Skip to contents

Generate a linear regression scatterplot for two sets of sample counts.

Usage

plotBarcodeRegression(
  dgeObject,
  sample1 = NULL,
  sample2 = NULL,
  title = NULL,
  trendline = TRUE,
  trans = NULL,
  rug = FALSE
)

Arguments

dgeObject

DGEList object with barcode counts.

sample1

Name of sample 1 (string).

sample2

Name of sample 2 (string).

title

Optional, title of plot (string).

trendline

Include linear trendline using `stat_smooth()` (boolean). Default = `TRUE`.

trans

Optional, the name of a transformation object or the object itself.

rug

Include geom_rug density information on the axes (boolean). Defaule = `FALSE`.

Value

Returns a scatterplot of the two samples

Examples

data(test.dge)
plotBarcodeRegression(test.dge, sample1 = "T0-1", sample2 = "T0-2")