SEQanswers

Go Back   SEQanswers > Bioinformatics > Bioinformatics



Similar Threads
Thread Thread Starter Forum Replies Last Post
Need help with DESeq2 KYR RNA Sequencing 13 03-11-2019 11:51 AM
Deseq2 luana RNA Sequencing 3 10-08-2016 12:38 AM
DESeq2 Simon Anders Bioinformatics 123 07-06-2015 01:45 AM
Example DESEQ2 David Lyon Bioinformatics 4 05-18-2015 04:28 AM
DESeq2 normalization tellsparck Bioinformatics 3 07-15-2014 03:10 PM

Reply
 
Thread Tools
Old 03-27-2017, 08:00 AM   #1
SDPA_Pet
Senior Member
 
Location: US

Join Date: Apr 2013
Posts: 222
Default DESeq2 questions

Hello, I have a function data sets. I want to look at which functions are over-representative. However, I am only interested in a subset of functions for example DNA synthesis.

Can I just use the functions related to DNA synthesis (subset table) to do the DESeq2 analysis or I have to use the whole dataset and later found which over-representative functions relative to DNA synthesis?

Thanks,

Last edited by SDPA_Pet; 03-27-2017 at 08:05 AM.
SDPA_Pet is offline   Reply With Quote
Old 03-27-2017, 08:11 AM   #2
GenoMax
Senior Member
 
Location: East Coast USA

Join Date: Feb 2008
Posts: 7,077
Default

Are you mixing up differential expression analysis (DESeq2) with GO term enrichment analysis (e.g. GeneSCF)?
GenoMax is offline   Reply With Quote
Old 03-27-2017, 08:15 AM   #3
SDPA_Pet
Senior Member
 
Location: US

Join Date: Apr 2013
Posts: 222
Default

Quote:
Originally Posted by GenoMax View Post
Are you mixing up differential expression analysis (DESeq2) with GO term enrichment analysis (e.g. GeneSCF)?
Well, I don't think so, because DESeq2 can give you log2 fold change of each functions and whether the change significant or not.

Last edited by SDPA_Pet; 03-27-2017 at 08:18 AM.
SDPA_Pet is offline   Reply With Quote
Old 03-27-2017, 01:46 PM   #4
wdecoster
Member
 
Location: Antwerp, Belgium

Join Date: Oct 2015
Posts: 97
Default

Quote:
Originally Posted by SDPA_Pet View Post
Can I just use the functions related to DNA synthesis (subset table) to do the DESeq2 analysis or I have to use the whole dataset and later found which over-representative functions relative to DNA synthesis?
You should use the whole dataset and later on check if the genes up and down regulated correspond to a certain function.
wdecoster is offline   Reply With Quote
Old 03-27-2017, 05:58 PM   #5
gringer
David Eccles (gringer)
 
Location: Wellington, New Zealand

Join Date: May 2011
Posts: 838
Default

DESeq2 expects raw read-level count data as input. I don't expect that it would be possible to get that from a function data set.
gringer is offline   Reply With Quote
Old 03-28-2017, 05:21 AM   #6
SDPA_Pet
Senior Member
 
Location: US

Join Date: Apr 2013
Posts: 222
Default

Quote:
Originally Posted by gringer View Post
DESeq2 expects raw read-level count data as input. I don't expect that it would be possible to get that from a function data set.
Hey guys, I am sorry about the confusion. The dataset that I use the gene counts dataset. (The gene coding for the functional genes).

So, should I use the subset or total dataset.
SDPA_Pet is offline   Reply With Quote
Old 03-28-2017, 05:37 AM   #7
GenoMax
Senior Member
 
Location: East Coast USA

Join Date: Feb 2008
Posts: 7,077
Default

@SDPA_Pet: I think you have metagenomics data that you have somehow mapped as counts --> a function (so not a specific gene). Is that interpretation correct? That may be important for other posters to know.

