I was trying exactly the same thing. I am exploring GEO2R. I think I can do that now using the generated R-script.
Is there an automatic way to detect the groups within the sample?
Is there a way to download microarray experiments in GEO, which have gene names and the corresponding p-value and fold change (optional). I need to find significant genes in those random experiments.
It is not possible to get this information in any general way from NCBI GEO. You could use the online GEO2R system or you could use GEOquery and limma, for example, to access and then perform analyses of your choice from R and Bioconductor.
I was trying exactly the same thing. I am exploring GEO2R. I think I can do that now using the generated R-script.
Is there an automatic way to detect the groups within the sample?
No, there is not an automated way to detect groups within the sample. This is simply a function of the GEO data model which does not describe groups in a computable way.
Use GEO datasets (not series). Look in the soft files corresponding to GEO datasets. Here, the samples are described and grouped into subsets.
Log in to answer this question.