Question: How To Identify Tissue Specific Genes From Mixed Samples In Microarray Data
gravatar for gundalav
5.8 years ago by
La La Land
gundalav270 wrote:

I have a microarray data which were derived from samples from mouse tissue (liver). We have intention in analyzing the liver specific genes. However since liver are located close to other tissue (ab.adipose), we highly suspect that some expressed genes are not related to liver.

What are standard ways to deal with this problem? Is there any tissue specific gene database?

ADD COMMENTlink modified 5.8 years ago by pld4.8k • written 5.8 years ago by gundalav270

I am not sure about the standard way but you should be able to get a list of highly expressed genes in adipose tissue and use this list to remove contaminant gene expression. BioGps should be one of the place that can help.

ADD REPLYlink written 5.8 years ago by Ashutosh Pandey11k
gravatar for pld
5.8 years ago by
United States
pld4.8k wrote:

It isn't possible to tell from the resulting data where it came from, the resulting probe intensity values provide no information as to the source of the RNA. Even if you find some data somewhere showing gene X is only present in adipose tissue, you have no way of demonstrating that the same phenomenon is present in your case. As Sean said, microarrays are about differential expression not absolute expression. What you'll end up subtracting is the difference between two conditions, which doesn't make much sense. You seem to be more interested in throwing out strong genes that don't fit your expectation of the experiment, not simply filtering noise, a reviewer will catch this.

Either the livers weren't cleaned of remaining fat and tissue before processing, or those genes really did come from the liver. I am not sure I would try open pandora's box: if gene x is contamination, how do we know for sure gene y isn't?

If you have any remaining tissue you can do qPCR on the adipose/liver tissue to show the source of that gene's signal.

ADD COMMENTlink written 5.8 years ago by pld4.8k
gravatar for Sean Davis
5.8 years ago by
Sean Davis25k
National Institutes of Health, Bethesda, MD
Sean Davis25k wrote:

Expressed genes are not an issue, as microarrays are not very useful for absolute expression quantification, in general. Most folks are looking for differential expression between two conditions/groups, so having some tissue admixture may not be so important. Before trying to get too fancy, it might be useful to simply move ahead with differential expression analysis and examine the results.

ADD COMMENTlink written 5.8 years ago by Sean Davis25k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 836 users visited in the last hour