Fig. 5: SARS-CoV-2 infection during pregnancy induces dysregulation of retrotransposons.

a, Dysregulated retrotransposon subfamilies from the bulk RNA-seq. Subfamilies that were significantly upregulated (Padj < 0.05 and log₂(fold change) > 0.25; n = 4) or downregulated (Padj < 0.05 and log₂(fold change) < −0.25; n = 38) are labelled in red and blue, respectively. b, Expression of retrotransposon subfamilies in the 21 cell types from the snRNA-seq. c, GREAT analysis of significantly downregulated individual retrotransposons from the bulk RNA-seq analysis. d, Genome browser screenshot of ERV3-1. The pseudo-bulk tracks from the snRNA-seq and snATAC are displayed as RPM with y axes ranging from zero to eight. The CUT&Tag H3K27ac tracks are displayed as aggregated RPM values with y axes ranging from zero to four. Blue shadings highlight the candidate enhancers defined by peak-to-gene analysis. e, Retrotransposon subfamilies with differential chromatin accessibility from the bulk ATAC-seq. Retrotransposon subfamilies with significantly increased (Padj < 0.05 and log₂(fold change) > 0.25; n = 44) or decreased (Padj < 0.05 and log₂(fold change) < −0.25; n = 44) accessibility are labelled in red and blue, respectively. a,e, The negative log₁₀(two-tailed Wald test Padj) and log₂(fold change, patient/control) values for each subfamily were calculated by DESeq2. Dashed lines represent the indicated thresholds. f, Heatmap showing chromatin accessibility of retrotransposon subfamilies across the nine cell types from the snATAC-seq. b,f, Each row is a retrotransposon subfamily and the colour scale represents the RPM-calculated row z-score. g, GREAT analysis of individual retrotransposons with decreased chromatin accessibility from the bulk ATAC-seq. c,g, The top relevant GO terms are shown. Ctrl, control and Cov, COVID-19; cell-type abbreviations as per Fig. 1b.