蛋白质结构型预测的研究

The Study of Protein Structural Class Prediction

以 nα、nβ、nβαβ、n4α、n4β等五个二级结构参数为基础 ,用离散增量研究了蛋白质的聚类和蛋白质的结构型预测 .α型 ,β型和 αβ型可以很好地分别聚在三个大支中 .蛋白质结构型预测总体预测正确率为 82 % ,用 Self-consistency和 Jake-knife这两种方法测试的结果没有明显的差别 .从结果可以看出利用蛋白质的二级结构参数能较好地体现出各种结构型蛋白质的特点 .

Based on secondary structural parameters n α?n β?n βαβ ?n 4α and n 4β , we use the method of increment of diversity to study the clustering and the structural class prediction of protein. The result of protein clustering shows that the α class, β class and αβ class in individual branch. The total rate of correct prediction of protein structural class is 82%. The result of self consistency test and jack knife test hasn′t difference. According to the result of protein clustering and protein structural class prediction , we suggested that the secondary structural parameters n α?n β?n βαβ ?n 4α and n 4β can represent the feature of protein structural classes.