[BioC] cluster a set of genes based on another set of genes

Sean Davis sdavis2 at mail.nih.gov
Mon Feb 12 12:02:57 CET 2007


On Monday 12 February 2007 05:42, Marcelo Laia wrote:
> Hi,
>
>    Suppose this situation:
>    There is a structure in a cell. This one is composed by five
> proteins. All that is essential. But, in that organism only three are
> known.
>
>     If I have the expression for all genes, is it possible to cluster
> all expression in relation to expression of this three genes?
>
>     I am looking if any of that genes group together this three. I can
> do a cluster with all genes, but I would like to cluster one group
> knowing that another group exist.
>
>     What you mean about my guess? You known a manner to do this? There
> is BioConductor package to do this?
>
>     My normalized data are in a limma MA object.

It sounds like you want to find genes that are similar in expression pattern 
to your three known genes.  You might look at using a correlation or a 
distance metric between your known genes and all other genes on your array.  
The cor() function computes correlation.  

Sean



More information about the Bioconductor mailing list