Wired DESeq2 error

[sindrle]

Probably me who has done something wrong, but I dont see what.

First run was fine, but when I ran it again some days later I get this error.

> sampleFiles <- list.files(path="/Volumes/timemachine/HTseq_DEseq2",pattern="*.txt")
> sampleFiles <- sort(sampleFiles,decreasing=TRUE)
> status <- factor(c(rep("Healthy",26), rep("Diabetic",22)), levels=c("Healthy", "Diabetic"))
> timepoints = as.factor(c(2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1))
> patients <- factor(c(88,87,86,83,82,81,75,72,71,70,13,08,01,88,87,86,83,82,81,75,72,71,70,13,08,01,79,77,76,73,67,62,61,55,53,21,04,79,77,76,73,67,62,61,55,53,21,04))
> sampleTable <- data.frame(sampleName = sampleFiles, fileName = sampleFiles, status=status, timepoints=timepoints, patients=patients)
> directory <- c("/Volumes/timemachine/HTseq_DEseq2/")
> design <- formula(~patients + timepoints*status)
> ddsHTSeq <- DESeqDataSetFromHTSeqCount(sampleTable = sampleTable, directory = directory, design = design)

Error in validObject(.Object) :
invalid class "DESeqDataSet" object: the model matrix is not full rank, i.e. one or more variables in the design formula are linear combinations of the others

[sindrle]

The error appears when:

timepoints = as.factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2));

is changed to:

timepoints = as.factor(c(2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1))

[sindrle]

I changed to:

timepoints = factor(c("After","After","After","After","After","After","After","After","After","After","After","After","After","Before","Before","Before","Before","Before","Before","Before","Before","Before","Before","Before","Before","Before","After","After","After","After","After","After","After","After","After","After","After","Before","Before","Before","Before","Before","Before","Before","Before","Before","Before","Before"), levels=c("Before","After"));

But when I change:

patients <-
as.factor(c(04,21,53,55,61,62,67,73,76,77,79,04,21,53,55,61,62,67,73,76,77,79, 01,08,13,70,71,72,75,81,82,83,86,87,88,01,08,13,70,71,72,75,81,82,83,86,87,88));

To:

patients <-
as.factor(c(88,87,86,83,82,81,75,72,71,70,13,08,01,88,87,86,83,82,81,75,72,71,70,13,08,01,79,77,76,73,67,62,61,55,53,21,04,79,77,76,73,67,62,61,55,53,21,04));

The error occurs again...

I really cant grasp this..

[dpryan]

This ends up being the same issue that you posted on the bioconductor email list, but there with edgeR (the solution is the same).

[sindrle]

Yes, Im sorry for that..

But I rally can't understand why:

timepoints = as.factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2));

Changed to:

timepoints = as.factor(c(1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2));

Suddenly gives the error...

[sindrle]

Sorry for spamming my own thread.. But now it worked again!

Its indeed wired.