Fig. 1

Comparative oxylipin metabolite profiling differentiates breast tumors according to receptor subtypes. a Heat map showing hierarchical clustering of paired tumor and normal mammary tissue specimens (N = 62) and oxylipin metabolites (N = 40) reveals that both tumor and normal samples are grouped into two distinct groups based on the Euclidian correlations. b Cross validated PLS-DA score and loading plots show the clustering of tumor and normal mammary tissues. Each biological replicate is represented by a single point. c Cross validated PLS-DA analysis comparing mammary tumors by hormone receptor status highlights the contribution of each oxylipin metabolite to the clustering in the score, loading and coefficients profile. Prediction matrix shows the classification results based on Fisher’s probability (91.9% correct assignment, P = 4.5E-23). d Total concentration of AA and LA-derived oxylipin metabolites from the CYP epoxygenases, CYP hydroxylases, sEH, cyclooxygenases, 5-LOX and 12/15-LOX catalytic pathways were quantified in the tumor and normal adjacent tissue specimens (N = 62). All analyses include 4 technical replicates per biological sample. Error bars indicate mean ± SEM. Significantly different values (P = 0.05, ANOVA) are denoted by letters