Study On The Identification Of Key Mi RNAs Of Disease And Relationship Between Mirnas And Drugs

miRNAs are a class of single-stranded non-coding RNAs with a length of approximately 22 nt.Although miRNAs cannot directly encode proteins,they can regulate gene expression through post-transcription,and then participate in various biological processes such as cell growth,differentiation,and apoptosis.Existing research shows that miRNAs are widely involved in the occurrence and formation of various diseases including cancer,which makes miRNAs a promising target for disease diagnosis and treatment.At present,most studies believe that miRNAs are involved in the process of disease by suppressing gene expression.In recent years,studies have found that miRNAs can also promote gene expression.In addition,due to the multi-targeting,miRNAs also have an important impact on drug sensitivity and drug resistance.In this paper,we will conduct research on the identification of key disease-related miRNAs and the prediction of potential miRNAs-drug resistance/sensitivity relationship based on miRNAs and gene high-throughput sequencing data,as well as public databases such as mi RNA-gene,PPI,and mi RNA-drug relationship networks.1.Identification of key miRNAs related to disease:We propose a method for systematic screening of key miRNAs for disease.This method comprehensively considers the ‘many-to-many' targeted regulatory relationship between miRNAs and genes,the complex regulatory mechanism of miRNAs on target genes(promotion/suppression),and the importance of target genes in protein networks.In this part of the study,we propose a linear model for calculating the contribution of specific miRNAs to changes in target gene expression.Based on the influence intensity of miRNAs on target genes and the importance of target genes,key miRNAs in diseases are identified.We appliy this method to the three cancer data sets of breast cancer,lung cancer and prostate cancer.The results confirm that miRNAs have two effects on target genes: promotion and inhibition,and the two mechanisms account for a similar proportion.The top ranked cancer key miRNAs screened by this method have been shown to significantly regulate cancer-related signaling pathways,affect the therapeutic effect of anticancer drugs,and have a significant relationship with the prognosis of cancer.2.Prediction of the relationship between miRNAs and drug sensitivity/drug resistance:We propose a method based on Bi-Random Walk(BiRW) to predict the relationship between potential miRNAs and drug sensitivity/drug resistance.The existing relationship between miRNAs and drug sensitivity/drug resistance is mainly based on the evidence collected in the experiment,and the relationship is relatively sparse.In this paper,a heterogeneous network based on miRNAs similarity,drugs similarity,and mi RNA-drug association is constructed.And then we use Bi RW to predict the potential mi RNA-drug sensitivity/drug resistance relationship.Through cross-validation and De novo verification,this method has high reliability.