predPhogly-Site: Predicting phosphoglycerylation sites by incorporating probabilistic sequence-coupling information...

Area of Science:

• Biochemistry
• Molecular Biology
• Computational Biology

Background:

• Post-translational modifications (PTMs) are critical for protein function and cellular processes.
• Lysine phosphoglycerylation is a newly identified PTM impacting glycolytic enzymes and linked to diseases such as heart failure and neurodegeneration.

Purpose of the Study:

• To computationally identify and characterize potential phosphoglycerylation sites in proteins.
• To enhance the understanding of phosphoglycerylation's role in cellular functions and disease causality.

Main Methods:

• Development of a novel computational tool, predPhogly-Site.
• Utilizing probabilistic sequence-coupling information of amino acid residues.
• Incorporating variable cost adjustment for imbalanced training data.

Main Results:

• Achieved approximately 99% accuracy in predicting phosphoglycerylation sites.
• Obtained a Matthews Correlation Coefficient (MCC) > 0.96 and Area Under the Curve (AUC) > 0.97.
• Demonstrated negligible standard deviation in 10-fold cross-validation, indicating robust performance.

Conclusions:

• predPhogly-Site significantly outperforms existing prediction tools for phosphoglycerylation sites.
• The tool offers a promising approach for studying PTMs and their associated diseases.
• A web interface is available for broader accessibility and application.