KinasePhos 2.0: a web server for identifying protein kinase-specific phosphorylation sites based on sequences and...

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Protein phosphorylation is vital for cellular regulation.
• Predicting kinase-specific phosphorylation sites is a significant research area.
• Previous tools like KinasePhos 1.0 utilized profile hidden Markov models (HMM).

Purpose of the Study:

• To develop an improved web server, KinasePhos 2.0, for predicting kinase-specific phosphorylation sites.
• To introduce a novel feature, the protein coupling pattern, for enhanced prediction accuracy.
• To provide a freely accessible web server for the scientific community.

Main Methods:

• Implemented support vector machines (SVM) for classification.
• Incorporated protein sequence profiles and a novel protein coupling pattern feature ([XdZ]).
• Computed coupling strength differences between phosphorylation sites and background sequences for SVM model training.
• Evaluated performance using k-fold and Jackknife cross-validation.

Main Results:

• Achieved high average predictive accuracies: 90% for serine, 93% for threonine, 88% for tyrosine, and 93% for histidine.
• KinasePhos 2.0 demonstrated superior performance compared to existing prediction tools.
• The protein coupling pattern proved to be an effective novel feature.

Conclusions:

• KinasePhos 2.0 offers a significant advancement in predicting kinase-specific phosphorylation sites.
• The novel use of protein coupling patterns enhances prediction accuracy.
• The freely available web server facilitates research in cell signaling and disease.