SEQanswers

Go Back   SEQanswers > Bioinformatics > Bioinformatics



Similar Threads
Thread Thread Starter Forum Replies Last Post
multi-factor vs. pair-wisw design frymor Bioinformatics 4 04-17-2015 01:40 AM
DESeq2 multi-factor designs id0 Bioinformatics 6 01-08-2015 12:23 PM
DESeq2 multi-factor designs id0 Bioinformatics 3 01-10-2014 06:25 AM
DESeq2: Multi-factor designs sindrle Bioinformatics 10 10-21-2013 07:47 AM
DEXSeq for multi-factor design alittleboy Bioinformatics 5 06-27-2013 09:43 AM

Reply
 
Thread Tools
Old 11-16-2015, 11:50 AM   #1
PMRoberts
Junior Member
 
Location: USA

Join Date: Nov 2015
Posts: 1
Default DESeq2 multi-factor design

Hi all,

I have just started using DESeq2 using a multi-factor approach. As I'm new to this (and to statistics), I was hoping to get feedback / corrections on what I think the analysis should look like:

I have the following colData:
Code:
sample	genotype	treatment	library_batch
S1	WT	vehicle	B1
S2	WT	vehicle	B2
S3	WT	vehicle	B1
S4	WT	vehicle	B2
S5	WT	Drug	B2
S6	WT	Drug	B1
S7	WT	Drug	B2
S8	KO	vehicle	B1
S9	KO	vehicle	B2
S10	KO	vehicle	B1
S11	KO	vehicle	B2
S12	KO	Drug	B1
S13	KO	Drug	B2
S14	KO	Drug	B1
S15	KO	Drug	B2
Blocking for library batch effects, I am interested in finding the following DE genes:
Q1) DE genes between vehicle-treated "WT" and "KO" (i.e. the 'baseline' difference);
Q2) Genes that are DE following treatment within each genotype ("WT" and "KO");
Q3) Genes that show a genotype-dependent response to treatment.

Looking through the DESeq2 tutorials and form posts, Q1 and Q2 would probably be easiest addressed by created a 'genotype_treatment' factor and working with contrast:

Code:
colData$condition <- paste(colData$genotype, colData$treatment, sep="_")
and using a design
Code:
design = ~ library_batch + condition
to extract results as:
Q1:
Code:
res1 <- results(dds, contrast=c("condition", "KO_vehicle", "WT_vehicle"))
Q2.1:
Code:
res2wt <- results(dds, contrast=c("condition", "WT_Drug", "WT_vehicle"))
Q2.2:
Code:
res2ko <- results(dds, contrast=c("condition", "KO_Drug", "KO_vehicle"))
As for Q3, I would then repeat the analysis including an interaction term:

Code:
design = ~ library_batch + genotype*treatment
to extract results as:

Code:
res3 <- results(dds, name="genotypeKO.treatmentvehicle")
Is this correct? I was wondering if there is a way to do this without running 2 separeate analysis designs...?

On a more general note, can anyone recommend a resource to get more familiar with these design formulas and what they 'mean'?

Last edited by PMRoberts; 11-16-2015 at 02:48 PM.
PMRoberts is offline   Reply With Quote
Old 11-17-2015, 04:32 AM   #2
gringer
David Eccles (gringer)
 
Location: Wellington, New Zealand

Join Date: May 2011
Posts: 836
Default

Q1/2 looks very similar to what we've done with our RNASeq experiments. Mike Love seemed to think the approach was correct:

https://support.bioconductor.org/p/74458/
gringer is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 08:08 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2019, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO