Study On Peripheral Blood Circular Rnas As Diagnostic Markers For Infectious Diseases

Circular RNAs(circRNAs)are a group of noncoding RNA molecules which are widely expressed in eukaryotes and form a closed loop structure through back-splicing.Existing studies have shown that circRNAs have great biomarker potential for application in liquid biopsy due to their high stability,ubiquity,spatiotemporal specificity and conservation.This thesis addresses the issues related to peripheral blood circRNAs as disease biomarkers,and explores the possibility of peripheral blood circRNAs as diagnostic biomarkers for infectious diseases.The main research contents of the thesis include:(1)Development of a bioinformatics tool,AQUARIUM,for accurate quantification of circRNAs expression in high-throughput RNA-seq data from peripheral blood samplesThe main strategy of the existing quantitative tools for circRNAs is to identify and count the number of sequencing reads that support back-splice junctions(BSJ).However,the sequencing reads are not evenly distributed in high-throughput sequencing.Therefore,using the number of sequencing reads that cross BSJ sites to represent the expression of circular transcript may introduce large biases.Sailfish-cir,the model-based quantification tool,counts reads from circRNA across the entire transcript by transforming circular transcript to pseudolinear transcript,which outperformed the above count-based tools.However,since it cannot reconstruct the internal structure of circRNA,Sailfish-cir cannot determine the true distribution of circRNA’ reads.AQUARIUM,we newly proposed pipeline,compiles and integrates sequences of circRNAs reconstructed by CIRI-full with different states,and then performs pseudo-linearization and employs Salmon,a quantification tool that does not require alignment,to observe the read distribution of circRNAs on the entire transcript and estimate their expression.AQUARIUM can estimate the abundance of circRNAs at the BSJ level and isoform level,and the quantification performance is significantly better than the existing commonly used quantification methods.In addition,we integrated the full-length sequences of circRNAs from blood samples parsed by Nanopore sequencing into the AQUARIUM reference sequence to complement the circular transcripts that could not be reconstructed by CIRI-full,thus AQUARIUM achieving further quantification optimization of circRNAs from RNA-seq data in peripheral blood samples.AQUARIUM is available at https://github.com/wanjun-groupseu/AQUARIUM.(2)Study of the effect of different sample processing on the expression of peripheral blood circRNAsSample incubation has been shown to produce significant changes in blood transcriptome expression,and circRNAs are more stable than linear RNAs,but there is no systematic study on how they behave in sample processing.We performed library construction and highthroughput sequencing by collecting peripheral blood from three healthy individuals,isolating PBMC and extracting RNA after incubation at 4 ℃ for 0 h,2 h,6 h,12 h,24 h,and 48 h,respectively.We systematically analyzed and compared for the first time the changes in expression of m RNAs,lnc RNAs and circRNAs caused by different treatments of the samples,the number of newborn or degraded transcripts and the changes of alternative splicing events of m RNAs and circRNAs.The results showed that circRNAs were significantly more stable than linear RNAs both at the transcriptional and post-transcriptional levels during the delayed sample processing.The blood storage times can be up to at least 48 h for circRNAs research.(3)Study of peripheral blood circRNAs as biomarkers for the diagnosis of two infectious diseases(Tuberculosis and Chikungunya)On the basis of our two works to reduce the noise of circRNA quantification(i.e.using AQUARIUM for circRNAs abundance estimation)and the noise of sample processing(i.e.controlling blood storage time),we analyzed the peripheral blood circRNA expression profiles of patients with tuberculosis caused by Mycobacterium tuberculosis and patients with Chikungunya caused by Chikungunya virus,and identified 14 circRNAs in tuberculosis and 12 circRNAs in Chikungunya,which could well distinguished patients from healthy controls in both the discovery set and the validation set,validating the possibility of peripheral blood circRNAs as diagnostic markers for infectious diseases.Meanwhile,we filled the gap of the interaction between Chikungunya virus and circRNAs.(4)Exploration of the possible mechanisms of peripheral blood circRNAs involved in the disease progression using tuberculosis as an exampleWe analyzed and compared the biological pathways of differentially expressed circRNAs enrichment in the peripheral blood of four patients(tuberculosis patients,latent tuberculosis patients,pneumonia patients,and lung cancer patients),with the same disease site and similar preexisting symptoms,and for the first time identified that circRNAs were specifically involved in γ-interferon-related regulatory pathways in TB.Based on common mi RNA binding sites and expression correlations,we constructed a circRNA-mi RNA-m RNA interaction network by combining differentially expressed m RNAs enriched into the γ-interferon regulatory pathway,candidate circRNAs identified by WGCNA and fitted linear models,and differentially expressed mi RNAs.Circ RNAs may regulate the expression of γ-interferon through the ce RNA regulatory axis and thus influence the onset and progression of tuberculosis.Moreover,five circRNAs in the ce RNA network showed good specificity in differentiating tuberculosis patients.The ce RNA regulatory axis provides a possible mechanism for the immune regulation of circRNAs in tuberculosis,and also provides circRNA candidates deserving to biomarker identification and functional validation.