Technical replicates for individual miRNAs were averaged using the median signal intensity. Box plots and cluster analyses were used to identify potential outliers (poor quality chips). This quality control check resulted in the elimination of one array from the analysis. Identification of differentially expressed miRNAs was carried out on the probe level as well as the miRNA level. The MAANOVA model included the sample identity as a random effect and the gene specific variance estimate (F1 test)

was used to test for differences between the controls and treated samples. In this analysis, parametric p-values were obtained and were then FDR corrected. All this website data are MIAME compliant and that the raw data have been deposited in a MIAME compliant database (GEO), as detailed on the MGED Society website http://www.mged.org/Workgroups/MIAME/miame.html.

For each sample, 1 μg total RNA (containing the small RNA fraction) was polyadenylated then converted to cDNA using an oligodT primer with a universal tag and miScript Reverse Transcription mix (The Qiagen miScript PCR system, Qiagen). Real-time PCR was performed in duplicate for each sample, using a primer complementary find more to the universal tag and a miScript primer (Qiagen) specific for each miRNA. PCR product was detected using SYBR Green and a CFX real-time detection system (Bio-Rad). Expression levels of miRNAs were normalized to expression levels of U6

snRNA. Statistical analysis of data was done by Student’s t-test. Approximately 800 ng of total RNA per sample (n = 5/group) was reverse transcribed using RT2 first Florfenicol strand kit (SABiosciences™). Reverse transcription and real-time PCRs were carried out using RT2 SYBR Green PCR Master Mix on 96-well PCR arrays designed for the evaluation of mouse T cell and B cell activation (SABiosciences™) and using a CFX real-time Detection System (BioRad). Threshold cycle values were averaged. Relative gene expression was determined according to the comparative Ct method and normalized to the Hprt and β-actin housekeeping genes. Fold changes were calculated using online PCR array data analysis software (SABiosciences™). Statistical significance was calculated using REST method ( Pfaffl et al., 2002). Statistically significant and differentially expressed (by both microarray and RT-PCR analyses) miRNA were further analysed for their functional implications in biological processes as described in Li et al. (2011). First, using TargetScan (Friedman et al., 2009 and Lewis et al., 2005), the predicted target genes of miR-150, miR-29b, miR-142-5p, miR-34c, miR-34b-5p and miR-122 were identified. TargetScan was specifically used because it is suggested to be more accurate than other available prediction software. Next, predicted targets that were also differentially expressed (p-value ≤ 0.05, fold change ± 1.