RegPhos

RegPhos: Human kinase–substrate phosphorylation network construction and analysis

RegPhos integrates experimentally verified human protein phosphorylation sites with protein–protein interaction data to construct and analyze kinase–substrate phosphorylation networks.

Key Features:

• Experimental phosphorylation dataset: Compiles 21,110 experimentally verified phosphorylation sites across 5,092 human proteins for network construction and analysis.
• Kinase annotation and prediction: Uses public database kinase annotations for ~20% of sites (4,138) and integrates KinasePhos to predict catalytic kinases for unannotated phosphorylation sites.
• Network exploration with localization context: Accepts user-specified groups of human proteins to generate associated phosphorylation networks and provides subcellular localization information for network members.
• Expression profile similarity analysis: Uses time-coursed microarray expression data to assess similarity among network members to support identification of co-regulated proteins.
• Pathway discovery support: Identifies known pathways and detects candidate novel signaling pathways interacting with established networks (demonstrated in insulin signaling network analysis).

Scientific Applications:

• Signal transduction network analysis: Constructs and analyzes intracellular kinase–substrate networks to study phosphorylation-mediated regulation.
• Protein kinase regulation: Supports identification of key regulatory nodes and potential cross-talk between signaling pathways.
• Disease mechanism investigation: Enables phosphorylation network analysis relevant to complex cellular processes and diseases where phosphorylation is a key regulatory mechanism.

Methodology:

RegPhos aggregates large-scale experimentally verified phosphorylation site data, integrates protein–protein interaction information to build kinase–substrate phosphorylation networks, applies KinasePhos to predict kinases for unannotated sites, incorporates subcellular localization for functional context, and uses time-coursed microarray expression similarity to contextualize network dynamics.