GEOquery issue

Question

[Jaber]

Hi community,

I've been lately facing an issue quering data from GEO database using GEOquery package, as when I want to get the expression data it didn't work for me and also I usually get olde version of the data, unlike from the direct link.

is there any suggestions?

 # Download GSE108022 dataset from GEO
gse <- getGEO("GSE108022", GSEMatrix = TRUE)
gse_data <- gse[[1]]
expr <- exprs(gse_data)

pheno <- pData(gse_data)

I also tried using this

getGEOSuppFiles()

This was not the case only for this GSE, I tried different GSEs, and it was same problem.

Thanks in advance

Answer 1

[Ming Tommy Tang]

sometimes, I also get no expression matrix with GEOquery. It could be how the author uploads the data to GEO. Go and see if you can find the expression matrix in GEO manually. If not, GEOquery can not do much.

[Jaber]

Thank you for your response, yes I guess so,

Answer 2

[ATpoint]

GEOquery can only retrieve array data, not sequencing data which this is. Anyway, if you visit https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE108022 and go to the bottom, there are counts and metadata provided in the supplemental section, which you can download.

[Jaber]

Thank you for your response,

Yes that what I usually do,

But I am still wondering why getGEOSuppFiles() doesn't work either, when I tried it it retrieved the clinical meta-data but didn't get me the expression data.

[Sean Davis]

This works for me:

library(GEOquery)
file_info <- getGEOSuppFiles('GSE108022')
rownames(file_info)
count_matrix <- readr::read_table(rownames(file_info)[1])
dim(count_matrix)
head(count_matrix)

[Jaber]

Thank you, I will try it

[Sean Davis]

GEO can handle counts from any assay, including sequencing. Unfortunately, though, NCBI does not require submitters to submit the count data in a format that GEO can include directly in the GEO record, meaning that GEOquery will come up short in those cases. In the case that the submitter does not do so, the data MAY be in the supplemental files or they may not even be present.

[Jaber]

thank you, I hope NCBI do require submitters to submit in GEO format in the future.