cummeRbund partitioning and heatmap

[Galgarad]

Hi guys,

I'm playing around with cummeRbund, and I have some questions I would like to ask you.

First, I would like to cluster a subset of genes and to plot them in a heatmap. I found on this forum how to create a list of genes, and how to draw a heatmap.

> myIDs<-read.table("text.txt")
> myIDs<-as.vector(myIDs$V1)
> myGenes<-getGenes(cuff,myIDs)

> h<-csHeatmap(myGenes,cluster='both')
> h



Now that gives a heatmap with the genes I want, but I somehow think that the clustering is not adequate. I tried to do the partitioning thing like so :

> ic<-csCluster(myGenes,k=4)

And then, I figured I could only draw a heatmap with the output of csCluster

> hic<-csHeatmap(ic)

And of course it doesn't work.

Do you guys know how I could plot a heatmap from the csCluster?

Thanks!

[Charitra]

did you get the answer?

Hi guys,

I'm playing around with cummeRbund, and I have some questions I would like to ask you.

First, I would like to cluster a subset of genes and to plot them in a heatmap. I found on this forum how to create a list of genes, and how to draw a heatmap.

> myIDs<-read.table("text.txt")
> myIDs<-as.vector(myIDs$V1)
> myGenes<-getGenes(cuff,myIDs)

> h<-csHeatmap(myGenes,cluster='both')
> h



Now that gives a heatmap with the genes I want, but I somehow think that the clustering is not adequate. I tried to do the partitioning thing like so :

> ic<-csCluster(myGenes,k=4)

And then, I figured I could only draw a heatmap with the output of csCluster

> hic<-csHeatmap(ic)

And of course it doesn't work.

Do you guys know how I could plot a heatmap from the csCluster?

Thanks!

[Galgarad]

No, I didn't.

However, I figured a way to do it in R, without cummeRbund. It is not elegant, but it works well! Let me know if you want to see it.

[Charitra]

Please tell me how you have managed to do. I need to make heatmap with -1 0 +1 range (not the log 10 RPKM value).
Can you reply please.
Thank you

No, I didn't.

However, I figured a way to do it in R, without cummeRbund. It is not elegant, but it works well! Let me know if you want to see it.

[Galgarad]

I will.

Are you familiar with R?

[Charitra]

Yes, I am using cummeRbund.

[anthurium]

Hello everybody,

I ran the cuffdiff analysis on 8 samples and I ran the csCluster command in cummeRbund in order to clusterize genes in my samples.
I would like to clusterize genes considering only a subset of my samples (ex. only four samples). Is there the possibility to select the samples for which I want to clusterize the genes?
thanks