DESeq2 interaction study
1
0
Entering edit mode
sokumoto • 0
@b23b32b7
Last seen 3.3 years ago

Hi, I am trying to compare differential expression between two cultivars, and two timepoints. My goal is primarily to identify the differentially expressed genes between the two cultivars, two timpoints, then also identify the genes that behave differently between the two cultivars and timepoints (interaction). It seems that I should be able to do this by setting up the interaction terms.

I have quasi-mappedt Illumina PE reads using salmon, which has been imported using tximport function.

Column data looks like this; (DAP stands for days after pollination)

>samples
library cultivar    DAP salmon_folder
L01 A   14DAP   Liu_cut_01.out
L02 A   14DAP   Liu_cut_02.out
L03 A   14DAP   Liu_cut_03.out
L04 A   5DAP    Liu_cut_05.out
L05 A   5DAP    Liu_cut_06.out
L06 A   5DAP    Liu_cut_07.out
L07 B   14DAP   Liu_cut_01.out
L08 B   14DAP   Liu_cut_02.out
L09 B   14DAP   Liu_cut_03.out
L010    B   5DAP    Liu_cut_05.out
L011    B   5DAP    Liu_cut_06.out
L012    B   5DAP    Liu_cut_07.out

Now trying to do construct the DEseq dataset from the txi object and run DESeq,

ddsTxi <- DESeqDataSetFromTximport(txi,
                                   colData = samples,
                                   design = ~ DAP + cultivar)


ddsTxi$group <- factor(paste0(ddsTxi$DAP, ddsTxi$cultivar))
design(ddsTxi) <- ~ group
ddsTxi <- DESeq(ddsTxi)

When I look at the results names however,

resultsNames(ddsTxi)
[1]"Intercept"                        "group_14DAPB_vs_14DAPA"
[3] "group_5DAPA_vs_14DAPB"  "group_5DAPB_vs_14DAPA" 

So the results does not seem to include all possible combinations of conditions. What am I doing wrong? How can I compare the combinations such as

cultivar A, 14DAP vs cultivar B, 14DAP
cultivar A, 5DAP vs cultivar B, 5DAP
cultivar A, 5DAP vs cultivar B, 14DAP
cultivar B, 5DAP vs cultivar B, 14DAP

and also identify those genes with signficant interaction?

DESeq2 • 728 views
ADD COMMENT
0
Entering edit mode
swbarnes2 ★ 1.3k
@swbarnes2-14086
Last seen 16 hours ago
San Diego

The resultsNames are normal; the software is assuming that the last group alphabetically is the reference, so it's giving you comparison names for each group against that one. You can use relevel to change the reference level, but you don't have to. You can compare any group to any group with the contrast option when you call results.

ADD COMMENT
0
Entering edit mode

Thanks! contrast function seems to work nicely to compare the two way comparison like cultivar A, 14DAP vs cultivar B, 14DAP

Now if I wanted to identify interaction (i.e. genes that behave differently between the cultivars when comparing the two timepoints) how would I do that?

ADD REPLY
0
Entering edit mode

This link lays it out:

https://www.biostars.org/p/211246/

Be sure to use relevel to make the right thing is the reference level.

ADD REPLY
0
Entering edit mode

Thanks, I think it worked!

ADD REPLY

Login before adding your answer.

Traffic: 765 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