Question

Pairwise comparison of multiple groups with different base

0

Entering edit mode

thjnant ▴ 10

@thjnant-23566

Last seen 5.7 years ago

Hello,

I have posted this question to the biostar forum but I couldn't get a thorough explanation. Sorry for repetition.

I have 4 different groups (species) that I want to look into their differential gene expression. I call them A, B, C and D. I have 5 - 8 replicates for each group and I am using DESeq2 for the analysis.

Which of the following ways is a recommended setting to continue with the analysis:

Create one dataset for each pairwise comparison, create the respective dds data matrix and run the DESeq function.
Create one dataset with all groups, create the dds data matrix, run the DESeq function and extract comparison of interest by contrast.

In the first approach, I get 215 significant genes for the comparison of A vs B while in the second approach, I get 49 significant genes. Of the top 50 genes with the lowest p-value in the two sets, 27 are common. How can I explain this difference and which approach is correct?

Thank you!

deseq2 rna-seq r • 2.8k views

ADD COMMENT • link updated 2 days ago by ja569116 • 0 • written 5.7 years ago by thjnant ▴ 10

0

Entering edit mode

Cross-posted on Biostars: https://www.biostars.org/p/439436/ thjnant, when you do this, in future, can you mention it in your question?

ADD REPLY • link 5.7 years ago Kevin Blighe ★ 4.0k

score 1 · Answer 1 · 2020-05-21

1

Entering edit mode

Michael Love 43k

@mikelove

Last seen 23 hours ago

United States

See the FAQ in the vignette, this is addressed directly.

ADD COMMENT • link 5.7 years ago Michael Love 43k

0

Entering edit mode

Thank so much. So sorry I missed this information, very clear explanation in the FAQ, thanks a lot.

ADD REPLY • link 5.7 years ago thjnant ▴ 10

0

Entering edit mode

Hi Michael Love I have a similar situation, and I feel that I have interpreted the results in the wrong way. In my case I would like to identify the microRNAS - working with that and not genes, still same data : read counts, that are differentially expressed across ALL (more than 2 groups). I did as stated in the manual with the LRT test

miRNA.cet.Jan.15 <- DESeqDataSetFromMatrix(countData = subset.miRNA.cetacean.Jan15,
                                                         colData = subset.metadata,
                                                         design = ~ sp)
sps.miRNA.cet.Jan.15<- DESeq(miRNA.cet.Jan.15, test="LRT", reduced=~1)
res.sps.miRNA.cet.Jan.15 <- results(sps.miRNA.cet.Jan.15)
res.sps.miRNA.cet.Jan.15

As stated in the manual, the results should be across all 4 groups, but the result table

>res.sps.miRNA.cet.Jan.15
log2 fold change (MLE): sp Zcav vs Ddel 
LRT p-value: '~ sp' vs '~ 1' 
DataFrame with 680 rows and 6 columns

indicates that that result table is only about the pair Zcav vs Ddel, which is not what I want/need.

Thus, DESeq2 can't find differentially expressed markers across all groups (3 or more), but rather just pairwise?

ADD REPLY • link 2 days ago ja569116 • 0