Skip to main content


Fig. 6 | BMC Microbiology

Fig. 6

From: Liver and blood cytokine microenvironment in HCV patients is associated to liver fibrosis score: a proinflammatory cytokine ensemble orchestrated by TNF and tuned by IL-10

Fig. 6

Computational bioinformatics analysis of serum cytokine levels in HCV patients according to the fibrosis score. Machine learning data mining was represented by heatmaps (a and b) and decision-trees (c and d) of z-score-normalized events. The heatmap computational method was applied to pre-process the serum cytokine data, in addition to identifying the attributes matching those across samples and clusters of individuals. a Serum cytokine attributes showed clear ability to cluster HCV patients, with up-regulated levels, apart from non-infected individuals, with basal levels of serum cytokines. b Heatmap analysis showed low performance to segregate the HCV patients according to their liver fibrosis scores. c Decision tree analysis provided the identification of “root” (TNF) and “secondary” (IL-10) attributes with high accuracy to categorize HCV patients aside from non-infected individuals. d Decision tree analysis showed moderate global performance of “root” (IFN-γ) and “secondary” (TNF) attributes to cluster HCV patients according to their liver fibrosis score

Back to article page