| In recent years,with the fast development of modern life science,biologists have discovered that,besides playing a crucial role in gene's expression,RNA widely participate into many kinds of cell's regulation.How to utilize the known RNA sequence information to predict its structure has become a hot spot in the research area.RNA secondary structure prediction is the first step while the key step in prediction of molecular structure of RNAs.In tradition,there exists mainly two kinds of algorithms for RNA secondary structure prediction.One is usually called sequence alignment,whose quality heavily depends on the strong relationship between query sequence and database.The other one can be called minimum free energy methods,which can overcome the heavy dependence on the database,but usually requires exponential computation time due to the complex free energy model.This article proposes a de novo RNA prediction algorithm using a comparably concise model to simulate the process of RNA folding.The algorithm's performance can reach a good balance between quality and efficiency.The proposed algorithm is based on such a biological truth:RNA chain folds to its structure within very short time.There are mainly two factors that determine RNA structure:bases types and distances.Firstly,an RNA database was set up to store RNA structural data,the original data comes from RSCB.Second,statistical methods were used to analyze the classes and the distances of RNA base pairs and generate parameters for the folding formula.At last,based on the folding formula,we have built an algorithm that searches the query sequence for high-score folding stem iteratively to generate the final secondary structure.After a comparative analysis between our method and other state-of-the-art works of RNA secondary structure prediction,we have found that our algorithm has competitive predictive quality.Because of the comparably concise folding model we use,our algorithm has lower time complexity than existing methods. |
PDF Full Text Request