Finding Common RNA Secondary Structures from RNA Sequences

Abstract

RNAs (Ribonucleic Acids) play an important role when organisms reproduce themselves. RNAs are single-stranded, however they tend to form higher order structures such as secondary or tertiary structures by folding onto themselves. It is the RNA structures that determine the functions of RNA sequences. Since it is very difficult to crystallize and/or get nuclear magnetic resonance spectrum data for large RNA molecules, reliable methods to determine RNA structures from the primary sequences is important. An important step toward the deter- mination of RNA structure is the prediction of RNA secondary structures. Based on a reliable RNA secondary structure, possible tertiary interactions that occur between secondary structural elements and between these elements and single- stranded region can be characterized. Thermodynamic stability methods have been developed [5] to fold a single RNA into secondary structures with minimum or near minimum energy with some success. Phylogenetic comparative methods are more successful which try to determine the common secondary structures from a set of RNA sequences by checking a large number of possible base pair- ings for their possible conservation. However this method is very tedious since it is basically performed manually. In this abstract, we propose an algorithm using dynamic programming trying to automate the phylogenetic comparative pro- cess. Given three RNA sequences, we first apply the folding algorithms for each sequence to determine the frequently recurring stems which are considered to be thermodynamically favourable. We then apply our algorithm to the three stem lists generated from the folding algorithm to determine the common secondary structures.We have applied our method to three viruses: cocksackievirus, human rhinovirus (type 14), and poliovirus (type 3). Our method successfully produced the main components of the common secondary structures of these viruses.