"mogene10stv1probe" probe_length issue

Question

[debajyoti.bionova]

Dear Concern,

This is Deb from Hong Kong. I would like to bring your kind attention in connection with executing microarray data analysis with "mogene10stv1probe" (based on the original experimental facts). The range for this probe length is showing 21x25, which is supposed to be 25x25. Therefore, I am stuck with processing my data (.CEL files) using following command lines:

probepkgname="mogene10stv1probe"
seq1=get(probepkgname) prlen=nchar(seq1$sequence) range(prlen)

[1] 21 25

celfiles.gcrma <- gcrma(celfiles) Adjusting for optical effect...............................................................................................................................................................................................................................................................................................Done. Computing affinitiesError: length(prlen) == 1 is not TRUE

showing the above error message while computing affinities

However, I tried running one test run with human data, supported with "hgu133plus2". It was working perfectly. And,

probepkgname="hgu133plus2 probe"
seq1=get(probepkgname) prlen=nchar(seq1$sequence) range(prlen)

[1] 25 25

this range returned with 25x25 # thus, next commands for "adjusting optical effects" and "computing affinities" etc. are working smoothly.

Please suggest me how to deal with this execution issue. Thanks in advance for your kind concern.

Thank you.

Best,

Deb

Answer 1

[Kevin Blighe]

Use the oligo package and then try these series of commands (using rma(), not gcrma()):

library("oligo")

#Read in the data into a dataframe
#readTargets will by default look for the 'FileName' column in the spcified file
targetinfo <- readTargets("Targets.txt", sep="\t")
CELFiles <- list.celfiles("SampleFiles/", full.names = TRUE)

#Raw intensity data
project <- read.celfiles(CELFiles)

#Background correct, normalize, and calculate gene expression
project.bgcorrect.norm.avg <- rma(project, target="probeset")

[debajyoti.bionova]

Thank you so much, Kevin, for your response. I am trying this now.

[debajyoti.bionova]

Hi Kelvin, I tried your commands, but following error is occurring:

> project <- read.celfiles(CELFiles)
       All the CEL files must be of the same type.    Error: checkChipTypes(filenames, verbose, "affymetrix", TRUE) is not TRUE

But my all .cel files are just unzipped from .tar files. Any suggestions, how to deal with this?

Thank you.

Best, Deb

[Kevin Blighe]

What are the contents of the CELFiles object?

[debajyoti.bionova]

It consists of Affymetrix data. Anyway, finally I could able to execute your commands, and all were working pretty fine. Solved my normalization issue with "rma()" instead of "gcrma()" as you mentioned. I have got 3 outputs, "intensitiesxxx.ff", "pmxxx.ff" and "rmaxxx.ff".

However, since I am kinda new in biological data analysis, therefore, not gained much expertise yet. Could you please guide me further, for filtering the normalized output, QC steps and gene symbol mapping. Would be great to receive your kindness!

Thank you.

Best, Deb

[Kevin Blighe]

Hi Deb,

The object created from the rma() command should be an Expression Set object, an object whose format is standardised and whose normalised and log (base 2) expression values can be accessed with the exprs() function.

For example:

expression <- rma ( ... )

exprs(expression)

Be aware that you can 'summarise' the expression values at the probe, exon, or gene level. This is performed via the rma() function. Please take a look at Section 3 and 4 of the document Preprocessing Affymetrix Exon ST and Gene ST Arrays in order to note the difference between these.

---

You can then do simple things with this normalised data, such as:

• Boxplots, violin plots, etc
• PCA
• A: PCA plot from read count matrix from RNA-Seq

To conduct differential expression analysis, you can use a package called limma. There are many tutorials online about how to do this, and also many posts on Biostars.

[debajyoti.bionova]

Thank you so much, Kevin. This worked well. :-) Please keep in touch.