How plotting wokrs in R

The plotting idea came to R when the computer displays were very low resolutions, the early 1990s. This time the only way to display scientific resolution graphic was the plotter machine. Which was basically a pen moved by two electronic motors in 2 axes. That’s why we have to work with devices and layers if we do plot in R.

Demo

Demonstration of graphics in R

demo("graphics")

Help page for main plot function

?plot

Basic Plots

Line charts

x=c(1:10)
plot(x)

y=c(x^3)
plot(x,y)

Labeling

plot(x,y, main="Title", xlab="X axis", ylab="Y axis")

Plot types

Detailed explanation: http://www.statmethods.net/graphs/line.html

type description
p points
l lines
o overplotted points and lines
b, c points (empty if “c”) joined by lines
s, S stair steps
h histogram-like vertical lines
n does not produce any points or lines 
plot(x,y, type="p")

plot(x,y, type="l")

plot(x,y, type="o")

plot(x,y, type="b")

Plotting symbols

Detailed description: http://www.statmethods.net/advgraphs/parameters.html

pch: Plotting CHaracters image: 

plot(x,y, type="b", pch="a")

plot(x,y, type="b", pch="*")

plot(x,y, type="b", pch=0)

plot(x,y, type="b", pch=23)

Line types

lty: Line TYpe

plot(x,y, type="b", pch=23, lty=2)

Text and Symbol Size

cex: Character EXtension

cex is a number indicating the amount by which plotting text and symbols should be scaled relative to the default. 1=default, 1.5 is 50% larger, 0.5 is 50% smaller, etc.

cex.axis: magnification of axis annotation relative to cex

cex.lab: magnification of x and y labels relative to cex

cex.main: magnification of titles relative to cex

plot(x,y, type="b", pch=23, cex=2, lty=2)

Colors

col: Default plotting color. Some functions (e.g. lines) accept a vector of values that are recycled.

col.axis: color for axis annotation

col.lab: color for x and y labels

col.main: color for titles

fg: plot foreground color (axes, boxes - also sets col= to same)

bg: plot background color

plot(x,y, type="b", pch=23, cex=2, col="red", lty=2, fg="gray", bg="yellow")

Overplotting

Create an empty plot and than add points and lines between points (copy the next command to the terminal line-by-line)

plot(NA, xlim=c(1,10), ylim=c(1,1000), xlab="x", ylab="y", main="Plot")
points(x,y, type="p", pch=23, cex=2, col="red", bg="yellow") # Add points
points(x,y, type="c", lty=2, col="blue") # Add line between points

Saving Graphs

Detailed description: http://www.statmethods.net/graphs/creating.html

You can save the graph in a variety of formats from the menu: File -> Save As.

You can also save the graph via code using one of the following functions.

Function Output format
pdf("mygraph.pdf") pdf file
win.metafile("mygraph.wmf") windows metafile
png("mygraph.png") png file
jpeg("mygraph.jpg") jpeg file
bmp("mygraph.bmp") bmp file
postscript("mygraph.ps") postscript file
svg("mygraph.svg") svg (Scalable Vector Graphics) file

Closing existing plot with:

dev.off()
## null device 
##           1
png("1_plot.png")
  plot(NA, xlim=c(1,10), ylim=c(1,1000), xlab="x", ylab="y", main="Plot")
  points(x,y, type="p", pch=23, cex=2, col="red", bg="yellow")
  points(x,y, type="c", lty=2, col="blue")
dev.off()
## png 
##   2
pdf("2_plots.pdf")
  plot(x,y, type="b", pch=23, cex=2, lty=2) # First plot to the 1st page
  plot(NA, xlim=c(1,10), ylim=c(1,1000), xlab="x", ylab="y", main="Plot") # Second plot to the 2nd page
  points(x,y, type="p", pch=23, cex=2, col="red", bg="yellow")
  points(x,y, type="c", lty=2, col="blue")
dev.off()
## png 
##   2

Different kind of plots

Import data

Read file with read counts

Counts=read.table(file="Dsim_count_table.tsv", header=T, row.names=1)

Keep those genes that were expressed in at a reasonable level (25 pairs) in all samples

K=Counts[rowSums(Counts>=25)==12, ]

Scatterplot matrix

The first 100 genes of all samples

plot(K[1:100,])

All genes of the first 6 samples

plot(K[, 1:6])

Histogram

Shows the distribution of variables in bins.

hist(K$C15_r1)

Histogram of logarithm of the raw counts.

