Prediction of Pseudoknotted RNA Structure by a Structural Alignment Using GeneticAlgorithm

Developing an efficient algorithm for predicting a pseudoknot structure of RNA has been a challenging problem because of its high computational complexity in both time and space. Structural alignment is one of the RNA secondary structure prediction methods, which predicts a secondary structure by aligning an RNA (called slave sequence) whose structure is unknown to an RNA (master sequence) with a known structure. So far, a few sophisticated structural alignment algorithms that can predict pseudoknots have been proposed [1, 2]. They have been, however, applied to relatively small RNAs (< about 200 nt) due to computational limitations; an exception is RAGA algorithm that is specifically designed for 16S rRNA [3]. In the present study, we have developed an efficient structural alignment program that can be applied to RNAs which are longer than 300 nt and have arbitrary pseudoknot structures; this program is not designed for a specific type of RNA. To show the performance of the developed program, we performed secondary structure predictions using 31 tmRNAs and six Cripavirus internal ribosome entry site (IRES) sequences.