Uncovering Phosphorylation-Based Specificities through Functional Interaction Networks.

Area of Science:

• Biochemistry and Molecular Biology
• Computational Biology
• Proteomics

Background:

• Protein kinases regulate cellular processes via phosphorylation, but their specificities are often unknown.
• Current knowledge covers binding sites for only ~30% of human kinases.
• Understanding kinase specificity is crucial for deciphering cell signaling mechanisms.

Purpose of the Study:

• To develop and validate a computational method for predicting protein kinase substrate specificities.
• To expand the known specificities for a significant portion of human kinases.
• To assess the method's applicability across different organisms and other phospho-recognition domains.

Main Methods:

• Utilized functional interaction and phosphorylation data for computational prediction of kinase specificities.
• Applied the method to human kinases, predicting substrate preferences for 57% of known kinases.
• Validated predictions using in vitro mass spectrometry for four understudied kinases.

Main Results:

• Successfully predicted substrate preferences for 57% of human kinases.
• Reconstructed well-established kinase specificities, confirming method accuracy.
• Experimental validation showed predicted models closely matched true specificities for understudied kinases.

Conclusions:

• The computational method effectively predicts protein kinase substrate specificities.
• This approach significantly expands the catalog of known kinase specificities.
• The method is applicable across species and adaptable to other post-translational modification (PTM) recognition domains, offering insights into signaling networks.