应用HMM和加权距离判别法的真核基因识别程序研究

根据现代科学对基因的认识，应用了隐马尔科夫模型(HMM)的算法，以大量核酸序列为信息来源，通过计算机计算来寻找未知基因的大体位置；再通过基因的固有结构特征及密码子使用的偏向性，使用加权距离判别法来准确地定位基因，以图形及文本的形式输出，从而极大地方便了实验室的研究工作。考虑到基因的许多特征还不为人们所了解，而且不同物种之间基因结构又有一定的差异，所以还开发了程序自学习功能，不断地存储已知的基因，再据此改变一些已有固有数据，以便更好地适应和了解不同生物基因结构的特异性，更加准确地寻找未知基因的位置。

A Hidden Markov Model analysis methods and Frequency of Triplet and Quadruplet for Genes Tracting

In this paper, Hidden Markov Model was applied to hunt for the presumable position of the unknown genes in the abundant sequences of DNA.Based on the inherent features of gene structure and relative synonymous codon usage(RSCU),Improved Weighted Average Algorithm was used to seek out the gene location accurately and then output the data with diagrams or texts.Because some gene features have not been understood and a great deal of small differences among different species, a kind of new self|learning program was established, which could update those original data after storing some known gene sequences,so as to adapt to and recognize the gene structures of those different species.