curvesplot: Plot of fitted curves

Description

Provides a plot of all the fitted curves from a dataframe of the main workflow results, possibly extended with additional information (e.g. groups from functional annotation) used to color and/or split the curves.

Usage

curvesplot(extendedres, xmin, xmax, y0shift = TRUE, scaling = TRUE,
                       facetby, facetby2, free.y.scales = FALSE, ncol4faceting,
                       colorby, removelegend = FALSE,  
                        npoints = 500, line.size = 0.5, 
                        line.alpha = 0.8, dose_log_transfo = TRUE,
                       addBMD = TRUE, BMDtype = c("zSD", "xfold"), 
                       point.size = 1, point.alpha = 0.8)

Value

a ggplot object.

Arguments

extendedres: the dataframe of results provided by bmdcalc (res) or a subset of this data frame (selected lines). This dataframe can be extended with additional columns coming for example from the annotation of items, and some lines can be replicated if their corresponding item has more than one annotation. This extended dataframe must at least contain the column giving the identification of each curve (id), the column model naming the fitted model and the values of the parameters (columns b, c, d, e, f) and a column coding for the chosen BMD, by default BMD.zSD or BMD.xfold BMDtype is to "xfold".
xmin: If not defined, a value just below the maximum BMD values is fixed if the x dose scale is in log, or 0 otherwise.
xmax: Maximal dose/concentration for definition of the x range (can be defined as max(f$omicdata$dose) with f the output of drcfit()). If not defined, a value just above the maximum BMD values is taken.
y0shift: If TRUE (default choice) curves are all shifted to have the theoretical signal at the control at 0.
scaling: If TRUE, default choice, curves are all shifted to have the theoretical signal at the control at 0 y0 and scaled by dividing by the maximal absolute signal change (up or down) from the signal at the control maxychange.
facetby: optional argument naming the column of extendedres chosen to split the plot in facets (no split if omitted).
facetby2: optional argument naming the column of extendedres chosen as an additional argument to split the plot in facets using ggplot2::facet_grid, with columns defined by facetby and rows defined by facetby2 (no split if omitted).
free.y.scales: if TRUE the y scales are free in the different facets.
ncol4faceting: Number of columns for facetting (not used if facetby2 is also provided.
colorby: optional argument naming the column of extendedres chosen to color the curves (no color if omitted).
removelegend: If TRUE the color legend is removed (useful if the number of colors is great).
npoints: Number of points computed on each curve to plot it.
line.size: Width of the lines for plotting curves.
line.alpha: Transparency of the lines for plotting curves.
dose_log_transfo: If TRUE a log transformation of the dose is used in the plot. This option needs a definition of a strictly positive value of xmin in input.
addBMD: If TRUE points are added on the curve at BMD-BMR values (requires to have BMD and BMD values in the first argument extendedres).
BMDtype: The type of BMD to add, "zSD" (default choice) or "xfold".
point.size: Size of the BMD-BMR points added on the curves.
point.alpha: Transparency of the BMD-BMR points added on the curves.

Author

Marie-Laure Delignette-Muller

Details

For each item of the extended dataframe, the name of the model (column model) and the values of the parameters (columns b, c, d, e, f) are used to compute theoretical dose-response curves in the range [xmin ; xmax].

Examples

Run this code


# (1) A toy example on a very small subsample of a microarray data set) 
#
datafilename <- system.file("extdata", "transcripto_very_small_sample.txt", 
  package = "DRomics")

o <- microarraydata(datafilename, check = TRUE, norm.method = "cyclicloess")
s_quad <- itemselect(o, select.method = "quadratic", FDR = 0.01)
f <- drcfit(s_quad, progressbar = TRUE)
r <- bmdcalc(f)

# (1.a) 
# Default plot of all the curves with BMD values added as points on the curve
#
curvesplot(r$res, xmax = max(f$omicdata$dose))

# \donttest{
# use of line size, point size, transparency
curvesplot(r$res, xmax = max(f$omicdata$dose), 
  line.alpha = 0.2, line.size = 1, point.alpha = 0.3, point.size = 1.8)

# the same plot with dose not in log scale
# fixing xmin and xmax
curvesplot(r$res, xmin = 0.1, xmax = max(f$omicdata$dose),
  dose_log_transfo = FALSE, addBMD = TRUE)
# or not
curvesplot(r$res, dose_log_transfo = FALSE, addBMD = TRUE)

# plot of curves colored by models
curvesplot(r$res, xmax = max(f$omicdata$dose), colorby = "model")

# plot of curves facetted by item
curvesplot(r$res, xmax = max(f$omicdata$dose), facetby = "id")

# plot of curves facetted by trends
curvesplot(r$res, xmax = max(f$omicdata$dose), facetby = "trend")

# the same plot with free y scales
curvesplot(r$res, xmax = max(f$omicdata$dose), facetby = "trend",
  free.y.scales =  TRUE)
  
