Disease-Related MiRNAs And LncRNAs Prediction Methods

Non coding RNA(ncRNA)is a class of RNA that cannot be translated into proteins.Non-coding RNA,although not translated into a protein,can regulate the target gene at the post-transcriptional level,which in turn affects the disease.A large number of research evidence suggests that mutations and disorders of non-coding RNA are important causes of disease.Therefore,the identification of disease-related non-coding RNA has become an important issue in recent years in the field of biological research.Non-coding RNA is a large family that can be divided into house ncRNA and regulate ncRNA.House ncRNA is closely related to cell function,mainly involved in gene translation,gene splicing and gene modification and so on.The main function of regulating ncRNA is to regulate the expression level of their target genes.In this dissertation,microRNAs(miRNAs)and long non-coding RNA(lncRNA)are selected as the research object,which is mainly based on the following reasons:miRNAs and lncRNAs are all ncRNAs that can regulate the expression of their target genes at post-transcriptional level;The study of miRNAs and lncRNAs is a hot issue of current research;Many miRNAs or lncRNAs mutations and disorders are closely related to a wide range of human diseases;miRNAs and IncRNAs have a relationship between each other,to participate in the impact of the occurrence of a variety of diseases.This paper focuses on the development of effective computational models to predict disease-related miRNAs or lncRNAs,or call it miRNAs-or lncRNAs-diseases association prediction,to provide more accurate experimental subjects for subsequent biological experiments.The main contents are as follows:1)Network Consistency Projection for Human miRNA-Disease Associations Inference.Prediction and confirmation of the presence of disease-related miRNAs help us to understand disease mechanisms at the miRNA level.However,the use of experimental verification to identify disease-related miRNAs is expensive and time-consuming.Effective computational approaches used to predict miRNA-disease associations are highly specific.In this study,we develop the Network Consistency Projection for miRNA-Disease Associations(NCPMDA)method to reveal the potential associations between miRNAs and diseases.NCPMDA is a non-parametric universal network-based method that can simultaneously predict miRNA-disease associations in all diseases but does not require negative samples.NCPMDA can also confirm the presence of miRNAs in isolated diseases(diseases without any known miRNA association).Leave-one-out cross validation and case studies have shown that the predictive performance of NCPMDA is superior over that of previous method.2)Network-based collaborative filtering recommendation model for inferring novel disease-related miRNAs.The current computational methods for predicting disease-related miRNAs,constructing miRNAs similarity network are mostly based on the known miRNAs-disease associations,and the predictive performance is overestimated when using cross validation to evaluate it.In response to this problem,we first propose a new method for measuring disease and miRNA similarities based on standardized mutual information.Subsequently,a network-based collaborative filtering recommendation model,network-based collaborative filtering(NetCF),is proposed for predicting potential miRNA-disease associations by integrating miRNA and disease similarities along with experimentally verified miRNA-disease associations.Leave-one-out cross validation is implemented to evaluate the predicted performance of NetCF.NetCF obtains a reliable AUC value of 0.8960,which is superior to other competitive methods.Case studies on lung and prostate cancer,which is 94%of the top 50 predicted miRNAs of each cancer,have been confirmed by other databases.Furthermore,we demonstrate that NetCF is applicable to diseases withoutany known associated miRNA.3)Global network random walk for predicting potential human IncRNA-disease associations.Effective computational approaches to identify disease-related lncRNAs are in high demand;and would benefit the detection of IncRNA biomarkers for disease diagnosis,treatment,and prevention.In light of some limitations of existing computational methods,we develop a global network random walk model for predicting IncRNA-disease associations(GrwLDA)to reveal the potential associations between IncRNAs and diseases.GrwLDA is a universal network-based method and does not require negative samples.This method can be applied to a disease with no known associated IncRNA(isolated disease)and to IncRNA with no known associated disease(novel IncRNA).The leave-one-out cross validation(LOOCV)method is implemented to evaluate the predicted performance of GrwLDA.As a result,GrwLDA obtained reliable AUCs of 0.9449,significantly outperforming previous methods.The case studies also show that the GrwLDA model has a good predictive performance.