基于SDTW算法的DNA序列相似性分析

Similarity analysis of DNA sequences based on SDTW algorithm

DNA序列相似性分析是生物信息学中最主要的内容之一,它的研究对于发现物种间的进化关系、判断血缘关系、治疗疾病等有着至关重要的作用。利用图形表示方式来分析DNA序列相似性时,局部差异是反映相似性的重要内容,但某些局部差异累积会导致本来十分相似的DNA序列在全局上呈现出较大的差异,从而导致误判。根据这一思想,本文提出基于SDTW算法的DNA序列相似性分析,该算法通过合理的分段既保持了局部差异的作用又在一定程度上控制了局部差异对全局差异的影响范围。文中以9个物种β-球蛋白第一个外显子的编码序列作为分析对象,将该算法与已有算法的分析结果进行比较。结果表明本文提出的算法相似性分析更加准确,分析结果的敏感性较高,有助于提高进化距离较近的分析对象间的区别度,可将其进一步应用于生物序列的信息分析。

When using graphical representation of DNA, partial differences are important elements to evaluate similarity of DNA sequence. With the accumulation of partial differences, some DNA sequences show great difference in global. It can lead to misjudgment. Based on the analysis of deficiency of existing DTW algorithm,a novel algorithm about similarity analysis of DNA sequences based on segmentation is proposed in this paper. By reasonable segmentation, this method maintains effects of partial differences, meanwhile, limits the scale of partial differences. Finally, the first exon coding sequences of the β-globin of nine species were analyzed. The result was compared with existing algorithm. Our method provides more exact result, higher sensitivity, which could help to promote categorization ability of species.