Figure 3

Hierarchical clustering analysis (HCA) of liver metabolites. HCA with predefined sample clusters, was performed for the 102 metabolites that were significantly different among groups as judged by ANOVA (without post hoc testing) to determine trends for individual metabolites among experimental groups. Each column cluster represents an experimental cohort as labeled (Normal, 4-h sham, 4-h AKI, 24-h sham, 24-h AKI, 7-day sham and 7-day AKI), whereas each individual column represents an experimental mouse subject (Normal: n = 10, 2 4 h sham: n = 10, 4 h AKI: n = 10, 24 h sham: n = 10, 24 h AKI: n = 10, 7 day sham: n = 9, and 7 day AKI: n = 6). Each row represents a metabolite. To the right of each row is the corresponding metabolite name. Red represents increased metabolite level whereas blue represents decreased metabolite level. The correlation metric used was the Pearson (n-1) correlation with an average linkage method. The major effect observed is for the 24-h AKI group which is characterized by metabolites that are both relatively decreased versus other groups (in blue, upper 2/3 of heat map) as well as increased levels versus other groups (in red, lower 1/3 of heat map).