Research On Method Of RNA Secondary Structure Prediction Based On The Features Of Native Structures

RNAs play various roles in body, they can carry genetic materials and regulate some vitalbiological process, such as RNA splicing and gene expression. Study show that the functionof RNA is determined by structure, one typical case is all tRNA have the similar cloverleafstructure. Research on RNA secondary structures is helpful for us to understand the functionof them.Now, people can get native secondary structure by X-RAY and NMR methods, however,these methods are expensive and time-consuming. To solve this situation, researchersdeveloped sequence alignment method. Though the accuracy of this method is high, it couldnot work without homologous sequence. Besides, manual operation may ignore someimportant information. Because of this, the computational method became the most suitablemethod.There are many perfect RNA secondary structure prediction algorithm, the most popularone is minimum free energy. We need do much more works to solve the problems of currentmethods.We were inspired by helical regions distribution method, in that method, the RNAsecondary structure can be viewed as combination of stems. The prediction of structure iscombining all potential stems based on some specific principle. Here we developed newcombining principle which is based on features of RNA native secondary structures. In stemselection process, the stem which makes the free energy of current structure and with highprobability is chose.To test the effectiveness of our method, we selected several sequences with nativepseudoknot-free structure to predict the secondary structures. The results show that ourmethod can perform well in sensitivity and positive predictive values. Besides, because of theoptimization of potential stems, our method perform better than helical regions distributionmethod.