8

Possible Duplicate:
How to overlay density plots in R?

I have recently started working with microarray datasets and am trying to get my hands on R. I wish to make some plots out of my result data, but however I am stuck at the following.

I have the following data (myData), 

cpg samp1 samp2 samp3 
cpg1 0.43 0.32 0.21 
cpg2 0.43 0.22 1.00 
cpg3 0.11 0.99 0.78 
cpg4 0.65 0.32 0.12 
cpg5 0.11 0.43 0.89

And I wish to obtain a density plot for this,

I did the following, 

plot (density(MyData$samp1), col="red") 
lines (density(MyData$samp2), col="green") 
lines (density(MyData$samp3), col="blue")

But doing this does not give me correct plots, because not all sample curves fit within the plot limits. I did try looking for answers, but honestly i am still not able to work this out. Can you help me know how do i set my scale for the above? Or what additional should I do to the above code, so that all the curves are in range?? I have got many samples, so i need a something that could also automatically assign a different colour curve for each of my sample, after scaling it right.

Thanks in advance..

Community
  • 1
  • 1
Letin
  • 1,255
  • 5
  • 20
  • 36
  • The answer to the question can be found in the comments to the accepted answer in the proposed duplicate. – Dason Oct 21 '12 at 06:24

3 Answers3

6

you will need to change the axis limits of the original plot to fit everything in. This is not done for you automatically when adding different lines:

e.g.

plot (density(MyData$samp1), col="red",ylim=c(0,3.5),xlim=c(-1,2)) 
lines (density(MyData$samp2), col="green") 
lines (density(MyData$samp3), col="blue")

enter image description here

user1317221_G
  • 15,087
  • 3
  • 52
  • 78
4

Using the lattice package (I'm sure a ggplot2 example isn't far behind), you could use the densityplot function:

library(lattice)
densityplot(~ samp1 + samp2 + samp3, data = myData, auto.key = TRUE)

enter image description here

Or use the reshape2 package to get your data into a tall form before plotting:

library(reshape2)

myDataM<-melt(myData)

densityplot(~ value, groups = variable, data = myDataM, auto.key = TRUE)
BenBarnes
  • 19,114
  • 6
  • 56
  • 74
2

You could also generate the limits of the x- and y-axis automatically:

ranges <-  apply(MyData, 2,
             function(x) { dens <- density(x); c(range(dens$x), range(dens$y)) })

plot(density(MyData$samp1), col="red",
       xlim = range(ranges[1:2, ]), ylim = range(ranges[3:4, ])) 
lines(density(MyData$samp2), col="green") 
lines(density(MyData$samp3), col="blue")

densities

Sven Hohenstein
  • 80,497
  • 17
  • 145
  • 168
  • Yes will work on it.. thanks to you as well for your help.. :) – Letin Sep 17 '12 at 09:57
  • 2
    A more automated approach (after calculating your ranges) might be something like: `plot(0, type = "n", xlim = range(ranges[1:2, ]), ylim = range(ranges[3:4, ])); COL = rainbow(ncol(MyData)); lapply(1:(ncol(MyData)), function(x) lines(density(MyData[, x]), col = COL[x]))`. In other words, create an empty plot, then use `lapply` to plot `lines` on this empty plot. – A5C1D2H2I1M1N2O1R2T1 Sep 17 '12 at 11:37