0

I have a file with genomic data for which I am trying to create a heatmap of the data together with a sidebar for the chromosome information. To make the heatmap I converted the numerical values from my file to create a a data matrix that I was then able to plot:

Heatmap <- heatmap.2(DataMatrix, trace="none", col=greenred, key=T, keysize=1.5, labRow=NA, density.info="none", Colv=FALSE)

However, I would like to add an additional color side bar to indicate the chromosome location of the said data. In particular I would like to have just two colors, say "black" if it's from "chrX" or "yellow" from any other chromosome. The column with this information is not in my data matrix and I'm not sure how to include this info into my heatmap. Any help would be greatly appreciated!

Gavin Simpson
  • 170,508
  • 25
  • 396
  • 453
user2165857
  • 2,530
  • 7
  • 27
  • 39

1 Answers1

5

The heatmap.2 function has a RowSideColors argument for this purpose, as you might infer from the documentation:

RowSideColors: (optional) character vector of length 'nrow(x)' containing the color names for a vertical side bar that may be used to annotate the rows of 'x'.

I've also recently discovered the aheatmap function in the NMF package, which makes nice heatmaps with an arbitrary number of annotation rows/columns on your heatmap. Included on that homepage you'll find several links to heatmap examples.

Steve Lianoglou
  • 7,183
  • 2
  • 25
  • 21
  • Thanks for the pointer to new heatmaps and packages. The `heatmap.2` code can be difficult to modify if there are no hooks of the sort needed for a particular task. And welcome to the Bay Area. – IRTFM Jun 24 '13 at 16:39
  • @SteveLianoglou - I saw the RowSideColors option but am not sure how to implement it into my code - will try the NMF package and see if it's easier to work with... – user2165857 Jun 24 '13 at 18:38
  • 2
    @user2165857: `RowSideColors` is a vector that has as many elements as `x` has rows. Each element in this vector is the color of the corresponding row of `x`. The colors in `RowSideColors` will get shuffled around with the rows of `x` as `x` is re-ordered via the clustering. – Steve Lianoglou Jun 25 '13 at 20:25
  • @SteveLianoglou: where did you get the information that x is automatically re-shuffled with the reordering? I can't find any information about this in the vignette. – tospo Mar 09 '15 at 15:03
  • @tospo: the documentation for `RowSideColors` says that these colors are used to * annotate the rows of x*. It may not explicitly say that the colors for the rows will follow the rows in the data as they are reordered due to clustering order, but it wouldn't be very helpful if it didn't ;-) – Steve Lianoglou Mar 10 '15 at 16:32