蛋白质中两种Strand-Loop-Strand模体的判断

Classifying of two kinds of Strand-Loop-Strand protein motifs

目的:蛋白质的结构决定其功能,为了解蛋白质功能,需要了解蛋白质结构。随着基因测序的发展,大量蛋白质一级结构被测出,但直接从蛋白质的序列出发来预测高级结构,仍很困难,尤其是三级结构的预测。然而由超二级结构获得的结构信息可用于三级结构的预测,对蛋白质三维结构及功能预测有重大意义。方法:由一级结构出发,考虑氨基酸序列信息以及氨基酸的亲疏水性,利用Fisher判别法区分两种Strand-Loop-Strand超二级结构模体。结果:采用7交叉检验,Loop长为2~8时,最后平均结果为Q=71.8%,Q_(L)=68.2%,Q_(H)=73.4%,MCC=0.39。如Loop长度相差不大时,预测结果更好,如选取Loop长为2、3、4的序列,7交叉检验结果为Q=75.2%,Q_(L)=69.4%,Q_(H)=77.7%,MCC=0.45。结论:以氨基酸信息为特征指标,利用Fisher判别法能较好地区分两种Strand-Loop-Strand超二级结构模体。

Objective:Protein structure determines its function.In order to understand protein function,it is necessary to understand protein structure.With the development of gene sequencing,a large number of primary structures of proteins have been detected,but it is still difficult to predict the advanced structures directly from the sequence of proteins,especially the prediction of tertiary structures.However,the structural information obtained from the super secondary structure can be used for the prediction of the tertiary structure,which is of great significance to predict the three-dimensional structure and function of proteins.Methods:Based on the primary structure,Fisher's criterion was used to distinguish two strand-loop-Strand supersecondary structure modules,considering the amino acid sequence information and hydrophilicity of amino acids.Results:When the Loop length was 2-8,the final average result was Q=71.8%,Q_(L)=68.2%,Q_(H)=73.4%,MCC=0.39.If the loop lengths differ little,the prediction results are better.For example,if the loop lengths are 2,3 and 4,the 7 cross test results are Q=75.2%,Q_(L)=69.4%,Q_(H)=77.7%and MCC=0.45.Conclusion:Using amino acid information as characteristic index,Fisher's criterion can distinguish two strand-loop-Strand supersecondary structure modules well.