中文摘要：

针对3-状态隐马尔可夫模型（HMM）预测蛋白质二级结构准确率不高的问题，提出了7-状态和15-状态HMM。研究对象为CB513数据集合中筛选出的492条蛋白质序列，将其随机均分7组。分别应用7-状态和15-状态HMM对以上数据集进行二级结构预测，对预测准确率进行了7-交叉验证，并将预测结果与应用3-状态HMM的预测结果进行了比较。结果表明，应用7-状态HMM，Q，准确率提高3．11％，SOV提高6．15％，QE提高6．49％；应用15-状态HMM，QF比7-状态HMM又提高5．74％。在15-状态HMM预测中加入序列的同源信息后，Q3准确率比单序列15-状态HMM增加8．76％。结果表明，7-状态HMM预测能力优于3-状态HMM，15-状态HMM总体预测能力和7-状态HMM相当，但8折叠预测能力强于7-状态HMM。

英文摘要：

In view of the lower accuracy of 3-state hidden Markov model （HMM） for protein secondary structure prediction, the study proposed 7-state HMM and 15-state HMM and applied them to prediction of the secondary structure of 492 proteins selected from the dataset CB513, and divided them into 7 even subsets. The prediction accuracy for the two models were evaluated by 7-fold cross validation and the prediction results were compared with 3-state HMM＇ s. It was found that the Q3, SOV and QE of 7-state HMM were increased by 3.11%, 6.15 % and 6.49% respectively compared with 3-state HMM. The QE of 15-state HMM was increased by 5.74% compared with 7-state HMM. The Q3 of 15-state HMM was 8. 76% higher when using multiple sequence alignments than using single sequence. The results show that the prediction ability of 7-state HMM is better than 3-state HMM. 15-state HMM is similar to 7-state HMM in prediction ability, but superior in β-strand prediction.