Hi all --
I've data from 12 affy arrays, 5 in one class, 7 in the other. I am
looking for genes that classify a sample into one or the other
classes.
One thing I'd like to do is to be able to visually evaluate the power
that exists to distinguish the classes, using clustered heatmaps. I've
done some of this in the dchip program and NCI's BRB arraytools, but I
haven't found an appropriate vignette or lab/tutorial example in the R
canon.
I am guessing that I need to filter the genes to those that are
differentially expressed, then use the 'heatmap' function. One thing I
can't see immediately is how to identify the different classes in the
call to heatmap...
I'm certain that I'm not the first to tackle this, can someone share
their experience?
Thanks in advance!
Marc
--
Dr. Marc Salit
Research Chemist
NIST
Chemical Science and Technology Laboratory
100 Bureau Drive, Stop 8391
Gaithersburg, MD 20899-8391
email: Salit@NIST.gov
voice: +1 301 975-3646
fax: +1 301 869-0413
Marc,
The heatmap function does the clustering from the data. In other
words, you
call heatmap with your matrix of gene expression values and heatmap
clusters
the samples and genes. You can optionally ask heatmap to not cluster
either
rows (genes) or columns (samples).
Sean
On 7/2/04 9:45 AM, "Marc Salit" <salit@nist.gov> wrote:
> Hi all --
>
> I've data from 12 affy arrays, 5 in one class, 7 in the other. I am
> looking for genes that classify a sample into one or the other
classes.
>
> One thing I'd like to do is to be able to visually evaluate the
power
> that exists to distinguish the classes, using clustered heatmaps.
I've
> done some of this in the dchip program and NCI's BRB arraytools, but
I
> haven't found an appropriate vignette or lab/tutorial example in the
R
> canon.
>
> I am guessing that I need to filter the genes to those that are
> differentially expressed, then use the 'heatmap' function. One thing
I
> can't see immediately is how to identify the different classes in
the
> call to heatmap...
>
> I'm certain that I'm not the first to tackle this, can someone share
> their experience?
>
> Thanks in advance!
> Marc
Hi Marc,
I've used the 'heatmap' function to display differential genes. I
filtered my data first using simpleaffy, and then created a heatmap as
in the example below:
>exprs.eset <- exprs(eset) #to get expression
values from normalised expression set
>heatmapgenes <- exprs.eset[filtered,] #to get just the values for my
differential genes
>samplenames <- as.vector(eset$class) #to get the names of the
classes
into a vector
>colnames(heatmapgenes) <- samplenames #this allows you to label your
samples by class in the heatmap
>heatmap(heatmapgenes)
If you alter the 'samplenames' vector, you can alter the labels on the
heatmap. You can also do this with the argument 'labCol'.
Hope this helps,
Laura
.......
Laura Hollins
PhD Student
Bioinformatics Group
Paterson Institute for Cancer Research
Hi all --
I've data from 12 affy arrays, 5 in one class, 7 in the other. I am
looking for genes that classify a sample into one or the other
classes.
One thing I'd like to do is to be able to visually evaluate the power
that exists to distinguish the classes, using clustered heatmaps. I've
done some of this in the dchip program and NCI's BRB arraytools, but I
haven't found an appropriate vignette or lab/tutorial example in the R
canon.
I am guessing that I need to filter the genes to those that are
differentially expressed, then use the 'heatmap' function. One thing I
can't see immediately is how to identify the different classes in the
call to heatmap...
I'm certain that I'm not the first to tackle this, can someone share
their experience?
Thanks in advance!
Marc
--
Dr. Marc Salit
Research Chemist
NIST
Chemical Science and Technology Laboratory
100 Bureau Drive, Stop 8391
Gaithersburg, MD 20899-8391
email: Salit at NIST.gov
voice: +1 301 975-3646
fax: +1 301 869-0413
--------------------------------------------------------
This email is confidential and intended solely for the use
o...{{dropped}}
