Phos3D

Phos3D predicts protein phosphorylation sites by integrating sequence profiles with three-dimensional structural information to identify kinase-specific target-site signatures.

Key Features:

• Support Vector Machine-Based Prediction: Employs Support Vector Machines (SVMs) trained on sequence profiles augmented with spatial context to predict P-sites.
• Integration of Spatial Context Information: Incorporates three-dimensional structural data to represent the spatial distribution of amino acids around phosphorylation sites beyond one-dimensional sequence information.
• Kinase-Specific Analysis: Groups experimentally verified P-sites according to kinase families to derive kinase-specific target-site characteristics.
• Utilization of 3D Structural Profiles: Derives signature 3D-profiles for phosphoserines, phosphothreonines, and phosphotyrosines from spatial distributions of surrounding residues.
• Performance Improvement: Inclusion of 3D-context yields a small but consistent improvement over sequence-only methods for recognizing kinases and their target sites.

Scientific Applications:

• Protein phosphorylation research: Improves prediction of phosphorylation sites by combining sequence and structural information.
• Kinase–substrate interaction analysis: Enables analysis of kinase-family-specific target-site signatures to inform studies of kinase specificity.
• Signaling pathway target identification: Supports identification of potential therapeutic targets within signaling pathways based on predicted P-sites.
• Drug development support: Aids development of drugs targeting specific phosphorylation events by providing more accurate site predictions.

Methodology:

Phos3D characterizes spatial context using Protein Data Bank (PDB) structures, analyzes 750 non-redundant experimentally verified P-sites to derive spatial amino-acid distributions around phosphoserines, phosphothreonines and phosphotyrosines, combines these 3D profiles with sequence data, and trains Support Vector Machines (SVMs) for prediction.