NTyroSite: Computational Identification of Protein Nitrotyrosine Sites Using Sequence Evolutionary Features

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Nitrotyrosine is a biomarker for cell damage and inflammation, resulting from tyrosine nitration by reactive nitrogen species.
• Identifying protein nitrotyrosine sites is vital for understanding disease mechanisms and oxidative stress.
• Traditional experimental methods for nitrotyrosine site identification are time-consuming and costly.

Purpose of the Study:

• To develop a novel computational tool, NTyroSite, for accurate in silico prediction of nitrotyrosine sites.
• To leverage protein sequence evolutionary information for improved prediction accuracy.
• To provide a freely accessible resource for high-throughput nitrotyrosine site prediction.

Main Methods:

• Utilized sequence evolutionary information as features for prediction.
• Optimized features using a Wilcoxon-rank sum test.
• Developed a random forest classifier for building the NTyroSite predictor.

Main Results:

• Achieved an Area Under the Curve (AUC) score of 0.904 in 10-fold cross-validation.
• Demonstrated superior performance compared to existing methods in an independent test set.
• Extracted predominant rules and informative features from the model for interpretability.

Conclusions:

• NTyroSite offers a highly accurate and efficient method for predicting nitrotyrosine sites.
• The predictor serves as a valuable computational resource for biological research.
• The tool is publicly available online for broad accessibility and application.