Fig. 6

diHLC-T outperform HepG2 cells in toxicometabolomic testing. A Unsupervised hierarchical clustering of samples and genes obtained from RNAseq data (see details in methods). B ROC curve with 95% confidence interval (shadowed) of the 10-times repeated threefold CV for (top) diHLC-T and (bottom) HepG2 cells. C RNAseq extracted dendrogram/heatmap (sample and gene-wise) of genes present in DETOX gene set (GOBP: detoxification). Genes significantly enriched (p < 0.05) in diHLC-T are highlighted in a red box, while those genes significantly enriched (p < 0.05) in HepG2 are highlighted in a green box (only ABCB6, top). D mRNA levels of genes involved in GSH metabolism (KEGG: glutathione_metabolism). diHLC-T data is extracted from this study. HepG2 values are extracted from GSE42643[8], GSE54066[12] and GSE112330[14]. E Intracellular levels of reduced glutathione (GSH) and S-adenosyl-L-methionine (SAM) in diHLC-T and HepG2 cells. diHLC-T, n = 6; HepG2, n = 3; *p < 0.05, **p < 0.01,***p < 0.005)