Question

DESeq2: Variables in the design section

1

Entering edit mode

Matan G. ▴ 60

@matan-g-22483

Last seen 3.3 years ago

Hi,

Is there a way to set the input in the design section of DESeq2 as variables?

The aim is that the colname chosen will be a variable that can be changed elsewhere in the script rather than the location of the design itself.

Best

deseq2 • 2.4k views

ADD COMMENT • link updated 4.2 years ago by Michael Love 42k • written 4.2 years ago by Matan G. ▴ 60

score 1 · Answer 1 · 2020-09-24

1

Entering edit mode

Michael Love 42k

@mikelove

Last seen 17 hours ago

United States

You can use formula() to convert character string to formula.

ADD COMMENT • link 4.2 years ago Michael Love 42k

0

Entering edit mode

Hi Michael, Thank you for replying.

I'm getting the message:

Error in DESeqDataSet(se, design = design, ignoreRank) : all variables in design formula must be columns in colData

Tried in various ways to apply your suggestion with no success. According to the error I understand that I have to define colData before running dds <- DESeqDataSetFromMatrix...

And then make sure that the variable in the design section would be a column in colData.

This is one of the methods I've tried:

  colData = coldata # where coldata is my metatdata table
  ccon = "cc" # the string I'd like to make as formula which is a function of colData column 
  con = 9 # the variable which will be the number of column to choose from
  formula(ccon ~ colData[con]) # the formula



# then I'm trying to run it this way:
    dds <- DESeqDataSetFromMatrix(
    countData = countdata,
    colData = coldata,
    design = ~ ccon)

I'm obviously using it wrong, I thought it has to do either with the fact that it has to be in the context of a dataset and not a seperate data frame. or with the fact that the column subsetting I'm using is not equivalent to using "$".

Hope you can help me solve this.

All the best

ADD REPLY • link 4.2 years ago Matan G. ▴ 60

0

Entering edit mode

I don’t think you got my suggestion. The design argument takes a formula. So you can provide a character string to formula() to pass to design.

ADD REPLY • link 4.2 years ago Michael Love 42k

0

Entering edit mode

Hi thank you for the input.

Unfortunately i still can't mange to apply it. struggling with the information about formula() usage.

Best regards

ADD REPLY • link 4.2 years ago Matan G. ▴ 60

0

Entering edit mode

I guess this is a bit out of scope for the support site, as there isn't an issue with using DESeq2 per se, but you want to write a program that calls DESeq2 programmatically. As you are stuck at this stage, I would recommend consulting with someone at your institute who has R programming experience.

ADD REPLY • link 4.2 years ago Michael Love 42k

0

Entering edit mode

Ok sure,

Thank you for your time and explanation.

Best

ADD REPLY • link 4.2 years ago Matan G. ▴ 60

0

Entering edit mode

For whoever is interested, what worked for me was:

  con = 4
  clin_con <- colnames(coldata[con])
  clin_con <-  paste("~",clin_con)



  dds <- DESeqDataSetFromMatrix(
    countData = countdata,
    colData = coldata,
    design = as.formula(clin_con))

Best

ADD REPLY • link 4.2 years ago Matan G. ▴ 60