RNA-sequencing, QC, and analyses

RNA quality and concentration were assessed using an Agilent 2100 bioanalyzer. RIN scores were 7.8 or better for all samples, with a mean RIN of 9.6. RNA was converted to cDNA using the SMART-Seq v4 Ultra Low Input RNA kit for sequencing (Takara Bio cat. 634898). Standard library preparation was completed using the Illumina Nextera XT DNA Library Preparation Kit and library quality and concentration were assessed using an Agilent 2100 bioanalyzer. Indexed samples were pooled and sequenced on the Illumina Hi-Seq 4000 with 100 base-pair, paired-end sequencing to a minimum of 10 million mapped reads per sample and a mean of 14 million reads per sample. RNA-seq reads were mapped to genome assembly GRCh37(hg19) reference sequence using STAR (version 2.6.1)⁵⁴. Read counts were adjusted to counts per million and normalized using TMM normalization⁵⁵. PCA was conducted in R (version 3.6.2) using prcomp to select covariates to adjust for in the model; library prep technician, cDNA concentration, cDNA prep date, library concentration, and culture plate orientation were significant and adjusted for in the model. Voom was used to adjust for significant technical covariates⁵⁶. Each sample was studied in triplicate (biological replicate), which was statistically accounted for using duplicateCorrelation in the Limma R package version 3.5 (https://bioconductor.org/packages/release/bioc/html/limma.html). Genes significantly associated with RV treatment or asthma were identified using Limma linear mixed modeling in R, while including sex as a covariate and blocking by sample to pair RV and vehicle treated samples from the same individual. An FDR-adjusted p value of 0.05 was used as a significance threshold, using the method of Benjamini and Hochberg⁵⁷.