【摘要】 采用氨基酸残基信息熵优化窗口,结合氨基酸的电荷性构建了蛋白质酪氨酸硝基化位点的预测模型。基于10-倍交叉验证,该模型的预测准确率和马氏相关系数分别达到了84.80%和69.69%。同时,对信息熵优化窗口和传统连续窗口进行了探讨,结果表明,信息熵窗口能够有效捕获酪氨酸硝基化肽段上的重要位点,克服短肽序列易丢失信息而单纯增大肽段长度又会引入冗余信息的矛盾,并最终提高模型的预测性能。特征分析揭示酪氨酸残基的局部静电环境、邻近的进化保守位点和长程位点对酪氨酸硝基化均有重要影响。更多还原

【Abstract】 A new method based on information entropy and charge of residue was developed to predict the tyrosine nitration sites.By using the 10-fold cross-validation,the predictive accuracy and Matthews Correlation Coefficient of the model were 84.80% and 69.69%,respectively.Some preliminary discussions were made to the window of information entropy and traditional continuous window.Our results showed that the window of information entropy could effectively capture the important sites in the nitro-tyrosine peptide,which overcomed the contradiction that the short peptide sequence was easy to lose information and the redundant information would be introduced by just increasing the length of peptide.The prediction performance of the model was ultimately improved.Feature analysis revealed that the local electrostatic environment of tyrosine residues,the adjacent evolutionarily conserved sites and long-range sites had some significant influences on tyrosine nitration.更多还原