Fig 1.
Example of the exploration of batch effects.
Four plots generated by ComBat to correct for batch effects. For the left panel plots, the red lines are the parametric estimates, and the black lines are the kernel estimates for the distribution of effects across genes. The right panel shows Q-Q plots with the red line for the parametric estimate and the ordered batch effects for each gene (black points). The bottom plots show the analyses for the variances and the top plots refers to the means. Plots were generated for batches TSS E9 and E2 to avoid batches containing only one sample.
Table 1.
Information on molecular subtypes for TCGA cancer studies as provided by the TCGA_MolecularSubtype function.
Fig 2.
The workflow illustrates the steps and TCGAbiolinks functions to be used for case study 1 on TCGA-BRCA luminal subtypes.
Fig 3.
DEA analyses of TCGA-BRCA data comparing luminal subtypes with normal samples.
A-B) Volcano plots are shown where only those genes with logFC higher than 6 or lower than -6 are labelled and only the significant up- or down-regulated genes are shown as dots. We carried out DEA using the limma (A) or edgeR pipelines (B) of TCGAbiolinks. C) The correlation plot between the logFC estimated by the two pipelines for the top 500 DE genes is shown. The genes discussed in the main text are highlighted in bold. D) The intersect between all the DE genes estimated by the two pipelines is shown using UpSet.
Fig 4.
DE genes in uterine cancer compared to healthy uterine tissue samples.
A) The workflow illustrates the steps and TCGAbiolinks functions to be used for this case study. B-C) In the volcano plot, the up-regulated genes with logFC higher than 5 (B) or the down-regulated genes with logFC lower than -5 (C) are shown as a result of DEA carried out using the limma pipeline comparing primary tumor samples from TCGA-UCS and normal uterine tissue samples from GTEx.