binomRegMethPredPlot: Plot the predicted methylation levels

In kaiqiong/SOMNiBUS: Smooth modeling of bisulfite sequencing

Plot the predicted methylation levels

Description

This function accepts the data.frame used as an input for the function binomRegMethModelPred with additional columns containing the predictions generated by the function binomRegMethModelPred and columns containing the name of each experimental group and returns a plot representing the predicted methylation levels according to each experimental group.

Usage

binomRegMethPredPlot(
  pred,
  pred.type = "proportion",
  pred.col = "pred",
  group.col = NULL,
  title = "Predicted methylation levels",
  style = NULL,
  save = NULL,
  verbose = TRUE
)

Arguments

pred	data.frame used as an input for the function binomRegMethModelPred (with columns 'Position', covariate names matching the original output from the function binomRegMethModel) with additional columns containing the predictions generated by the function binomRegMethModelPred and columns containing the name of each experimental group. Rows without a valid group name (empty character "" or NA) are ignored
pred.type	type of prediction returned by the function binomRegMethModelPred: proportion or link.scale. The default value is "proportion".
pred.col	character defines the name of the column containing the prediction values. The default value is "pred".
group.col	character defines the name of the column containing the experimental groups. If the group.col is set to NULL, the resulting plot will be a simple scatter plot representing all predicted values disregarding any experimental design. The default value is NULL.
title	the text for the title
style	named list containing the wanted style (color and line type) for each experimental groups. The first level list is named according each experimental group and for each experimental group there is a list containing the color and the type of the line. The line types should be among the following types: twodash, solid, longdash, dotted, dotdash, dashed, blank. The function accepts color name and its hexadecimal code. The default value is NULL meaning that the colors will be chosen randomly and the line style will be set to solid.
save	file name to create on disk. When the value is set to NULL, the plot is not saved. The default value is NULL.
verbose	logical indicates if the algorithm should provide progress report information. The default value is TRUE.

Value

This function prints out a plot of the predicted methylation levels according to preset experimental groups.

Author(s)

Audrey Lemaçon

Examples

#------------------------------------------------------------#
data(RAdat)
RAdat.f <- na.omit(RAdat[RAdat$Total_Counts != 0, ])
BEM.obj <- binomRegMethModel(
  data=RAdat.f, n.k=rep(5, 3), p0=0.003307034, p1=0.9,
  epsilon=10^(-6), epsilon.lambda=10^(-3), maxStep=200,
  Quasi = FALSE, RanEff = FALSE, verbose = FALSE
)
pos <- BEM.obj$uni.pos
newdata <- expand.grid(pos, c(0, 1), c(0, 1))
colnames(newdata) <- c("Position", "T_cell", "RA")
my.pred <- binomRegMethModelPred(BEM.obj, newdata, type = "link.scale", 
verbose = FALSE)
newdata$group <- ""
newdata[(newdata$RA == 0 & newdata$T_cell == 0),]$group <- "CTRL MONO"
newdata[(newdata$RA == 0 & newdata$T_cell == 1),]$group <- "CTRL TCELL"
newdata[(newdata$RA == 1 & newdata$T_cell == 0),]$group <- "RA MONO"
newdata[(newdata$RA == 1 & newdata$T_cell == 1),]$group <- "RA TCELL"
pred <- cbind(newdata, Pred = my.pred)
style <- list("CTRL MONO" = list(color = "blue", type = "dashed"),
"CTRL TCELL" = list(color = "green", type = "dashed"),
"RA MONO" = list(color = "blue", type = "solid"),
"RA TCELL" = list(color = "green", type = "solid"))
g <- binomRegMethPredPlot(pred, pred.col = "Pred", group.col = "group",
style = style, save = NULL, verbose = FALSE)

kaiqiong/SOMNiBUS documentation built on Feb. 24, 2023, 5:38 a.m.