normalizing CEL files from Affymetrix Human Genome U133 Plus 2.0 Array and Affymetrix Human Exon 1.0 ST Array [transcript (gene) version]

Question

[zizigolu]

Hi,

I found a collection of cancerous samples from Affymetrix Human Genome U133 Plus 2.0 Array (GSE2109) and some normal samples of Affymetrix Human Exon 1.0 ST Array (GSE19163) . does it make sense if I normalize these CEL files together because I am going to perform t-test between normal and cancerous samples while they from different platforms.

thank you for any comments

[Shamim Sarhadi]

Take a look at incilicodb , I think you could retrieve your GSEs from insilicodb and then merging them with insilicomerging package(there are some methods for merging cross platform data and remove batch to batch variation like COMBAT ), if your dataset does not exist in insilicodb you could request to insert them,then try with limma for DEG analysis

Answer 1

[ATRX]

The platform used for GSE2109 is GPL570/ [HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array and the platform used for GSE 19163 is GPL5175/ [HuEx-1_0-st] Affymetrix Human Exon 1.0 ST Array [transcript (gene) version. Take a look at the probe ids and so will realize that they are different and therefore, you cannot normalize them together. What you may do is:

1. Normalize them separately
2. Map the ids to genes symbols
3. Combine the dataset (do check if the number of probes that map to gene symbols are around the same range)
4. Do limma or t-test

Be cautious about the batch effects as these samples come from different experiment. Do check the amount of cdna used and if the processing of samples are consistent or not.