Microarrays analyses of mRNAs, microRNAs and lncRNAs

Cultured avian BMDCs were randomly divided into either control or IBV-stimulated groups. For IBV stimulated group, cells were incubated at a multiplicity of infection of 1 (MOI = 1) for 12 h. Each group consisted of three wells of BMDCs from three chickens. Total RNA and microRNA were separately isolated using the RNeasy Total RNA Isolation Kit (QIAGEN, Germany). The avian microarray (containing mRNA and lncRNA, RiboArray™ Custom Array (A10000–1-90)) was hybridised. Approximately 3900 lncRNAs and 15,081 mRNAs are detected using the mRNA microarray. In each microarray, we have three replication per slide. Also, we have used the microRNA array (A10000–1-40) to detect the experssion of microRNA. Approximately 991 microRNAs are detected using the microRNA microarray Raw data were normalized using the RMA method [20, 33].

Differentially expressed (DE) mRNAs between control avian DCs group and IBV stimulated group were determined with a cut off of at least 2-fold change and a P value less than 0.01. Such genes were subject to GO categorization, KEGG (Kyoto Encyclopedia of Genes and Genomes) and BioCartapathway analyses. Analyses were performed with DAVID (the Database for Annotation, Visualization and Integrated Discovery) by using an independent list of differentially-expressed genes.

DE microRNAs were chosen with a cut off of at least 2-fold change and a p value less than 0.05. Potential targets of these microRNAs were predicted using the microRNA target prediction and functional study database (miRDB) and TargetScan. Taking the intersection of these two predictions, we obtain the optimal potential target genes. To further understand the potential functions of microRNA - target genes, GO categorization and pathway analysis were assigned using the DAVID gene annotation tool.

DE lncRNAs between IBV stimulated group and control group were first chosen with a cut off of at least 2-fold change and a P value less than 0.01. Since transcriptional regulation by lncRNAs could work either in cis or trans model, we then predicted the cis and trans target gene of difference expressed lncRNAs as previous published manuscript [34]. To further classifly lncRNA trans-target genes, the RNAplex program (RNAplex < − 100) was then used to identify possible trans-target genes of the lncRNAs. Thirdly, the Pearson correlation coefficient absolute value over 0.9 together with P value less than 0.01 were used to predict the lncRNA’s co-expression target genes. Finally, the lncRNAs and co-expression genes relationship networks were drawn using Cytoscape software [35].