Question

DESeq2 time-series data of multiple treatments with common time 0

0

Entering edit mode

Ting • 0

@ting-12258

Last seen 7.5 years ago

Hi all,

I'd like to use DESeq2 for treatment-specific differences over time.

Here is my colData with a common time 0 for Mock, treatment1 and treatment 2:

> as.data.frame(colData(cds))
                  times    treatments
Control_a            00        Mock
Control_b            00        Mock
Control_c            00        Mock
Mock_03hr_a           03        Mock
Mock_03hr_b           03        Mock
Mock_03hr_c           03        Mock
Mock_06hr_a           06        Mock
Mock_06hr_b           06        Mock
Mock_06hr_c           06        Mock
Mock_12hr_a           12        Mock
Mock_12hr_b           12        Mock
Mock_12hr_c           12        Mock
treatment1_03hr_a      03      treatment1
treatment1_03hr_b      03      treatment1
treatment1_03hr_c      03      treatment1
treatment1_06hr_a      06      treatment1
treatment1_06hr_b      06      treatment1
treatment1_06hr_c      06      treatment1
treatment1_12hr_a      12      treatment1
treatment1_12hr_b      12      treatment1
treatment1_12hr_c      12      treatment1
treatment2_03hr_a      03      treatment2
treatment2_03hr_b      03      treatment2
treatment2_03hr_c      03      treatment2
treatment2_06hr_a      06      treatment2
treatment2_06hr_b      06      treatment2
treatment2_06hr_c      06      treatment2
treatment2_12hr_a      12      treatment2
treatment2_12hr_b      12      treatment2
treatment2_12hr_c      12      treatment2

When I tried this:

cds <- DESeqDataSetFromMatrix(countData = data, colData = colData, design = ~ treatments+times+treatments:times)

I got the error message:

Error in checkFullRank(modelMatrix) : the model matrix is not full rank, so the model cannot be fit as specified. One or more variables or interaction terms in the design formula are linear combinations of the others and must be removed.

I've been searching for the reasons and solutions like this post: With DESeq2 "Not full rank" Error with design ~ line + time + condition, and went through its comments. However, since my data is unpaired, using:

colData$times.nested = factor(c(rep(1,3),rep(rep(2:5,each=3),3)))

cds <- DESeqDataSetFromMatrix(countData = data, colData = colData, design = ~ treatments + times + treatments:times.nested + treatments:times)

or

mm = model.matrix(~ treatments + times + treatments:times, colData(cds))

cds = DESeq(cds, full=mm, betaPrior=FALSE)

gave me the same error.

How do I capture differentially expressed genes of each treatment vs. Mock over time with common time 0 incorporated?

Best regards,

Ting

deseq2 multiple time points multiple treatments • 4.1k views

ADD COMMENT • link updated 7.5 years ago by Michael Love 42k • written 7.5 years ago by Ting • 0

score 1 · Answer 1 · 2017-05-15

1

Entering edit mode

Michael Love 42k

@mikelove

Last seen 7 hours ago

United States

hi Ting,

There was a similar question posted recently, which didn't have any treatment samples at time 0. (There was also a second level of grouping in that question, but I didn't address this in my reply to the question.) Can you take a look at my answer and then see if you have further questions:

A: deseq2 multifactor timeseries help

For building the DESeqDataSet you can use a design of ~1, and then you can use matrices for the arguments 'full' (and 'reduced' if you want to perform a likelihood ratio test).

ADD COMMENT • link 7.5 years ago Michael Love 42k

0

Entering edit mode

Hi Michael,

With your solution, it works well now! Thank you so much for the prompt reply.

ADD REPLY • link 7.5 years ago Ting • 0

0

Entering edit mode

Hi Michael,

As I checked the comparisons of the analysis, it showed:

> resultsNames(cdsFull)
"Intercept" "times03" "times06"                
"times12" "times03.treatmentstreatment2" "times06.treatments treatment2"
"times12.treatments treatment2" "times01.treatmentsYeast" "times03.treatments treatment1"
"times06.treatments treatment1" "times12.treatmentstreatment1"

Does that shows there's no statistical comparisons between treatments as you said in A: deseq2 multifactor timeseries help ? Can genes with p.adj<0.05 in "times03.treatmentstreatment2" be interpreted as significant differences between treatment2 and Mock at times03 or just significant differences in treatment2 at times03 over time? If it's the latter case, is there other way to do the differential analysis between one of the treatments and Mock over time including common time 0? I once thought copy-paste the data of common time 0 to each treatment, thus the data set will be paired. But I guess this will change size factors?

Or sacrifice time 0 by removing the common time 0 and then just compare treatments over times03, times06, and times12.

ADD REPLY • link 7.5 years ago Ting • 0

0

Entering edit mode

times03.treatmentstreatment2 => treatment 2 vs mock at time 3

To compare treatment 2 vs treatment 1 at time 3, you would use the 'contrast' argument:

res <- results(dds, contrast=list("times03.treatmentstreatment2", 
                                  "times03.treatmentstreatment1")

ADD REPLY • link 7.5 years ago Michael Love 42k