An Efficient Linear Seeded Tree Algorithm For Finding The Similarity Score of Two RNA Secondary Structures

Abstract

The comparison between two RNA secondary structures is done to predict class and functionalities of unknown organisms and to analyze similarities between them. In previous method of finding similarity scores between two RNA secondary structures, seeded tree alignment algorithm of time complexity O(n3) is used. Here a large RNA tree is converted into a smaller tree known as seeded tree. After that seeded tree alignmentalgorithmisusedtofindsimilaritybetweeneachseedpairsandfinallytheoverall trees. Butinthatmethodnodesateachlevelofsubtreesofseedsarenotalignedandthus itisnot alwayspossibleto obtainaccurateresult. Here, weintroduceanewmethodfor comparing and finding similarity scores between two RNA secondary structures using linearalignmentalgorithm. WeusethepreviousmethodsofconvertingRNAsecondary structure to RNA tree, RNA tree to seeded tree and seeded tree to normalized weighted tree. Then we use our proposed linear alignment algorithm to align each and every nodes of subtrees of seed pairs and compute similarity score between every seed pairs. Finally seeded tree alignment algorithm is used to align and compute similarity score of the whole seeded trees. Although it has the same time complexity as the previous method, it works linearly with multiple alignments and yields to probable correct result. The algorithm is justified by computing similarities among several well-known RNA secondary structures.