Is It Possible To Find Sample Batch # In Cel Files?

Question

[Brian Tsai]

Hi,

I've been downloading raw CEL files from the Gene Expression Omnibus (GEO), and have been trying to process them -- i'd like to account for batch effect when computing differential expression, but the authors didn't provide the information explicitly in their annotations. Is this information stored/retrievable through the CEL files through Bioconductor?

[Leonor Palmeira]

Could you give us the accession number of the experience you are working on? This information should be present in the annotations and in the data describing the experiment (or in the paper). Your last chance, if the information remains inaccessible, is to write an email to the authors of the experiment. Either way, you won't find this information in the CEL files, for sure.

Answer 1

[okko.clevert]

Hi, readout the cel-file header using the function readCelHeader in the affxparser package and take a look at the scan date.

Answer 2

[Sean Davis]

From Wolfgang Huber on the Bioc email list:

the 'readAffy' function in the 'affy' package automatically populates the 'ScanDate' field in the resulting AffyBatch object, which you can access with syntax like protocolData(a)$ScanDate where I have assumed that 'a' is an AffyBatch.

Besides explicitly providing the batches, one can use the sva bioconductor package to "discover" batch effects in the data, so you should consider looking there, also.

Answer 3

[Irsan]

If you provide a GEO accession number (like Leonor said) we can look it up for you but as long as you don't I suggest you take a look at the GEOquery package and read the vignette on how to deal with GEO data.