Prediction Of RNA Secondary Structure Based On Machine Learning And Long-range Interaction Quality

As a macromolecule in organisms,ribonucleotide molecular RNA is an important substance present in organisms.It not only collaborates with deoxyribonucleotide molecular DNA and proteins to maintain the activities of organisms.It also plays an important role in DNA and protein synthesis.The research on the structure of RNA can help us to understand the function of RNA molecules more comprehensively,which will help biological researchers to explore the relationship between RNA,DNA and proteins,understand the function of organisms and treat the diseases.The RNA molecular structure consists of three parts: a primary sequence,a secondary structure,and a tertiary spatial structure.The RNAtertiary spatial structure is a stable structure formed in space by the interaction between secondary structural units,distortion,folding,etc..Therefore,the prediction of RNA secondary structure plays an important role in RNA structure.How to effectively predict RNA secondary structure has become one of the important research problems in the field of bioinformatics.Conventional methods for predicting RNA secondary structures are all based on experimental physics,chemistry,and other methods.However,RNA molecules themselves have the characteristics of difficult to obtain crystals and rapid molecular degradation.Therefore,the use of physical experiments to predict its secondary structure is relatively time-consuming and costly.In recent years,artificial intelligence algorithms have increasingly emerged,and the use of artificial intelligence methods has achieved very good results in predicting the application of various types of data such as stocks,house prices,and so on.In this paper,the different algorithms in machine learning in artificial intelligence are compared and analyzed.Firstly,the biological concept of RNA secondary structure and the research status of RNA secondary structure prediction are introduced.According to the correlation algorithm of machine learning,the main research work of this paper is determined:(1)Conduct in-depth research on various algorithms in machinelearning in the prediction of RNA secondary structure,analyze and compare the principles and performance of these algorithms.How these algorithms predict RNA secondary structure and its prediction efficiency results are further compared and analyzed to select the best algorithm model.(2)We further discovered the characteristic vector of the number of base pairs of RNA through the characteristic information of the internal long-term correlation of RNA,and incorporated it into our previous algorithm model,thus greatly improving the efficiency of the prediction of RNA secondary structure.(3)There are some secondary structural features of RNA that have not been found in biology.Therefore,we use the cyclic neural network(RNN)algorithm in the deep learning algorithm as a new model.The advantage of this model is that it can map data to high-dimensional space.The model independently discovers the intrinsic features that we can not find with the naked eye or simple algorithms in the RNA primary sequence.