Cross-platform of microarray or analysis separately

Question

[Morteza Hadizadeh]

Dear all,

I want to find common genes for 4 different diseases, I have 4 microarray datasets with the same platforms (Affymetrix). Which is the best way for the determination of common genes? I merge all datasets and remove batch effects or analysis each dataset individually and then detect common genes?

I would be very pleased if you could guide me.

[ATpoint]

You can only remove batch effects if you have replicates of the samples in each batch (so each study) which you don't if I get this right. Therefore option two is what you should do, followed by some kind of meta-analysis.

[Hamid Ghaedi]

The second option is more robust here. you can undertake a ranked-based meta-analysis approach like This package do was intorduced to resolve similar issues.