Similarity Analysis Of RNA Secondary Structures Based On Graphical Representation And Its Application

With the development of transcriptome and epigenetics, the important roles of RNAs played in life processes have been explored repeatedly. Comparative study of the RNA structures is an important approach to assess their biological functions.The researches of genomics show that human genome contains about 3 billion base pairs, and only 2% can encode protein, the rest 98% are non-coding RNAs with variety functions. The functionalities of non-coding RNAs heavily depend on their structures, therefore, it is meaningful to discover and determine the functions of non-coding RNA in the view of structure. Predicting the secondary structure of non-coding RNAs is more difficult than obtaining the original sequence. Recently, Biologists found that comparative study of the non-coding RNA structures can not only provide support for finding new non-coding RNAs, but also offer valuable information to predict the functions of non-coding RNAs. So it is important for us to find a novel and effective method to make similarity analysis of non-coding RNA structures.Pseudo-knot is a kind of RNA tertiary structures formed by stem nesting. Scientists found that some important biological activities are dependent on the pseudo-knot structure of RNA molecule. Pseudo-knot is so complex that predicting it has become a challenge. Therefore, designing algorithms to make comparison for the RNA secondary structures with pseudo-knots have became a more economical and efficient approach of understanding the pseudo-knot functionalities.The research content of this paper is the similarity of RNA secondary structures. We provide a novel 3D graphical representation for similarity analysis of RNA structures. The main work of this paper is as follows:1) We propose a new dynamic 3D graphical representation of RNA secondary structures, and introduce three properties of the 3D graphical representation. According to the 3D graphical representation, we show the projection graphs of the points corresponding to the characteristic sequence, and we can get some information of the base distribution directly from the graphs.2) Based on the properties of the 3D graphical representation of RNA secondary structures, we extract a 36-dimensional vector to represent an RNA secondary structure. To evaluate our algorithm, we perform the test on three sets of RNA secondary structures. Set I are nine different virus; Set II contains 17 complex RNA secondary structures and 16 RNA secondary structures with pseudo-knots; Set III are 60 non-coding RNA secondary structures. The results show the proposed 3D graphical representation method can not only make an effective similarity analysis for RNA secondary structures with pseudo-knots, but also can distinguish the non-coding RNA secondary structures from different families in Rfam. Finally, we compare our similarity measure with the other nine popular RNA comparison methods, and the results illustrated the utility of our approach.