Extended Data Fig. 10: Trajectory analysis reveals the association of KLF2 and MYC with the quiescent and activated state, respectively, in HIV+ donor CD4 + T cells.
From: HIV infection reprogrammes CD4+ T cells for quiescence and entry into proviral latency
a. UMAP projection with the coloring scheme showing the clusters corresponding to the functional clustering marking different classes of activated and resting cells, which are derived using the leiden algorithm. b. Force Atlas projection of the same clusters (leiden clustering). c and d. Force Atlas projection with the clustering used for the trajectory analysis. The thin lines connecting cells (shown as dots) represent the deduced trajectories. The new clusters are marked by numbers, with the largest cluster (the one with the highest number of cells in it) being cluster zero and the smallest, cluster 10. Colors show the cells belonging to each cluster. e. Simplified representation of the clusters shown in panel d, as a graph plot. The numbers next to each node correspond to the number associated with the clusters in panel d. Node size correlates with the number of cells in the cluster represented by each node, with larger nodes corresponding to clusters with more cells. The nodes are colored to match the coloring scheme in c and d panels. f. Force Atlas projection, with the pseudotime calculations shown as a color transition. The color bar on the right shows the colors corresponding to the start of the trajectory (dark blue) and its end (red). g. The pseudotime superimposed on the linearized graph plot. h-m. Force Atlas projections showing the expression of markers of resting and activated CD4 + T cells on the trajectory. The color bar on the right of each panel shows the expression level. For matched graph plots, see Fig. 5m-r. n-q. Matched Force Atlas and graph plots depicting the expression patterns of TIGIT and CDKN1A. r. Graph plot with UMI counts per cell represented by color. As expected, nodes corresponding to activated cells (nodes 4, 9 and 10) have higher UMI counts, which generally correlates with the number of transcripts found in each cell. s. Neither of the clusters on the trajectory contain a high percentage of mitochondrial RNA, which is associated with dead/dying cells. t-w. Violin plots showing the expression level of key markers of resting and activated state in resting and activated CD4 + T cells from six HIV+ donors. The numbers shown are based on raw UMI counts. The use of raw, non-normalized counts is needed as transition between the resting and activated state results in changes in the size of the transcriptome. Violin plots display the distribution of expression values for each group; the outline depicts the probability density of the data at different values, red horizontal lines indicate the first and third quartiles, red circles denote the medians, and blue diamonds indicate the means. Comparisons between activated and resting cells for each transcript were performed using a two-sided Mann–Whitney U test with no adjustment for multiple comparisons. For KLF2, U = 135,959,975.5, rank-biserial correlation = 0.84, 95% CI [0.83, 0.85]; for CD40LG, U = 737,704,123.5, r = 0.79, 95% CI [0.78, 0.80]; for MYC, U = 447,742,560.5, r = 0.15, 95% CI [0.14, 0.16]; and for IL2, U = 461,422,717.5, r = 0.69, 95% CI [0.68, 0.70].
