Oligo: GeneChip miRNA 4.0 analysis only returns control probes

Question

[Keiron O'Shea]

This is the first time I've ever posted on biostars, and I'm still not 100% sure of the etiquette - please feel free to berate me if I've forgotten to mention something as I'm here to learn.

I've been trying to understand this problem for the best part of 3 days now, and would like some help.

I've been given a number of CEL files that were produced using a GeneChip™ miRNA 4.0 Array. The samples have come from human tissue. In an attempt to obtain the expression data, I have written the following R script:

library(oligo)
# I've seen some issues with this failing to load on oligo::read.celfiles.
library(pd.mirna.4.0)

celFiles <- oligoClasses::list.celfiles("/home/uid/path/to/data", full.names=TRUE)
rawData <- oligo::read.celfiles(celFiles)

expressionSet <- oligo::rma(rawData)

This returns a "Large ExpressionSet" of 2.5Mb in size (from the original 17.9Mb loaded in as rawData). Sadly, when I want to take a look at the probe identifiers, only control probes seem to be returned (as in everything ending with "_st").

Biobase::featureNames(expressionSet)

I'm not 100% sure where I've gone wrong here as I've analysed quite a lot of miRNA in the past - although this is my first attempt at analysing GeneChip miRNA 4.0 Array data.

Any help would be massively appreciated.

Best wishes,

Keiron.

[Keiron O'Shea]

I think I've confused myself a bit here. I would like some clarification, but would I be right in saying that st is just appended to the end of the probeset identifier to affirm that it's a Affymetrix annotation?

[iraia.munoa]

Hi! I am in the same point. Is it posible when doing the rma() for normalization to obtain more columns added in the output file, I mean other information apart from the ID name of the miRNA, such as the specie, transcript ID, etc. Or the only way to know this info is to take the output and do an intersect with the annotation table you provide in the previous link?

Thanks in advance,

Iraia

[Kevin Blighe]

These will be contained in the NatAffx annotation file, which you can obtain from Affymetrix / ThermoFisher's website.

[anand]

Hi Keiron,

Thanks for the detailed response. I am working on Affymetrix miRNA.3.0 and have normalized my data using 'oligo' followed by differential expression analysis using limma package.

I was wondering how can I retrieve miRNA expression p-values after oligo normalization?

Anand

[Kevin Blighe]

Hi Anand. please use limma to perform the differential expression analysis. see the toptable() function.

Answer 1

[Kevin Blighe]

Hey Keiron,

Yes, you are correct: The probes will have a _st appended to them. This indicates that they are ST array probes. You can see (and download) an annotation table for this array here: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GPL19117

Note that you can access the background corrected, quantile normalised, log (base 2) transformed expression values with exprs(expressionSet).

If you next need to perform differential expression analysis, you can try to use limma.

Note that you should not have to explicitly reference functions with :: - this is only important where namespace issues exist, i.e., where a function has the same name across 2 or more packages.

Kevin