Figure 2. Transcriptomic profile of mock-infected, uninfected bystander and HIV-1-infected cells.

A) Volcano plots from the limma analysis of the transcriptome of cells fractions described in Figure 1D, at 1% FDR and 1.7 fold-change cut-off (represented by lines). The log10 of the p-value adjusted for multiple comparisons (Benjamini and Hochberg) is used for the Y axis and the log2 of the fold change is used for the X axis. Left columns: comparison between mock-infected versus uninfected bystander cell populations shows no statistically significant genes passing both thresholds at any time point or in aggregate analysis (labelled “all”). Center columns: comparison between mock-infected and virus-infected cells. Right columns: comparison between uninfected bystander and HIV-1-infected cells. B) Proportional Venn diagrams of above data, illustrating the overlap of DEGs at 24, 48 and 72 h for mock vs infected and bystander vs infected comparisons. C) Unsupervised hierarchical clustering (per sample and per gene) of differentially-expressed genes in the dataset. The similarity between mock-infected and uninfected bystander cells is obvious. Moreover, HIV-1-infected cell samples cluster together at the 24 h time point, but diverge on a per-sample basis at 48 and 72 h, illustrating that, although a common subset of T cells are infected early on, the precise context preferred by the virus can evolve differently over time on a per-sample basis.