病毒与宿主细胞在遗传信息上具有相似的字模式(k-tuple)，病毒的 DNA 序列与 其可感染的宿主细胞的 DNA 序列通过字模式的统计打分值往往比与随机宿主细胞的打 分值高，也就是病毒和其可感染的宿主细胞的 DNA 序列有一定的相似性． 基于此原理，文 中利用序列非比对统计方法 D^S₂ 和 D^*₂ 对病毒的 DNA 序列和宿主细胞的 DNA 序列基于 字模式进行比对打分，将打分值与获得的阈值进行比较，判断该病毒是否能感染宿主细 胞． 实验结果表明，当 k =5(k 为字模式的的大小)、马尔可夫阶次为 1 时，D^S₂ 和 D^*₂ 统计 量均能较好地反映病毒与宿主细胞在基因上的相似性，而且通过 ROC(受试者工作特征 曲线)获得的最佳阈值可以作为一种判断病毒是否可感染宿主细胞的方法．

A virus and its host cell have a similar word pattern (k-tuple).The scores of the DNA sequences of the virus and its host cell,which are obtained by means of the word pattern,are often higher than those of random host cells,that is to say,the DNA sequence of the virus is similar to that of its host.On the basis of this principle,two alignment-free statistics D^S₂ and D^*₂ are adopted to acquire the scores between the DNA sequence of the virus and that of its host cell in this paper.Then,the scores are compared with the threshold，so as to judge whether the vi- rus can infect the host cell.Experimental results show that,when k =5 (k is the size of k-tuple) and Markov order is 1,both of the statistics and can describe the similarity between the virus and its host cell in genes，and that the optimal threshold of D^S₂ and D^*₂ from the ROC (Receiver Operator Characteristic) curves can be used to judge whether the virus can infect the host cell.