Question: Preprocessing Of Microarray Datasets Derived From Different Platforms
1
gravatar for Biologist
8.4 years ago by
Biologist10
Biologist10 wrote:

Hello everybody,

I want to pre-process some microarray datasets and I would like to know how I can compare them if the datasets are from different platforms. Specifically, one of the datasets is the Hedenfalk et al. 2001 dataset in which the gene expression values are given as the ratio of Cy5/Cy3 intensities. The other datasets I have are of Affymetrix-type and I have been using RMA preprocessing (Bioconductor) for them. As far as I understand the RMA preprocessing outputs the expression values as log2 transformed values. Since I am new on the topic, could you explain me how I can compare (or what steps are further needed) the RMA processed data with the Hedenfalk data?

Thanks a lot,

microarray • 2.7k views
ADD COMMENTlink written 8.4 years ago by Biologist10
1
gravatar for Leonor Palmeira
8.4 years ago by
Leonor Palmeira3.7k
Li├Ęge, Belgium
Leonor Palmeira3.7k wrote:

What you are trying to do is what people usually call "meta-analyses" on microarray data from different experiments, and possibly different platforms.

One suggestion would be to combine p-values of DE genes from the two experiments, you can find more information on this in this paper: http://bioinformatics.oxfordjournals.org/content/25/20/2692.full and in its associated R package: metaMA.

ADD COMMENTlink modified 8.4 years ago • written 8.4 years ago by Leonor Palmeira3.7k
0
gravatar for Sean Davis
8.4 years ago by
Sean Davis26k
National Institutes of Health, Bethesda, MD
Sean Davis26k wrote:

http://www.biostars.org/post/show/1216/what-kinds-of-software-can-be-used-for-multi-platform-microarray-normalization-for-meta-analysis/ http://www.biostars.org/post/show/4418/combining-gene-expression-microarray-datasets/

As for the two-color issue, that may, indeed, be tricky. If the data all use a common reference, then you may be able to treat those data as single-channel data. If not, you may not be able to compare the datasets. In any case, what you are proposing is not straightforward and just because one can perform the analysis does not mean the results are going to be reliable.

ADD COMMENTlink written 8.4 years ago by Sean Davis26k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1765 users visited in the last hour