[BioC] Comparison of murine and human microarray data

[Sylvain Brohée]

Hi Bas,

First of all, I would say that I would not recommend to to this kind of 
stuff as it seems kind of dirty to me.

However, recently, I was 'kindly' asked to perform this type of analyze 
and I used the good old ComBat function from the sva package to remove 
the batch effect between human and mouse. In order to have a reliable 
matrix from the beginning, I used only those genes that had the same 
gene names (in capital letters for mouse).

This is the code I used :

Let's eset be the mouse dataset and human.exp.names.agg be the human 
dataset. Genes are in rows and experiments in columns.

row.names(eset) <- toupper(row.names(eset))
human.mouse.complete <- merge(human.exp.names.agg, eset, by = 'row.names')
row.names(human.mouse.complete) <- human.mouse.complete[,1]
human.mouse.complete <- human.mouse.complete[,-1]
pheno.mod0 <- data.frame(row.names = names(human.mouse.complete), fact.1 
= rep(1, ncol(human.mouse.complete)), fact.2 =  rep(2, 
ncol(human.mouse.complete)))
mod0 <- model.matrix(~1, data = pheno.mod0)
human.mouse.complete.combat <- ComBat(human.mouse.complete, c(rep(1, 
human.exp.nb), c(rep(2,22))), mod = mod0)

It seemed to give satisfactory results.

If there are more "clever" ways, I would be happy to hear about them!

Cheers,

Sylvain


On 04/22/2014 04:04 PM, Bas van Gestel wrote:
> Dear all,
> For a project I would like to compare the gene expression of different immune cells in both mouse and human. For the immune cells of interest, microarray data is available. The microarray data for the human immune cells have been generated with the same platform. The microarray data for the murine immune cells have been generated with the same platform, although with a different platform than used for the human immune cells. I performed RMA normalization using the rma function in the affy package separately for the human and the murine datasets. However, I would like to compare the gene expression levels of mouse and human immune cells. I therefore would like to ask you the following questions:What is the recommended way to normalize the RMA normalized datasets of human and mouse, so that I can compare/combine both datasets?
> Thanks a lot for your help.
> Kind regards, Bas 		 	   		
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor