构建基于折叠核心的全α类蛋白取代矩阵

氨基酸残基取代矩阵是影响多序列比对效果的重要因素，现有的取代矩阵对低相似序列的比对性能较低.在已有的 BLOSUM 取代矩阵算法基础上，定义了基于蛋白质折叠核心结构的序列 结构数据块；提出一种新的基于全α类蛋白质折叠核心结构的氨基酸残基取代矩阵——TOPSSUM25，用于提高低相似度序列的比对效果.将矩阵TOPSSUM25导入多序列比对程序，对相似性小于25%的一组四螺旋束序列 结构数据块的测试结果表明，基于 TOPSSUM25的多序列比对效果明显优于BLOSUM30矩阵；基于一个BAliBASE子集的比对检验也进一步表明, TOPSSUM25在全α类蛋白质的两两序列比对上优于BLOSUM30矩阵.研究结果可为进一步的阐明低同源蛋白质序列 结构 功能关系提供帮助.

Construction of All-alpha Protein Substitution Matrix Based on Fold Core

Amino acid substitution matrix is a key factor to protein sequence alignment. However, the existing matrices are not competent for aligning low identity sequence. Based on the algorithm used to construct BLOSUM series, sequence structure “blocks” are defined based on protein fold cores. We have constructed a new matrix, TOPSSUM25, based on all α protein fold core structure of proteins to improve the alignment of low identity sequences. Aligning a four updown helical bundle sequence structure block(＜25% sequence identity) based on TOPSSUM25 and BLOSUM30 respectively indicates that TOPSSUM25 is better than BLOSUM30 in multiple sequence alignment. The alignment test conducted on a BAliBASE subset also indicates that TOPSSUM25 gives better result in pairwise sequence alignment of all α proteins. Our research indicates that some improvements could be managed to the protein sequence alignment under certain conditions, therefore might contribute to the understanding of relationship among sequence, structure and function.