Next: Software platforms and environment Up: Algorithms and Thermodynamics for Previous: Abstract

Introduction

RNA is ubiquitous in the cell and is important for many processes. The activity of RNA is determined by its structure, the way it is folded back on itself. Secondary structure modeling of RNA predicts, or otherwise determines, the pattern of Watson-Crick (WC), wobble and other, non-canonical pairings that occur when the RNA is folded. For the purposes of this article, a non-canonical base pair is defined as non Watson-Crick (non W-C) and non-wobble (not GU or UG). There are many different kinds of RNA. Ribosomal RNA (rRNA) is a crucial part of the ribosome which is found in all living cells and in organelles such as mitochondria and chloroplasts. Small nuclear RNAs (snRNA) form a vital part of sliceosomes that process mRNAs in eukaryotes. These are 2 examples of structural RNAs.

Messenger RNAs (mRNA) do more than just carry information. Secondary structures can be used in part to explain translational controls in mRNA [1,2], and replication controls in single-stranded RNA viruses [3]. Although the vast majority of known mRNAs code for proteins or structural RNAs, some do not [4,5]; and it is likely that the secondary structures of these transcripts play an important role in their regulatory function in the cell. It is to be expected that many more such functional RNAs will be discovered in the future.

RNA is not just a passive structural element or a regulator. It is also an active component in many situations. Thus RNA acting alone is able to catalyze RNA processing [6,7]. In a protein-RNA complex, the RNA component of ribonuclease P is an active component of tRNA processing [8].

The function of an RNA can only be understood in terms of its secondary or tertiary structure. For the understanding of catalytic activity, knowledge of secondary structure alone is insufficient. However, few large structures have been determined by crystallography [9,10,11] and the need for modeling is great. Secondary structure modeling can reasonably be viewed as a first step towards three dimensional modeling. For example, in small and large subunit rRNA, all tertiary interactions, including base triples, involve only 3% and 2% of the nucleotides, respectively [12]. In contrast, nucleotides in secondary structure comprise 60% and 58% of these rRNAs.

Secondary structure modeling is therefore a significant first step to the far more difficult process of three dimensional atomic resolution modeling. Knowledge of secondary structure, together with additional information on structural constraints or tertiary interactions, can be used to construct atomic resolution structural models [13,14,15].

---

Next: Software platforms and environment Up: Algorithms and Thermodynamics for Previous: Abstract

Michael Zuker Center for Computational Biology Washington University in St. Louis 1998-12-05