[DESeq] invalid class "CountDataSet" object

[Azazel]

I encountered problems after re-installing DESeq. I had DESeq 1.4.1 installed in R 2.13.0 (Win 7) for a couple of weeks now, running my scripts without problems. Today I re-installed DESeq (though it's still at 1.4.1) and now I run into problems. I am using exactly the same script which used to work previously:

> setwd("C:/Documents and Settings/azazel/Desktop/workspace")
> library(DESeq)
> countsTable <- read.table(file="DESeq.in", header=TRUE, stringsAsFactors=TRUE)
> rownames(countsTable) <- countsTable$gindex
> countsTable <- countsTable [,-1]
> conds <- c("x1","x2","x3","x4","x5","x6","x7","x8")
> cds <- newCountDataSet(countsTable,conds)
Error in validObject(.Object) : 
  invalid class "CountDataSet" object: undefined class for slot "experimentData" ("MIAxE")

I do not understand this error (the last two lines in above code-block).

Does anyone know how I can fix this?

[Wolfgang Huber]

Azazel,

thanks for the feedback. It is difficult to tell because you do not give enough detail (see below on how to fix that), but one reason could be that your versions of Bioconductor packages on which DESeq builds (e.g. Biobase) are out of sync.

My recommendation would be to install R2.14 from http://cran.r-project.org and then follow the installation instructions at http://www.bioconductor.org/packages...tml/DESeq.html, i.e. run

source("http://www.bioconductor.org/biocLite.R")
biocLite("DESeq")

There have been changes to DESeq in the newer version that you will get, and you may have to adapt some of your scripts. I find some of these changes quite important, they are worth the trouble.

How to make an informative post (which increases the likelihood and speed of getting a good answer)? See http://www.bioconductor.org/help/mai...posting-guide/
IMHO this is a useful posting guide also for posts to Seqanswers.

Hope this helps
Wolfgang

[Azazel]

It is difficult to tell because you do not give enough detail (see below on how to fix that)

How to make an informative post (which increases the likelihood and speed of getting a good answer)? See http://www.bioconductor.org/help/mai...posting-guide/

Thank you very much for your answer regarding my question. However, regarding your criticism of my original post: I have asked a well-formulated question including the relevant details, provided a proper subject, I have listed versions used, and provided sourcecode! Care to explain what you mean by "not enough detail"? Do you want a coredump of my machine or what. The link you mention is just superfluous bla-bla and does not contain anything which I have not done already. I am always thankful for the great information and help I receive through SEQanswers, and I am honestly thankful for your insightful posts at SEQanwers... but referring me to that link is simply patronizing and not helpful or productive at all.

[Wolfgang Huber]

Azazel,

you did not provide the output of sessionInfo(). This is relevant because one possible reason for the error that you got is that your packages are out sync. But as I said, it is difficult to tell.

Did running source("http://www.bioconductor.org/biocLite.R"); biocLite("DESeq") solve the problem?

Best wishes
Wolfgang.

[Azazel]

Azazel,

Did running source("http://www.bioconductor.org/biocLite.R"); biocLite("DESeq") solve the problem?

I managed to update by simply doing a fresh install of R version 2.14.0 (2011-10-31), and then DESeq 1.6.0. The problems were caused by bioconductor dependencies.