蛋白质二级结构序列片段的识别

RECOGNITION OF SECONDARY STRUCTURE SEGMENTS OF PROTEINS

识别蛋白质二级结构片段，实质上是从序列片段的水平上对二级结构进行预测，用离散量的方法，通过一系列的计算发现：选择氨基酸（20种氨基酸加一个空位）和其紧邻关联共同为参数，从N端截取固定长序列片段为6或7个连续氨基酸残基时，10交叉检验平均预测精度能达到75％～77％，jack—knife检验平均预测精度能达到72％；当固定长序列片段为9或10个连续氨基酸残基时，10交叉检验平均预测精度能达到82％～83％，jack—knife检验平均预测精度能达到74％～76％.

The recognition of secondary structure segments of proteins is,in substance ,a prediction of secondary structure at the level of segments. By use of the measure of diversity, a series of calculations are made. It is found that,with amino acid composition and dipeptide composition taken as parameters,in fixed-length sequence with 6 or 7 amino acids from N terminal,the average prediction accuracy of 10-cross validation test reaches 75%-77%, while with jack-knife test the average prediction accuracy is no more than 72%. In case the fixed-length sequence chosen contains 9 or 10 amino acids,the average prediction accuracy of 10-cross validation test will reach 82%,while with jack-knife test it will be 74%-76%.