Method for predicting RNA secondary structure.

Abstract

We report a method for predicting the most stable secondary structure of RNA from its primary sequence of nucleotides. The technique consists of a series of three computer programs interfaced to take the nucleotide sequence of any RNA and (a) list all possible helical regions, using modified Watson-Crick base-pairing rules; (b) create all possible secondary structures by forming permutations of compatible helical regions; and (c) evaluate each structure for total free energy of formation from a completely extended chain. A free energy distribution and the base-by-base bonding interactions of each possible structure are catalogued by the system and are readily available for examination. The method has been applied to 62 tRNA sequences. The total free-energy of the predicted most stable structures ranged from -19 to -41 kcal/mole (-22 to -49 kJ/mole.) The number of structures created was also highly sequence-dependent and ranged from 200 to 13,000. In nearly all cases the cloverleaf is predicted to be the structure with the lowest free energy of formation.We have developed a technique for predicting the secondary structure of RNA from its primary sequence. The method uses thermodynamic and structural criteria to generate all possible secondary interactions in the molecule and evaluates each for free energy. The technique can be used in conjunction with experimental procedures to elucidate the most favorable conformation of a polyribonucleotide chain.The secondary and tertiary structure of RNA is assumed to play an important role in determining the interactions of these macromolecules with proteins. In the most studied case, that of the tRNAs, it has been found that some form of structural integrity other than the primary sequence of nucleotides must be maintained in order to ensure the biological activity of these molecules (1, 2). This is not surprising, since the tRNAs must be able to interact specifically with a myriad of proteins, including those involved in (a) tRNA maturation (i.e., methylases, thiolases, nucleases, and other modifying enzymes); (b) amino-acid activation (the aminoacyl synthetases); and (c) other translational factors (ribosomal proteins, initiation factor 2, and probably others). In addition to their function in translation, some tRNAs have been shown to be involved in the autogenous regulation of certain cistrons and, thus, must also interact with certain transcriptional components (3).Likewise the rRNAs are believed to possess a definite secondary and tertiary structure which serves to govern their interaction with ribosomal proteins (4). The genomes of certain RNA bacteriophages have also been shown to fold owing to secondary and tertiary interactions (5, 6). In these cases too, the evidence demonstrating the functional significance of specific structure is convincing.Studies attempting to demonstrate structure-function relationships in the RNAs have been hindered by a lack of knowledge concerning the exact nature of the structure of 2017 these molecules in solution. The use of c...

Computational methods in noncoding RNA research

“…A straightforward approach is to enumerate all possible structures and then select the one with the minimum value for the free energy [109], but the exponential time complexity spent in the enumeration step is unfeasible but for the smallest sequences. To deal with this complexity issue, all current methods use a dynamic programming method first proposed by Nussinov et al [104] that reduces the time complexity to O(n 3 ).…”

Predicting rna secondary interactions from primary sequence

“…In all, fifty sequences yielded a cloverleaf or close variant as the lowest free energy structure. Further details of specifics of the method and general tRNA results are available elsewhere [9].…”

An Algorithmic Framework for the Study of Behavior of siRNA Sequences

“…1.1). The initialization of the matrix is carried out as follows: Set G(i,i-1) = 0 for i = 2 to L, and set G(i,i) = 0 for i = 1 to L. (Pipas 1975), thereby ensuring that base pairs do not occur between a residue and the immediate next one. In addition, the entire lower left half of the matrix is blanked out; since we don't use this.…”