KinasePhos

KinasePhos predicts kinase-specific phosphorylation sites in protein sequences to identify kinases responsible for phosphorylation events.

Key Features:

• Data source: Leverages publicly available data on known phosphorylation sites categorized according to their associated protein kinases.
• Kinase-specific grouping: Organizes phosphorylation sites into kinase-specific groups for targeted model training.
• Modeling approach: Utilizes profile hidden Markov models (HMMs) to learn sequence patterns characteristic of kinase-specific phosphorylation sites.
• Model selection and evaluation: Applies a rigorous evaluation process to select the most accurate model from each kinase-specific group, optimizing sensitivity and specificity.
• Prediction output: Produces kinase-specific phosphorylation site predictions for input protein sequences.

Scientific Applications:

• Phosphorylation site prediction: Identifies candidate phosphorylation sites with assigned likely kinases.
• Signal transduction analysis: Supports elucidation of kinase–substrate relationships in signal transduction pathways.
• Cellular regulation studies: Aids investigation of cellular regulation mechanisms mediated by phosphorylation.
• Disease and therapeutic research: Facilitates analysis of disease-related signaling aberrations and the identification of potential therapeutic targets linked to specific kinases.

Methodology:

Uses publicly available, kinase-annotated phosphorylation site data to train profile hidden Markov models (HMMs) on kinase-specific groups and selects the most accurate model from each group after rigorous evaluation.