Abstract
A pair of proteins is called a circular permutation (CP) if it has similar sequence compositions and most likely share the same fold but the C-terminal and N-terminal regions of the protein sequence are interchanged. CP is useful in protein engineering; however, the biological functions and origination of naturally occurred CP phenomenon are not clear. Most CP detecting methods are based on structural comparison strategies such as GANSTA+ and CPSARST. But now the amount of 3D protein structures (about 70,000) is much less than the amount of protein sequences (about 15,000,000). To our knowledge, there are only two sequence-based CP detecting methods developed by Uliel et al. [1] and Weiner et al. [2], and these two methods lack obvious criteria to distinguish CPs and non-CPs. Here, we purposed a high efficient and accurate sequence-based detection method based on sequence local alignment with secondary structural elements information. In this thesis, we evaluated our method by different substitution matrices and variant sequence identity data. The result shows that the accuracy of our prediction is highly positive correlated with sequence identity of paired protein. In the future, this developed method is helpful for possible CP site determination and large-scale protein screening.