Hi, 

I am new to the differential expression analysis field. I am trying to compare the DE genes in cells treated with varied concentrations of a drug, with the cell's RNA being sequenced at different times throughout the treatment. I have 20 condtions (I'll list the time of RNA extraction after adding the drug, and the drug concentration in [ ]) : 0h [0] ; 45min [0], 45min [1000], 90min [0], 90min[500], 90min [1000], 90min [2000], 3hr [0], 3hr[1000], 6hr [0], 6hr[500], 6hr[1000], 6hr[2000], 12hr[0], 12hr[1000], 24hr[0], 24hr[500], 24hr[1000], 24hr[2000]. I am trying to figure out how to correctly implement this into the design matrix when making the DESeq. 

Please let me know if you have any suggestions on how to improve my design matrix. 

For my colData, I have included the text found in the sample.conditions file. The two columns are Treatment and Time. 

Sample Treatment Time

0.0A 0 0h

3h0 0 3h

24h0 0 24h

0.0B 0 0h

3h1000 1000 3h

24h500 500 24h

45m0 0 45m

6h0 0 6h

24h1000 1000 24h

45m1000 1000 45m

6h500 500 6h

24h2000 2000 24h

90m0 0 90h

6h1000 1000 6h

90m500 500 90h

6h2000 2000 6h

90m1000 1000 90h

12h0 0 12h

90m2000 2000 90h

12h1000 1000 12h

dds <- DESeqDataSetFromMatrix(countData = count_matrix,
                              colData = sample.conditions,
                              design = ~ Treatment+Time)

I'd recommend you meet and discuss your experimental design and analysis plan with a statistician. Your design is not a standard one (not all time points have all concentrations), and so you need to plan out how you'll do the modeling of expression over time. There are multiple options and choices to make, it's not straightforward from knowing the time points and concentrations tested at each time point. You can still use DESeq2, but you'll need to come up with a meaningful design for answering whatever research question you have and it's not so much a software question, as an analysis plan question.