Skip to main content
. 2018 Aug 27;14(8):e7862. doi: 10.15252/msb.20177862

Figure 6. Differentially expressed host transcripts across timepoint and clinical variable contrasts in natural CHIKV infections.

Figure 6

Human whole‐blood transcriptomic signatures for timepoint, CHIKV viral titer at the acute phase, symptom severity, and convalescent (15 d post‐symptom onset) anti‐CHIKV antibody (Ab) titer across the two observed phases of natural CHIKV infection. Expression was measured with RNA‐seq and quantified at the transcript level, with all models of differential expression adjusting for patient age and gender as covariates. For all panels, 42 patients were sampled at 2 timepoints, each with 2 technical replicates; q‐values are Benjamini–Hochberg‐adjusted P‐values from a moderated paired t‐test under the mixed‐effects model.
  1. Volcano plot of differentially expressed host transcripts between acute‐ and convalescent‐phase samples, with negative log10‐scaled q‐values on the Y‐axis, and the fitted β coefficient for each transcript (corresponding to modeled log2 fold change) on the X‐axis. Transcripts to the right of the vertical dashed line were comparatively upregulated in acute‐phase samples, while transcripts to the left were upregulated during the convalescent phase. Transcripts that pass FDR < 0.05 and fold change > 2 are colored salmon and turquoise for acute‐associated and convalescent‐associated transcripts, respectively. Top transcripts by q‐value or fold change are individually labeled by their corresponding gene symbol.
  2. Heatmap of expression in units of Z‐scores per transcript for the top 50 differentially expressed host transcripts between acute‐ and convalescent‐phase samples. Clinical variables are depicted for all samples across the top of the heatmap; convalescent post‐symptom onset anti‐CHIKV antibody titer and viral titer (which was measured during the acute phase) are both in units of log10 dilutions. Hierarchical clustering (using complete linkage) was applied to both samples (X‐axis) and transcripts (Y‐axis). Two major clusters of samples (largely separating acute and convalescent samples) are highlighted.
  3. Volcano plot as in (A) but for differentially expressed host transcripts between samples with higher and lower CHIKV viral titer. Transcripts to the right of the vertical dashed line are associated with higher viral titer, while transcripts to the left are associated with lower viral titer. Transcripts significant at FDR < 0.05 are colored red.
  4. Volcano plot as in (C) but for differentially expressed host transcripts between patients with more severe and less severe acute‐phase symptoms. Transcripts to the right of the vertical dashed line were comparatively upregulated in more severe cases, while transcripts to the left were upregulated in less severe cases.
  5. Heatmap of expression as in (B) but for 20 differentially expressed host transcripts (all significant at FDR < 0.05) under a model that includes a severity–timepoint interaction. Transcripts have been filtered to those differentially expressed between the timepoints in more severe cases but not in less severe cases.
  6. Volcano plot as in (C) but for differentially expressed host transcripts between patients with higher and lower convalescent (15 d post‐symptom onset) anti‐CHIKV antibody (Ab) titers. Transcripts to the right of the vertical dashed line were comparatively upregulated in patients with a higher convalescent anti‐CHIKV antibody titer, while transcripts to the left were upregulated in patients with a lower convalescent anti‐CHIKV antibody titer.