| Summary: | 碩士 === 國立臺灣大學 === 生物產業機電工程學研究所 === 92 === This study develops a novel method for sequence alignment based on correlation appoach. This approach can be broken down into three stages – correlation, searching and scoring which are decoupling processes. A scoring process is incorporated into the alignment process in the final stage. Therefore, the variation of scoring matrix doesn’t affect the correlation and searching procedures. Moreover, all feasible optimal sequences are ensured to be obtained by using the developed approach. The correlation approach for sequence alignment is mainly based on exclusive NOR operation. If two symbols are matching, output 1as a result, if mismatch, output 0. Subsequently, all feasible paths of optimal sequences are searched and their corresponding sequences are found from the correlation matrix and cumulative correlation matrix. Finally, the score of optimal the sequence is calculated with the specified scoring matrix. Due to the 1st and 2nd stages not considering scoring matrix, the variation of scoring matrix couldn’t affect the aligned results with the correlation approach. Yet, dynamic programming, which is widely used in sequence alignment, often suffers from the recalculation of the whole alignment process while scoring matrices change. Therefore, different optimal sequences are obtained. Compared with dynamic programming, the developed alignment approach with three stages of decoupling processes is proved to provide complete sets of optimal sequences and a more efficient and comprehensive way for sequence alignment.
|
|---|
