Download Microarray GSE files GEO

Question

[E]

Hi,

I have troubles downloading the file "GSE21344" from GEO (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE21344). When I run the code below, I would expect to get the expression Matrix. Instead, the variable exprMatrix contains the names of the GSM files. Do you know how to download this file correctly?

gset <- getGEO("GSE21344", GSEMatrix =TRUE)
if (length(gset) > 1) idx <- grep("GPL1322", attr(gset, "names")) else idx <- 1
gset <- gset[[idx]]

exprMatrix <- exprs(gset)

Thanks in advance, E

[Kevin Blighe]

Hi,

The authors have not uploaded any processed expression data; so, you will have to re-process from the raw data CEL files, which are available. You can, of course, contact the authors to see if they can send to you the processed data.

You can see all files that are available here: https://ftp.ncbi.nlm.nih.gov/geo/series/GSE21nnn/GSE21344/

Kevin

[E]

Hi Kevin, this makes sense. I actually never re-processed microarray data from raw files. Do you know a good tutorial on this? (I looked before but did not find any I liked.) E

[Kevin Blighe]

Hey again, I posted an answer below. Hopefully you can do the remainder of the analysis.

Answer 1

[Kevin Blighe]

[following from comment trail]

Hi, it may be your lucky day because I have been processing other microarray studies all morning. Here is how you could process this (after you download the CEL files and unzip them):

[note that you may be asked to install one or both of these packages, while you'll notice another being automatically downloaded when you run rma()]

  # workspace setup
    require(oligo)
    require(drosophila2.db)

    gseID <- 'GSE21344'

  # read in and prepare data
    # raw CEL files
      message('--loading CEL files for ', gseID)

      raw <- read.celfiles(
        filenames = list.files('GSE21344_RAW/', pattern = '*cel', full.names = TRUE),
        sampleNames = gsub('\\.cel$', '', list.files('GSE21344_RAW/', pattern = '*cel')))

    # RMA
      message('--RMA normalising...')
      gset <- rma(raw)
      message('--Done.')

      gset <- exprs(gset)
      probes <- rownames(gset)
      samIDs <- colnames(gset)

    # annotate
      annotLookup <- select(drosophila2.db, keys = probes,
        columns = c('PROBEID', 'ENSEMBL', 'SYMBOL'))

      # remove probes with any NA mapping
        annotLookup <- annotLookup[!is.na(annotLookup$ENSEMBL) & !is.na(annotLookup$SYMBOL),]
        annotLookup <- annotLookup[!duplicated(annotLookup$PROBEID),]

      # look up the ensembl ID and gene symbol
        probes <- probes[which(probes %in% annotLookup$PROBEID)]
        gset <- gset[probes,]
        all(rownames(gset)==probes)
        all(probes == annotLookup[match(probes, annotLookup$PROBEID),'PROBEID'])
        geneid <- annotLookup[match(probes, annotLookup$PROBEID),'SYMBOL']
        ens <- annotLookup[match(probes, annotLookup$PROBEID),'ENSEMBL']

  # finalise the dataset
    final <- data.frame(ens = ens, symbol = geneid, gset)
    head(final)
                         ens  symbol GSM533369 GSM533370 GSM533371 GSM533372
    1616608_a_at FBgn0001128   Gpdh1  3.075607  2.987812  3.121998  2.794935
    1622892_s_at FBgn0035889   mkg-p 11.411883 11.415364 11.400395 11.779498
    1622893_at   FBgn0040736     IM3  3.061561  3.233674  3.483630  3.316230
    1622894_at   FBgn0034454 CG15120  3.930225  3.829407  3.770370  3.873030
    1622895_at   FBgn0052075 CG32075 10.966323 10.857122 10.971413 10.743497
    1622896_at   FBgn0038966   pinta 13.159087 13.212790 13.309891 13.335578
                 GSM533373 GSM533374 GSM533375 GSM533376 GSM533377 GSM533378
    1616608_a_at  2.827791  2.918710  3.072793  3.108860  2.946866  3.167440
    1622892_s_at 11.803834 11.717940 11.691620 11.720805 11.688627 11.429850
    1622893_at    3.207360  3.162072  3.193889  3.101028  3.034796  3.258224
    1622894_at    4.131397  3.974184  3.777796  3.719844  3.854079  3.788482
    1622895_at   10.745481 10.780364 10.729215 10.632876 10.657724 11.078941
    1622896_at   13.271371 13.348126 13.644451 13.747381 13.740111 11.819523
                 GSM533379 GSM533380 GSM533381 GSM533382 GSM533383 GSM533384
    1616608_a_at  3.171195  3.078535  3.147919  3.067052  3.046781  3.056133
    1622892_s_at 11.484802 11.503235 11.868650 11.820685 11.868820 11.618499
    1622893_at    3.468744  3.510779  3.456605  3.573743  3.275730  2.915941
    1622894_at    3.904618  3.906584  3.704304  3.772455  3.656190  3.888913
    1622895_at   11.129996 11.103732 10.970281 10.920947 10.922826 10.798052
    1622896_at   11.855323 11.790097 11.928086 11.896962 11.949402 12.527205
                 GSM533385 GSM533386
    1616608_a_at  2.987426  2.987495
    1622892_s_at 11.594103 11.531154
    1622893_at    3.274847  3.020663
    1622894_at    3.529930  3.878161
    1622895_at   10.853586 10.790171
    1622896_at   12.542879 12.492722

Hopefully you can retrieve the metadata to match these GSM IDs. Please check the GEO record.

Kevin

[1769mkc]

gset = gcrma(affyBatch)

​
Adjusting for optical effect................................................................................................................................................................................Done.
Computing affinities.Done.
Adjusting for non-specific binding................................................................................................................................................................................Done.
Normalizing
Error in rma(object, subset = subset, background = FALSE, normalize = normalize, : ERROR; return code from pthread_create() is 22

Traceback:

1. gcrma(affyBatch)
2. rma(object, subset = subset, background = FALSE, normalize = normalize, 
 .     verbose = verbose)

I tried this as well but same error

My R and os platform

R version 4.0.3 (2020-10-10)
Platform: x86_64-conda-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

What is going wrong here ?