蛋白质中钙离子结合残基的识别

钙结合蛋白在诸多重要的生命进程中实现着不可替代的生物学功能。而这些功能的实现,均取决于蛋白质中配体结合残基与钙离子的相互作用。因此,对蛋白质中钙离子结合残基的识别是理解这种重要分子机制的有效手段。建立了396条非冗余蛋白质链共包含1952个钙离子结合残基的钙离子结合蛋白数据集,通过统计分析确定以17个氨基酸残基作为最佳片段长度。使用10交叉检验,以氨基酸组分为特征参数的离散增量算法的预测精度为62.4%,相关系数为0.25;以位点氨基酸保守性信息为特征参数的矩阵打分算法的预测精度为69.9%,相关系数为0.40;以离散增量值、矩阵打分值和自协方差值为特征参数的支持向量机算法的预测精度为75.0%,相关系数为0.50。

THE RECOGNITION OF CALCIUM BINDING RESIDUES IN PROTEINS

Calbindin plays a critical role in many important life processes .The realization of this biology function depends on interaction between the ligand binding residues and calcium ions .Thus ,the recognition of calcium binding residues is an effective method to understand this molecular mechanism . Based on 396 calcium binding proteins containing 1952 calcium binding residues ,we built a non -redundant calbindin database .Through statistical analysis ,optimum length of sequences segments was selected as 17 amino acid residues .With amino acid composition as the inputted parameter for increment of diversity algorithm ,the overall prediction accuracy and MCC was 62 .4% and 0 .25 by 10-fold cross-validation ,respectively .When with conservation of position as the parameter for the matrix scoring value algorithm ,the overall prediction accuracy and MCC achieved 69 .9% and 0 .40 by 10 -fold cross -validation ,respectively .While with increment of diversity values ,matrix scoring values and auto covariance of physicochemical property index as the parameters for support vector machine , the overall prediction accuracy and MCC achieved 75% and 0 .5 by 10-fold cross -validation ,respectively .