GPS: a comprehensive www server for phosphorylation sites prediction.

Area of Science:

• Molecular Biology
• Biochemistry
• Bioinformatics

Background:

• Protein phosphorylation is crucial for cellular regulation, but experimental identification of kinase substrates and phosphorylation sites is challenging.
• Mass spectrometry-based phosphoproteomics generates vast data, yet distinguishing kinase-specific phosphorylation sites remains difficult.
• In silico prediction of kinase-specific phosphorylation sites can guide experimental validation and interpretation of phosphoproteomic data.

Purpose of the Study:

• To develop a comprehensive and accurate in silico prediction tool for kinase-specific phosphorylation sites.
• To provide a resource for researchers to predict phosphorylation sites and their cognate kinases from protein sequences.
• To improve the interpretation of mass spectrometry-based phosphoproteomic data.

Main Methods:

• Collected and curated verified phosphorylation sites from public databases and literature.
• Clustered 216 unique protein kinases (PKs) into 71 groups based on shared sequence motifs and functional patterns.
• Applied a group-based phosphorylation scoring (GPS) method for prediction.

Main Results:

• Developed the GPS prediction server, capable of predicting kinase-specific phosphorylation sites for 71 PK groups.
• The GPS server utilizes protein primary sequences for predictions.
• The tool is implemented in PHP and accessible via a web server.

Conclusions:

• The GPS server offers a valuable resource for identifying kinase-specific phosphorylation sites.
• This tool can significantly aid in the experimental design and interpretation of phosphoproteomic studies.
• GPS enhances the understanding of cellular regulatory pathways governed by protein phosphorylation.