Analysis And Prediction Of LncRNA-gene Regulatory Relationship

Many studies concentrate largely to investigate of microRNA(miRNA)targeting massager RNA(mRNA).Observations have revealed non-coding RNA(ncRNA)mimic not by their gene of target only,but additionally associate with one another to operate on biological features.In plants,long non-coding RNAs(lncRNAs)play some critical roles in a variety of biological processes by undergoing dynamic regulation and acting in development and stress regulation.In this study,two independent novel models have been proposed for the analysis and prediction of lncRNA-gene regulatory relationship.The first model is based on Traditional Support Vector Regression(SVR);the second model is based on Deep Ensemble Learning.In the first phase,we examined miRNA and their corresponding long non-coding RNA(lncRNA)target applying SVR.In order to interpret remarkable regulatory roles with stress response,there is a need to identify new interactions.We proceeded by the network creation of miRNA targeting mRNA,miRNA targeting lncRNA,and finally,miRNA targeting both mRNA and lncRNA.We examined these network and further illustrated to confirm miRNA with a lower sequence number,targeted lncRNA with a higher sequence number and miRNA with a higher sequence number targeted lncRNA with a lower sequence number.The outcomes of the experiment reveals that there exist a regulatory relationship associating miRNA and lncRNA.Utilizing SVR we predicted hidden interactions and the functions related with stress were annotated.In the second model,we examined the interaction of miRNA-lncRNA sequence by applying the Long Short Term Memory Auto encoder(LSTM-AE)with the same dataset.The experimental outcomes reveal that the LSTM-AE betters the SVR model with accuracies of 99.87%,and 97.13% respectively.These models can correctly predict miRNA interacting with lncRNA and can serve as a foundation for future genomic researches,which is a move towards applying ensemble learning to biological research since the strategy can be perfect for aggregate data taking into consideration of the multifariousness miRNA and lncRNA sequences.