一种基于信息容量的模体比较非比对度量算法

An alignment-free metric algorithm for comparing motifs based on information content

为了有效度量2个模体的相似度,提出一种带有信息容量的用于模体比较的非比对度量算法ICBSM.该算法通过将一个模体的位置频率矩阵在另一个模体的带有信息容量的位置赋权矩阵上滑动,计算出2个模体间的相似度,算法依赖参数少,计算效率高.与其他7种度量法进行比较分析,结果表明:ICBSM可以在数据库查找中更准确地匹配模体,并且能够更有效地聚类相似模体,区分出不相关的模体,为找回模体和分组密切相关模体等方面的研究工作提供有效的计算工具.

In order to measure the similarity of two motifs effectively, an alignment-free metric algorithm with information contents for comparing similarity which named ICBSM was proposed. ICBSM calculated the similarity of two motifs by sliding one motifs position frequency matrix （PFM） on position weight matrix （PWM） of another motif which carried with information contents. This metric needs less parameters and its computational efficiency is high. By comparing with another seven metrics, the results show that ICBSM can retrieve the matching motif while searching database more accurately, it also can group the similar motif and distinguish the irrelevant motif effectively, and it provides an effective tool for retrieving motifs and grouping closely relative motifs.