DESeq2 for repeated/clustered samples
1
0
Entering edit mode
jperin • 0
@jperin-22588
Last seen 4.9 years ago

How does DESeq2 account for the dependency between samples in repeated 16S data, for example, relative abundance for samples from the same child?

Following the vignette section on time series, I've written the code below, which runs fine but I'm nervous that the variance is being underestimated, since I don't see a way to include the info about which samples are from the same child.

dds <- DESeqDataSetFromMatrix(countData = count.table, colData = coldata, design = ~ case + time + time.case)

dds <- DESeq(dds, test="LRT", reduced=~time.case) results(dds)

deseq2 repeated measures • 733 views
ADD COMMENT
0
Entering edit mode
@mikelove
Last seen 1 day ago
United States

If you want to test the effect on a condition (or in this case time) and account for patient level baseline, you would just use ~patient + condition. For an LRT you would use a reduced design of ~patient.

If you have a single sample per patient per condition (or time point) you cannot estimate something like an interaction between patient and condition.

ADD COMMENT
0
Entering edit mode

I want to see whether the change over time is similar between two groups of patients. That's the interaction between group and time, but this is not full rank when including patient as an additional factor. That is, the model, ~patient + case+ time+ case.time is not estimable.

ADD REPLY
0
Entering edit mode

We have a section in the vignette on how to compare across groups when individuals are nested within groups:

https://bioconductor.org/packages/release/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#group-specific-condition-effects-individuals-nested-within-groups

ADD REPLY
0
Entering edit mode

OK, that looks more reasonable. Thank you for your help!

ADD REPLY

Login before adding your answer.

Traffic: 742 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6