In any case cherry picking data is not the way to go with DESeq2.
GenoMax is offline   Reply With Quote
Old 03-28-2017, 05:41 AM   #8
SDPA_Pet
Senior Member
 
Location: US

Join Date: Apr 2013
Posts: 222
Default

Quote:
Originally Posted by GenoMax View Post
@SDPA_Pet: I think you have metagenomics data that you have somehow mapped as counts --> a function (so not a specific gene). Is that interpretation correct? That may be important for other posters to know.

In any case cherry picking data is not the way to go with DESeq2.
Right? It's kind of tricky? In our field, we the conception of gene/functions are exchangeable.Anyhow, it is just counts table.

Just curious,GenoMax? What is difference between the gene table and function table in your field (I would guess your field is biomedical field)? Do you have an example?
SDPA_Pet is offline   Reply With Quote
Old 03-28-2017, 07:38 AM   #9
GenoMax
Senior Member
 
Location: East Coast USA

Join Date: Feb 2008
Posts: 7,077
Default

DESeq2 is not operating on the identifiers but is considering the counts associated with them.

Code:
DNA_synthesis      50
TCA_Cycle        100
OR
Code:
BRCA1      50
TRPV1        100
should produce equivalent results from DESeq2.
GenoMax is offline   Reply With Quote
Old 03-28-2017, 07:50 AM   #10
SDPA_Pet
Senior Member
 
Location: US

Join Date: Apr 2013
Posts: 222
Default

Right, my questions about the dataset choosing.

In you example, let say DNA_systhesis total have 50 counts. I can go deeper hierarchical level
DNA systhesis

sub Function 1 5
sub function 2 3
...

TCA total 100
sub Fucition 1 5
sub fuction 2 3

....

I could choose two way to do this. First, use the total datasets which is 150 functions total, and look at which sub function related to DNA synthesis is overabundant.

Or, I could only extract the 50 subfunctions from DNA synthesis and run DEseq to find out which one is overabundant.

To me, I only care about one large category. However, the sample size wouldn't be same in the two different analyses

Quote:
Originally Posted by GenoMax View Post
DESeq2 is not operating on the identifiers but is considering the counts associated with them.

Code:
DNA_synthesis      50
TCA_Cycle        100
OR
Code:
BRCA1      50
TRPV1        100
should produce equivalent results from DESeq2.

Last edited by SDPA_Pet; 03-28-2017 at 07:52 AM.
SDPA_Pet is offline   Reply With Quote
Old 03-28-2017, 09:01 AM   #11
GenoMax
Senior Member
 
Location: East Coast USA

Join Date: Feb 2008
Posts: 7,077
Default

You may want to post this question on Bioconductor DESeq2 support site where people with the statistical chops will answer (someone on here may do that as well).
GenoMax is offline   Reply With Quote
Old 03-28-2017, 09:02 AM   #12
SDPA_Pet
Senior Member
 
Location: US

Join Date: Apr 2013
Posts: 222
Default

Quote:
Originally Posted by GenoMax View Post
You may want to post this question on Bioconductor DESeq2 support site where people with the statistical chops will answer (someone on here may do that as well).
Can you give the link? I google it and find different sites. Not sure which one is the authentic
SDPA_Pet is offline   Reply With Quote
Old 03-28-2017, 09:22 AM   #13
GenoMax
Senior Member
 
Location: East Coast USA

Join Date: Feb 2008
Posts: 7,077
Default

Quote:
Originally Posted by SDPA_Pet View Post
Can you give the link? I google it and find different sites. Not sure which one is the authentic
Here is the bioconductor support site. Tag your question with deseq2.
GenoMax is offline   Reply With Quote
Old 03-30-2017, 10:38 AM   #14
SDPA_Pet
Senior Member
 
Location: US

Join Date: Apr 2013
Posts: 222
Default

Quote:
Originally Posted by GenoMax View Post
Here is the bioconductor support site. Tag your question with deseq2.
Hi GenoMax,

Thanks. BTW, do you know is there any good forum to discuss R application. Bioconductor.org focus on molecular bioinformatics. I need some forums such as help people with plots (ggplots 2), multivariate statistics, etc.
SDPA_Pet 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 06:00 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO