Hello All,

I am working on time-series RNA-seq data. I have four individuals and collected RNA-seq samples for each individual at 4 different time points (T1, T2, T3 and T4) within one year. The experiments shown below.

> SampleInfo

           Individual   Time
Sample1.1   Sample1          T1
Sample1.2   Sample1          T2
Sample1.3   Sample1          T3
Sample1.4   Sample1          T4
Sample2.1   Sample2          T1
Sample2.2   Sample2          T2
Sample2.3   Sample2          T3
Sample2.4   Sample2          T4
Sample3.1   Sample3          T1
Sample3.2   Sample3          T2
Sample3.3   Sample3          T3
Sample3.4   Sample3          T4
Sample4.1   Sample4          T1
Sample4.2   Sample4          T2
Sample4.3   Sample4          T3
Sample4.4   Sample4      T4

I would like to identify differentially expressed genes associated with seasonal changes.

For that I use the LRT test and the following model

dds <- DESeqDataSetFromMatrix(countData = counts,
           colData = SampleInfo,
           design= ~Time)
dds <- DESeq(dds, test="LRT", reduced=~1)

Would this give me genes that are differentially epressed across all time points?

Thanks

The LRT here will identify genes with any change across the groups. If you say seasonal changes then you might be more interested in distinct patterns, such as cyclic effects, so with a periodicity. Common approaches for this are cosinor regression, but just four timepoints might not be enough for this. See seasonal and circadian literature for this and this guide https://bioconductor.org/packages/release/workflows/vignettes/RNAseq123/inst/doc/designmatrices.html which is a great reads for different sorts of models. Also be sure to include the individual to the design since this seems like a paired design which is powerful. For hands-on guidance you will need to ask elsewhere since the support site is not meant for this.