摘要
以chou等人提出的伪氨酸组成方法为基础,从蛋白质序列的氨基酸组成信息和顺序信息着手,提出一种新的伪氨酸组成方法,即利用新伪氨酸序列的自相关函数、氨基酸的平均中程接触和氨基酸频率构造了23维向量来描述蛋白质序列,进而建立多元线性回归函数对蛋白质折叠速率进行预测,经jackknife检验相关系数达到了0.84.并与其他两种方法进行比较使本文的结论得到较好的验证.同时验证了本文提取的特征参数对蛋白质折叠速率有一定的影响.
Based on Pseudo-acid composition,by chou a new pseudo-acid composition is proposed from the amino acid composition information and the order of the protein sequence. Combining the autocorrelation function with the Nm and frequency of amino acids, 23-dimensional vector is constructed, and a protein sequence can be described by the 23-dimensional vector and to create multiple linear regression function to predicte protein folding rate. By jackknife test, the correlation coefficient is 0.84. Comparison proves that the new method is batter than the other two methods.
出处
《大连交通大学学报》
CAS
2015年第3期113-115,共3页
Journal of Dalian Jiaotong University
关键词
蛋白质折叠
伪氨酸
线性回归函数
predicte protein folding rate
Pseudo-acid
linear regression function