有效预测RNA二级结构是生物信息学中的重要研究领域.提出一种基于隐Markov模型预测RNA二级结构的新方法.首先,应用前后缀匹配算法快速找到所有可能(包括假结)的茎区,建立RNA-HMM,寻找最优的茎区组合方法,得到包含假结的RNA二级结构.实验结果表明,提出的新方法降低了计算复杂性,提高了预测的特异性和敏感性,具有较高的准确率,可以预测RNA的假结结构.%In multicore system, system execution efficiency presently has gottn increasing concerns. Generally, a whole system includes several modules and some optimization work has been done on these modules. Given an integrated system including Tomcat, Httpd and Lucene, each of them has processed some optimization to reach the favorable performance. However, when they constitute an integrated system, the system can not have good performance. Based on the deep research for the characteristic of each subtask in the system, several parallel ways are presented to improve the whole execution efficiency. The proposed methods involve: 1) Cancelling the lock of shared object or files; 2) Rearranging subtask; 3) Removing the system call from the multi-thread operation. Experimental results show that the whole performance gets improved and each function the subtask focuses on is more distinct.
展开▼
