Question: DESeq2 Paired samples Before and after treatment
gravatar for JK Kim
11 days ago by
JK Kim0
JK Kim0 wrote:


I am new to mRNAseq data analysis and I would like to know how to run deseq2 for my data. (I googled it and I saw some posts like DESEq2 Paired samples Before and after treatment , but I didn't understand it. I am sorry to ask a basic question and take your time...)

Here is my data. (6 patient data and they are paired. They have the same disease and took the same drug. I would like to know different gene expression between before and after condition.)

> coldata
                           samples condition
1_3_12_clamp_0_buffy        1_3_12    before
1_3_12_clamp_200_buffy      1_3_12     after
9_clamp_0_buffy                  9    before
9_clamp_200_buffy                9     after
010_0_buffy                     10    before
010_200_buffy                   10     after
11_15_006_0m_buffy       11_15_006    before
11_15_006_200m_buffy     11_15_006     after
12_27_11_clamp_buffy_0    12_27_11    before
12_27_11_clamp_buffy_200  12_27_11     after
013_clamp_0m_buffy             013    before
013_clamp_200m_buffy           013     after

> head(txi.rsem$counts) # I removed some of columns for a better look.

  1_3_12_clamp_0_buffy 1_3_12_clamp_200_buffy 9_clamp_0_buffy 9_clamp_200_buffy 010_0_buffy 010_200_buffy
ENSG00000000003 1 1 1 3 0 0
ENSG00000000005 0 0 0 0 1 0
ENSG00000000419 470 528 388 563 36 180
ENSG00000000457 306.02 243.22 254.25 273.97 248 186
ENSG00000000460 398.98 484.78 262.75 307.03 62 96
ENSG00000000938 3221 2756 1750 1920 17516 7642


I used the rsem output for tximport and ran deseq2 and I got this result. 

>ddsTxi <- DESeqDataSetFromTximport(txi.rsem, colData = coldata, design = ~samples + condition)
>dds <- ddsTxi[ rowSums(counts(ddsTxi)) > 1, ]
>dds <- DESeq(dds)
>res <- results(dds, pAdjustMethod = "fdr")
> res
log2 fold change (MLE): condition before vs after 
Wald test p-value: condition before vs after 
DataFrame with 26959 rows and 6 columns
                    baseMean log2FoldChange     lfcSE        stat    pvalue      padj
                   <numeric>      <numeric> <numeric>   <numeric> <numeric> <numeric>
ENSG00000000003    1.5229213     0.29551456 3.0266631  0.09763709 0.9222205        NA
ENSG00000000005    0.9538215     1.13808315 3.0745585  0.37016149 0.7112622        NA
ENSG00000000419  265.9276159     0.10175892 0.6761213  0.15050394 0.8803670 0.9987058
ENSG00000000457  230.5449617    -0.09454739 0.3944541 -0.23969175 0.8105692 0.9987058


So, I would like to know what I did is a right way of analysis my data?

Sorry for my English and thank you in advance.


ADD COMMENTlink modified 10 days ago • written 11 days ago by JK Kim0
gravatar for Michael Love
10 days ago by
Michael Love14k
United States
Michael Love14k wrote:

Yes, this is the recommendation in the FAQ in the DESeq2 vignette:


Except you need to take care of factor levels of condition. See the Note on Factor Levels in the vignette.

ADD COMMENTlink written 10 days ago by Michael Love14k

Dear Michael,

Thank you for your comment! I am glad what I did is not wrong. And I have some questions to ask if you don't mind. (Sorry again for wasting your time, but hopefully it might help for some new users.)

1. So as the reults shows this is comparison between before and after samples, which means 6 biological replicates for before samples and after samples. Am I right?

2. About factor levels. I saw the deseq2 vignette and if I want to know the difference between "9" and "1_13_2", my code should be ```res <- results(dds, contrast=c("samples", "9", "1_3_12"))``` which means two biological replicates for "9" samples and "1_13_12" samples. Am I right?

Thank you again in advance.


ADD REPLYlink written 10 days ago by JK Kim0

I just meant that it’s more standard to have after vs before. You have before vs after. But it’s up to you how to present the comparisons.

ADD REPLYlink written 10 days ago by Michael Love14k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 184 users visited in the last hour