GEO study selection for analysis in R

Question

[786]

Hi I just want to ask that when we are going to analyze micro-array data in R then can we consider those studies in which disease and control states of samples are not defined?? Please help me out it'll be very helpful.

[GZ1995]

Could you please post the GEO number where the state is not defined? Because generally it is a required information for GEO data sets.

[786]

GSE54279 in this study control and diabetic aren't taken properly. How to find the groups of this series on the basis of HbA1c levels??

[GZ1995]

I'm not sure whether the author include the diabetic samples in this dataset. The author intended to show that CD59 is highly expressed in the human pancreatic islets and silencing CD59 (intracellular) can suppress insulin secretion. From Figure 1A we can see that CD59 is highly expressed in the islets, implying that all the samples were from normal people. Otherwise they should not pool the expression pattern from two different groups without comparision. Also, the authors did not mention patients anywhere in the article. Thus I think there is only one group in the data set. You may e-mail the author for direct clarification.

[786]

Thank u so much fr your guidance :)

Answer 1

[Shicheng Guo]

How to get the sample information?

library("GEOquery")
GSE56851 <- getGEO("GSE56851",destdir="/home/sguo/monod/data/geo")
data <- as.data.frame(exprs(GSE56851[[1]]))
phen <- pData(phenoData(GSE56851[[1]]))

Then you can find all the sample information in phen, as Eldon said, these information is essential when you upload array to GEO

Answer 2

[ugandhar.ch1]

i am able to get phenodata information from GSE8977, But i am bit confused how to create a text file? how can i create COVDESC FILE ?