Fig. 2

Networks of candidate immune-related biomarkers during malaria, dengue or co-infection. Plasma levels of several immune-related (cytokines, chemokines) biomarkers were measured in malaria, dengue and co-infection subjects. Each connecting line represents a significant interaction (P < 0.05) detected by Spearman’s correlation test (a). All interactions had positive correlations. A heat map was designed to depict the overall pattern of expression of immune markers in the different outcomes by the median value of each parameter (b). A two-way hierarchical cluster analysis (Ward’s method) of immune molecules by clinical group was performed (b). Biomarkers that had the same median in the three groups were excluded from the heat map and cluster analysis. The colours shown for each symbol represent the fold variation from the median values calculated for each marker (a, b). The distribution of haemoglobin (HB), haematocrit (HT), platelets (PTL), aspartate aminotransferase (AST), alanine aminotransferase (ALT) in different clinical groups is shown in red symbols (medians and interquartile ranges) whereas the values for network densities are shown as black bars (c). The variation of HB, HT, PTL, AST, and ALT according to the groups was assessed using the Kruskal–Wallis test (***P < 0.001; **P < 0.01; *P < 0.05; ns = non-significant) (c). The five immune-related biomarkers with the highest number of interactions in all three groups were chosen (IFN-γ, IL-6, IL-13, TNF, and IL-12) and the relative number of interactions of these biomarkers was calculated according to each group (d). Dark grey rectangles represent the highest relative number of connections, light grey rectangles the medium relative number and white rectangles the lowest relative number of hits between molecules (d).