EdgeR problem

[yifangt]

Hello,

Tried using edgeR and met a problem that bugs me for a while, could not go further with edgeR following to the manual step-by-step.
The experiment consists of four lanes of controls and three lanes of cold treatment to see the DE of the cold treatment.

Here is my screen output following exactly the manual.

> raw.data <- read.delim(file="Cold-CK.csv")
> names(raw.data)
[1] "FeatureID" "s1FC36"    "s1FC77"    "s1FC79"    "s2FC77"    "s2FC36"   
[7] "s2FC79"    "s3FC77"   
> 
> data <- raw.data[, 2:8]
> rownames(data) <- raw.data[, 1]
> group <- c(rep("CK", 4), rep("Cold", 3))
> data <- DGEList(counts=data, group=group)
Calculating library sizes from column totals.
> dim(data)
[1] 28775     7
> 
> data <- data[rowSums(1e+06*data$counts/expandAsMatrix(data$samples$lib.size, dim(data)) >1) >=3, ]
Error in inherits(x, "data.frame") : 
  could not find function "expandAsMatrix"

The expandAsMatrix function seems built-in. Not sure if any body has met this before. Any clue with this problem is greatly appreciated.

Yifang

[Gordon Smyth]

Hard to say what the problem is, because expandAsMatrix() certainly is an edgeR function, although not one that users usually need to call directly.

I suggest that you try re-installing the current edgeR from Bioconductor. If that doesn't solve your problem, please ask for help on the Bioconductor mailing list. The edgeR maintainers regularly read and reply to that list, but much less often to SEQanswers. You'll need to include sessionInfo() output.

[Balat]

I am also getting the same error. I have downloaded the latest version of the software from the Bioconductor site.

[Balat]

It looks like the latest version of edgeR is built for R 2.13.0. I now don't get the error when I used the edgeR in the latest version of 'R'.

[Davis McC]

I cannot reproduce this error. I strongly suspect that you are not using edgeR version 2.2.5, which is the latest release. edgeR version 2.2.5 was released with Bioconductor 2.8. The previous release of edgeR did not contain the expandAsMatrix() function, so if you are using an older version then I would expect you to see this message.

One trick with Bioconductor is that it installs the version of Bioconductor that corresponds to the version of R that you are running. The latest release of Bioconductor is 2.8, which aligns with R 2.13.

Thus, to obtain the latest release of edgeR from Bioconductor with the biocLite() command you must be running R 2.13. If you are running R 2.12 and install using biocLite then you will get the version of edgeR from Bioconductor 2.7.

This is why Gordon (and others!) will ask you to provide the output from sessionInfo() when you are reporting a possible bug. If we can't tell exactly what version of the software you are using then we have to guess at the cause of the problem. If we know exactly what versions of software you are using then we can pinpoint a bug and/or a solution to your problem much more easily.

To reiterate Gordon's request: the Bioconductor mailing list is a better forum for this sort of question. It's a more dynamic community to respond. I feel SeqAnswers is better for more general discussion. But then that just might be me.

HTH.

Best wishes
Davis

[yifangt]

edgeR: Error gone!

Thanks, the error has gone after update the edgeR to 2.2.5 and R to 2.13.0.