Question

DESeq2 / agricolae packages conflict

0

Entering edit mode

etienne.rifa • 0

@etiennerifa-23302

Last seen 4.6 years ago

Hello,

I'm developping package with DESeq2 and agricolae as dependencies. But DESeq() function seems to not work anymore when package agricolae is called.

library(agricolae)
library(DESeq2)

## DESeq() Example commands 
cnts <- matrix(rnbinom(n=1000, mu=100, size=1/0.5), ncol=10)
cond <- factor(rep(1:2, each=5))

# object construction
dds <- DESeqDataSetFromMatrix(cnts, DataFrame(cond), ~ cond)

# standard analysis
dds <- DESeq(dds)

Error:

Error in model.matrix.formula(design(object), colData(object)) : 
  data must be a data.frame

How can I handle this issue?

Thanks for your help Etienne

deseq2 dependendies conflict agricolae • 824 views

ADD COMMENT • link updated 4.6 years ago by Michael Love 43k • written 4.6 years ago by etienne.rifa • 0

score 1 · Answer 1 · 2020-04-10

1

Entering edit mode

Michael Love 43k

@mikelove

Last seen 6 days ago

United States

Thanks for this report. I'm actually not sure why there is this conflict. I downloaded their package but I couldn't find where there would be some functions that change the behavior of DESeq2.

Anyway, I think I have modified the DESeq2 package calls to model.matrix that avoid the error. I was able to reproduce the error before the fix, and it is solved afterward.

This is then fixed in the latest version of DESeq2 on github (which will be the released version of DESeq2 on Bioconductor later in April). For now you can obtain this with:

install_github("mikelove/DESeq2", dependencies=FALSE)

I think this will work fine if you had the current release of Bioconductor, and then you can upgrade to the latest Bioconductor once it is released in end of April.

ADD COMMENT • link 4.6 years ago Michael Love 43k

0

Entering edit mode

I see

> dds <- DESeq(dds)
Error in model.matrix.formula(design(object), colData(object)) :
  data must be a data.frame
> traceback()
5: stop("data must be a data.frame")
4: model.matrix.formula(design(object), colData(object))
3: stats::model.matrix(design(object), colData(object))
2: designAndArgChecker(object, betaPrior)
1: DESeq(dds)
> class(design(dds))
[1] "formula"
> methods(model.matrix)
[1] model.matrix.coxph*    model.matrix.default   model.matrix.formula*
[4] model.matrix.lm        model.matrix.reStruct* model.matrix.survreg*
[7] model.matrix.terms*
see '?methods' for accessing help and source code
> getAnywhere("model.matrix.formula")
A single object matching 'model.matrix.formula' was found
It was found in the following places
  registered S3 method for model.matrix from namespace AlgDesign
  namespace:AlgDesign
with value

function (frml, data = sys.frame(sys.parent()), ...)
{
    if (!missing(data)) {
        if (!inherits(data, "data.frame"))
            stop("data must be a data.frame")
        if (!inherits(frml, "formula"))
            stop("frml must be a formula")
        frml <- expand.formula(frml, colnames(data))
    }
    model.matrix.default(frml, data, ...)
}
<bytecode: 0x7ff6e320e538>
<environment: namespace:AlgDesign>

so it looks like AlgDesign (poorly) implements a formula method that masks the (better written) default method. It seems like the AlgDesign people should be contacted...