ggplot2_TPR_FPRPlot: Plot TPR and FPR of each combination

In zshao1/ERSSA: Empirical RNA-seq Sample Size Analysis

Plot TPR and FPR of each combination

Description

ggplot2_TPR_FPR function uses the full dataset list of DE genes as the ground truth to calculate the True Positive Rate (TPR) and False Positive Rate (FPR) for each sample combinations tested. The TPR and FPR are then plotted with FPR on x-axis and TPR on y-axis similar to a ROC curve.

Usage

ggplot2_TPR_FPRPlot(
  deg = NULL,
  count_table.filtered = NULL,
  stat = "mean",
  path = ".",
  save_plot = TRUE
)

Arguments

deg	The list of DE genes generated by one of ERSSA::DE_*.R scripts.
count_table.filtered	The filtered count table with non- and low-expression genes removed. Used to identify the genes found to be non-DE.
stat	The statistics used to summarize TPR and FPR at each replicate level. Options include 'mean' and 'median'. Default = 'mean'.
path	Path to which the plot will be saved. Default to current working directory.
save_plot	Boolean. Whether to save plot to drive. Default to TRUE.

Details

Using the list of DE genes generated from the full dataset as the ground truth should be done with caution. Since the true list of DE genes is not known, this is the best alternative. This plot enables the visualization of the sensitivity and (1-specificity) of the DE gene detection at the tested replicate levels. At a sufficient replicate level, a relatively high TPR can be reached with reasonable low FPR. Such a replicate level is sufficient for most studies as additional replicates produce little improvement in TPR.

Value

A list is returned containing:

• gg_object the ggplot2 object, which can then be further customized.

• TPR_FPR.dataframe the tidy table used for plotting.

• list_TP_FP_genes lists of TP and FP genes. Follow the format of deg object with each comb_n now as a list contain two vectors, one for each of TP and FP list of DE genes

Author(s)

Zixuan Shao, [e-mail]

References

H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York, 2009.

Examples

# load edgeR deg object generated by erssa_edger using example dataset
# example dataset containing 1000 genes, 4 replicates and 5 comb. per rep.
# level
data(deg.partial, package = "ERSSA")
data(count_table.filtered.partial, package = "ERSSA")

gg_TPR_FPR = ggplot2_TPR_FPRPlot(deg.partial, count_table.filtered.partial)

zshao1/ERSSA documentation built on July 19, 2023, 9:20 p.m.