L=log10(K)
hist(L$C15_r1)

Colored histogram with different number of bins.

hist(L$C15_r1, breaks=20, col="red")

colors: https://www.nceas.ucsb.edu/~frazier/RSpatialGuides/colorPaletteCheatsheet.pdf

Kernel density plots

Kernel density plots are usually a much more effective way to view the distribution of variables. It an alternative to histograms.

plot(density(L$C15_r1))

Overplot the density of two samples.

Cretae an empty plot and add lines.

plot(NA, xlim=c(1,6), ylim=c(0,1), xlab="Log10 counts", ylab="Density", main="Density plots")
lines(density(L$C15_r1))
lines(density(L$C15_r2))

Overplot the density of all samples.

plot(NA, xlim=c(1,6), ylim=c(0,1), xlab="Log10 counts", ylab="Density", main="Density plots")
for( i in colnames(L)) {
  lines(density(L[,i]))
}

Plot the density of all samples and use different colors.

# color vector
Color_4for12=rep(c("blue", "cyan", "orange", "red"), each=3)
Color_4for12
##  [1] "blue"   "blue"   "blue"   "cyan"   "cyan"   "cyan"   "orange"
##  [8] "orange" "orange" "red"    "red"    "red"
# plot
plot(NA, xlim=c(1,6), ylim=c(0,1), xlab="Log10 counts", ylab="Density", main="Density plots")
for( i in 1:12) {
  lines(density(L[,i]), col=Color_4for12[i])
}

Add legend to the plot.

Leg=c("C15", "C23", "W15", "W23")

plot(NA, xlim=c(1,6), ylim=c(0,1), xlab="Log10 counts", ylab="Density", main="Density plots")
for( i in 1:12) {
  lines(density(L[,i]), col=Color_4for12[i])
}
legend("topright", legend=Leg, lwd=2, col=c("blue", "cyan", "orange", "red"), bty="n")

Exercises

Task 1

Create a boxplot diagram from the counts in log10 scale. Colour it with the colour scheme used in density plots. Make the x-axes label to horizontal or 45° angle.

Task 2

Create a density plot from all of the samples in counts in log10 scale with colored lines: c15 | light blue c23 | bark blue w15 | light red w23 | dark red

You can try the RColorBrewer Paired palette as well.

Task 3

Try to find a solution to fill a colour the area under the curve of a density plot. Create a density plot for C15_r1 sample (log10 scale) with coloured under curve area.

Appendix

Creating nice colors with RColorbrewer library

library(RColorBrewer)

Show all the color schemes available.

display.brewer.all()

View a single RColorBrewer palette by specifying the number of colors and its name.

display.brewer.pal(8,"Set1")

Save their Hexa code to a vector.

Color_1=brewer.pal(8,"Set1")

Coloring point of a scatterplot according to their density

Plotting the log counts of C15_r1 (x axis) and C15_r2 (y axis).

plot(L[, 1:2])

Creating a vectors of blue colors

display.brewer.pal(9,"PuBu")[3:9]

## NULL

colorRampPalette interpolate a set of given colors to create new color ramps.

Col_blues=colorRampPalette(brewer.pal(9,"PuBu")[3:9])

densCols produces a vector containing colors which encode the local densities at each point in a scatterplot.

Col_dens=densCols(L$C15_r1, L$C15_r2, colramp=Col_blues)
plot(L[, 1:2], col=Col_dens)