Prediction of secondary structure of RNAs with pseudoknots using matched filter

Prediction of RNA secondary structure is one of the pivotal tasks in bioinformatics. Several computational methods based on dynamic programming, statistical models, have been proposed with considerable success. A typical substructure that occurs in several classes of RNAs, called pseudoknot, plays vital role in many biological processes. Prediction of the pseudoknots in RNA secondary structure is still an open research problem. In this paper, we employ matched filtering approach to determine the secondary structure of a target RNA. The central idea is to match the stem patterns in the base-pairing matrix of RNA with unknown secondary structure. The proposed approach predicts number of stems, loops and also the presence of pseudoknot in the secondary structure of RNA. Illustrative examples on real RNA sequences illustrate the effectiveness and accuracy of our proposed method.