Fig. 4

In-silico analyses of KCNH2, SCN5A, ITGB1 and SLC9A1 in BCa samples. A) Joint expression correlation plots for KCNH2, SCN5A, ITGB1 and SLC9A1 in BCa (green dots and curves), healthy breast (blue dots and curves) and human heart (orange dots and curves). The curves on the diagonal panels represent plots of kernel density distributions of expression. B) Contour plot showing KCNH2 and SCN5A expression in BCa (orange curves) and healthy breast (blue curves). Human heart (orange curves) is also reported as a reference since it is well known that in this tissue both channels are highly expressed and functionally associated. C) Variable-subgroup correlation analysis (see text for details) between KCNH2 and SCN5A in healthy breast (blue curves) and BCa (green curves). In healthy breast no correlation is evident, while in BCa patients the expression of the two genes follows a similar trend, displaying a consistently increasing correlation in higher and higher-expression subgroups, reaching a maximum quasi-steady correlation value of 0.3 in a relatively wide region of expression. An almost identical result is observed using both Pearson and Spearman correlation tests. Data are reported as fpkm-uq. D) Correlation analysis of KCNH2 and SCN5A in BCa datasets from TCGA according to KCNH2 expression levels. The threshold was set at the median value. E) Correlation analysis of KCNH2 and SCN5A in BCa datasets from TCGA according to SCN5A expression levels. The threshold was set at the median value. The analyses were performed with Breast Cancer Gene-Expression Miner v5.0 (bc-GenExMiner v5.0)