不同折叠类型蛋白中基于密码子的二级结构偏向性分析

在传统的Chou-Fasman蛋白质二级结构预测方法的基础上引入同义密码子使用的信息，计算了200个蛋白(49种全α结构蛋白，69种全β结构蛋白，38种仅α β结构蛋白，44种α/β结构蛋白)中不同密码子对应的氨基酸形成不同二级结构(α：螺旋，β：折叠，C：卷曲)的偏向性参数．通过对这些密码子对应氨基酸二级结构偏向性的分析，得到了氨基酸二级结构偏向性分析中所忽略的同义密码子的蛋白结构信息．这些新的信息量对于指导蛋白质设计以及提高蛋白质二级结构预测的准确率有着一定的作用．

Codon-based amino acid propensities for secondary structural units in proteins with different folding types

We have integrated synonymous codon usage information into the amino acid propensities for different protein secondary structural units, which are traditionally used in protein secondary structure prediction method. Codon based amino acid propensities for different protein secondary structural units in proteins with different folding types have been derived from 200 genes coding for proteins with different folding types (49 all alpha proteins, 69 all beta proteins, 38 alpha+beta proteins and 44 alpha/beta proteins). Protein secondary structure information, which is contained not only in amino acid sequences but also in mRNA sequences, is proposed in these synonymous codon specific amino acid propensity parameters. Detailed analysis shows that codon based amino acid propensities are not all the same among different synonymous codons, while they are also different in different protein classes with different folding types. These codon based amino acid propensities we proposed here are helpful to in vitro protein design and to improve the accuracy of protein secondary structure prediction.