Hi All, I am trying to represent my data as a heatmap without scaling it, in such a way that the genes that are upregulated (respect to the control) are "green" and those that are downregulated (respect to the control) are "red". In order to do this, I normalized the data respect to the control, now the upregulated genes have a value >1, and the downregulated genes have a value <1. However, if I do a simple heatmap.2 (defining: scale = "none"), most of the genes show the same color, and this is because the data is not centered, and its distribution is skewed to the right. But I would like to capture the differences on gene expression without removing the higher values, and without centering the data. So, my guess is that I have to modify the color key, but after trying different things and couldn't find a way to do it. Could you please give me some clues? I appreciate any help! Avhena [[alternative HTML version deleted]]

Hi All, After trying different things, I think it works if instead of plotting the raw data, I apply the *log* function to the data normalized by the control condition... Do you have any other suggestion? Thank you! Avhena On Tue, May 31, 2011 at 1:30 AM, avehna <avhena@gmail.com> wrote: > Hi All, > > I am trying to represent my data as a heatmap without scaling it, in such a > way that the genes that are upregulated (respect to the control) are "green" > and those that are downregulated (respect to the control) are "red". > > In order to do this, I normalized the data respect to the control, now the > upregulated genes have a value >1, and the downregulated genes have a value > <1. However, if I do a simple heatmap.2 (defining: scale = "none"), most of > the genes show the same color, and this is because the data is not centered, > and its distribution is skewed to the right. > > But I would like to capture the differences on gene expression without > removing the higher values, and without centering the data. So, my guess is > that I have to modify the color key, but after trying different things and > couldn't find a way to do it. > > Could you please give me some clues? > > I appreciate any help! > > Avhena > [[alternative HTML version deleted]]