# (1.b) 
# Plot of all the curves without shifting y0 values to 0
# and without scaling
curvesplot(r$res, xmax = max(f$omicdata$dose),  
  scaling = FALSE, y0shift = FALSE)

# (1.c) 
# Plot of all the curves colored by model, with one facet per trend
#
curvesplot(r$res, xmax = max(f$omicdata$dose), 
  facetby = "trend", colorby = "model")

# changing the number of columns
curvesplot(r$res, xmax = max(f$omicdata$dose), 
  facetby = "trend", colorby = "model", ncol4faceting = 4)

# playing with size and transparency of lines
curvesplot(r$res, xmax = max(f$omicdata$dose), 
  facetby = "trend", colorby = "model", 
  line.size = 0.5, line.alpha = 0.8)
curvesplot(r$res, xmax = max(f$omicdata$dose),  
  facetby = "trend", colorby = "model", 
  line.size = 0.8, line.alpha = 0.2)
curvesplot(r$res, xmax = max(f$omicdata$dose),  
  facetby = "trend", line.size = 1, line.alpha = 0.2)
  
# (2) an example on a microarray data set (a subsample of a greater data set)
#
datafilename <- system.file("extdata", "transcripto_sample.txt", package="DRomics")

(o <- microarraydata(datafilename, check = TRUE, norm.method = "cyclicloess"))
(s_quad <- itemselect(o, select.method = "quadratic", FDR = 0.001))
(f <- drcfit(s_quad, progressbar = TRUE))
(r <- bmdcalc(f))

# plot split by trend and model with BMR-BMD points added on curves
# adding transparency
curvesplot(r$res, xmax = max(f$omicdata$dose), 
  line.alpha = 0.2, line.size = 0.8,
  addBMD = TRUE, point.alpha = 0.2, point.size = 1.5,
  facetby = "trend", facetby2 = "model")

# same plot without scaling and not in log dose scale
curvesplot(r$res, xmax = max(f$omicdata$dose), 
  line.alpha = 0.2, line.size = 0.8, dose_log_transfo = FALSE,
  addBMD = TRUE, point.alpha = 0.2, point.size = 1.5,
  scaling = FALSE, facetby = "trend", facetby2 = "model")

# (3) An example from data published by Larras et al. 2020
# in Journal of Hazardous Materials
# https://doi.org/10.1016/j.jhazmat.2020.122727

# a dataframe with metabolomic results (output $res of bmdcalc() or bmdboot() functions)
resfilename <- system.file("extdata", "triclosanSVmetabres.txt", package="DRomics")
res <- read.table(resfilename, header = TRUE, stringsAsFactors = TRUE)
str(res)

# a dataframe with annotation of each item identified in the previous file
# each item may have more than one annotation (-> more than one line)
annotfilename <- system.file("extdata", "triclosanSVmetabannot.txt", package="DRomics")
annot <- read.table(annotfilename, header = TRUE, stringsAsFactors = TRUE)
str(annot)

# Merging of both previous dataframes
# in order to obtain an extenderes dataframe
# bootstrap results and annotation
extendedres <- merge(x = res, y = annot, by.x = "id", by.y = "metab.code")
head(extendedres)

# Plot of the dose-response curves by pathway colored by trend
# with BMR-BMD points added on curves
curvesplot(extendedres, facetby = "path_class", npoints = 100, line.size = 0.5,
  colorby = "trend", xmax = 7, addBMD = TRUE) 

# The same plot not in log scale
curvesplot(extendedres, facetby = "path_class", npoints = 100, line.size = 0.5,
  dose_log_transfo = FALSE, colorby = "trend", xmin = 0, xmax = 7) 

# The same plot in log scale without scaling
curvesplot(extendedres, facetby = "path_class", npoints = 100, line.size = 0.5,
  colorby = "trend", scaling = FALSE, xmax = 7) 

# Plot of the dose-response curves split by pathway and by trend
# for a selection pathway
chosen_path_class <- c("Membrane transport", "Lipid metabolism")
ischosen <- is.element(extendedres$path_class, chosen_path_class)
curvesplot(extendedres[ischosen, ],
  facetby = "trend", facetby2 = "path_class",
  npoints = 100, line.size = 0.5, xmax = 7) 

# Plot of the dose-response curves for a specific pathway
# in this example the "lipid metabolism" pathclass
LMres <- extendedres[extendedres$path_class == "Lipid metabolism", ]
curvesplot(LMres, facetby = "id", npoints = 100, line.size = 0.8, point.size = 2,
  colorby = "trend", xmax = 7) 
# }

Run the code above in your browser using DataLab