School of Life Sciences and the State Key Lab of Agrobiotechnology, The Chinese University of Hong Kong, Shatin, Hong Kong.
Laboratory of Bioinformatics, Department of Statistics, University of Rajshahi, Rajshahi 6205, Bangladesh.
Molecules. 2018 Jul 9;23(7):1667. doi: 10.3390/molecules23071667.
Nitrotyrosine is a product of tyrosine nitration mediated by reactive nitrogen species. As an indicator of cell damage and inflammation, protein nitrotyrosine serves to reveal biological change associated with various diseases or oxidative stress. Accurate identification of nitrotyrosine site provides the important foundation for further elucidating the mechanism of protein nitrotyrosination. However, experimental identification of nitrotyrosine sites through traditional methods are laborious and expensive. In silico prediction of nitrotyrosine sites based on protein sequence information are thus highly desired. Here, we report a novel predictor, NTyroSite, for accurate prediction of nitrotyrosine sites using sequence evolutionary information. The generated features were optimized using a Wilcoxon-rank sum test. A random forest classifier was then trained using these features to build the predictor. The final NTyroSite predictor achieved an area under a receiver operating characteristics curve (AUC) score of 0.904 in a 10-fold cross-validation test. It also significantly outperformed other existing implementations in an independent test. Meanwhile, for a better understanding of our prediction model, the predominant rules and informative features were extracted from the NTyroSite model to explain the prediction results. We expect that the NTyroSite predictor may serve as a useful computational resource for high-throughput nitrotyrosine site prediction. The online interface of the software is publicly available at https://biocomputer.bio.cuhk.edu.hk/NTyroSite/.
硝酪氨酸是由活性氮物种介导的酪氨酸硝化反应的产物。作为细胞损伤和炎症的指标，蛋白质硝酪氨酸用于揭示与各种疾病或氧化应激相关的生物学变化。硝酪氨酸位点的准确鉴定为进一步阐明蛋白质硝酰化机制提供了重要基础。然而，通过传统方法对硝酪氨酸位点进行实验鉴定既费力又昂贵。因此，基于蛋白质序列信息的硝酪氨酸位点的计算预测受到高度期望。在这里，我们报告了一种新的预测器 NTyroSite，用于使用序列进化信息准确预测硝酪氨酸位点。使用 Wilcoxon-rank sum 检验对生成的特征进行了优化。然后，使用这些特征训练随机森林分类器来构建预测器。在 10 折交叉验证测试中，最终的 NTyroSite 预测器的接收者操作特征曲线 (AUC) 得分达到 0.904。它在独立测试中也明显优于其他现有实现。同时，为了更好地理解我们的预测模型，从 NTyroSite 模型中提取了主要规则和信息特征，以解释预测结果。我们期望 NTyroSite 预测器可用作高通量硝酪氨酸位点预测的有用计算资源。该软件的在线界面可在 https://biocomputer.bio.cuhk.edu.hk/NTyroSite/ 上公开获取。
