularly improved prediction performance for kinases from the CMGC family when contextual information is included.
• Functional overlap detection: Identifies functional overlaps between predicted substrates of CDK2 and targets of the E2F family of transcription factors.

Scientific Applications:

• Protein phosphorylation analysis: Serves as a resource for researchers studying protein phosphorylation as a post-translational modification.
• Regulatory mechanism elucidation: Aids in describing cellular events that regulate protein phosphorylation and in elucidating molecular mechanisms in biological pathways and disease states.

Methodology:

Integrates protein-protein interaction data and protein abundance across cell cycle phases with sequence-based phosphorylation site prediction and evaluates performance using AUC across 59 human kinases.