Skip to content

[Bioc-devel] differential expression tools for proteins

5 messages · Cardin Julie, Diego Morais, Wolfgang Huber +1 more

#
Hi,
   I have experienced very good results with DESeq2 for my RNASeq analysis. As far as I understand, it is a tool that normalise our data from sequencing to make them comparable.

I have a new project implicating proteins counts.
I have  couple of data sets. For each sample we have:
rows with proteins names (instead of genes), with their respective counts.

My goal is again to make a differential expression between treated groups versus controls.
I wonder if I can use DESeq2 to do a differential expression for proteins? 
Or if the correcting factor that is used by DESeq2 to correct counts for RNASeq is specific to DNA sequencing and it is not applicable to proteins?

Is there a tool that do the exact same thing as DESeq2 but for proteins?

Thank you very much for your help and time,
Best regards and happy new year!

Julie Cardin
Bioinforamatician
IRCM
#
Hi Cardin,
this package does not do the same as DESeq2,
http://bioconductor.org/packages/3.7/bioc/vignettes/transcriptogramer/inst/doc/transcriptogramer.html,
but you can use it to do a differential expression for proteins (case x
control).

2018-01-06 14:45 GMT-03:00 Cardin Julie via Bioc-devel <
bioc-devel at r-project.org>:

  
    
#
Dear Julie

Thank you for your question. It would however be more appropriate for 
the support forum, not for the developer mailing list. Would you mind 
moving it there, perhaps also the responses so far?

There is no "in-principle" reason why DESeq2 shouldn't produce useful 
results also for count data from technologies that are not 
DNA-sequencing based. It's error model (Gamma-Poisson, GP) is quite generic.

As always, you should do model fit diagnostics though, to see whether 
the residuals for each protein across replicates and conditions (after 
fitting the GLM) are reasonably consistent with the GP, in particular, 
that they look unimodal.

One issue to check is also whether the normalization (size factors) is 
appropriate.

There is another bit of irony afaIu: If you have enough replicates (or: 
degrees of freedom) that you can actually "see" deviations from the GP 
assumption (i.e. >=dozens), then you probably don't need a parametric 
method, and could switch to something non-parametric.

Kind regards
		Wolfgang

6.1.18 18:45, Cardin Julie via Bioc-devel scripsit:

  
    
#
7.1.18 12:46, Wolfgang Huber scripsit:
I just saw you posted on the forum, after browsing it.
I had searched for your name using the "search" function of the forum 
webpage before sending the previous, but that did not turn up this post.

Wouldn't it be great if our search box worked better? Or would just 
convert the search into an URL like

https://www.google.com/search?q=Cardin+site%3Asupport.bioconductor.org&sort=date

Wolfgang
#
On 01/07/2018 08:13 AM, Wolfgang Huber wrote:
use the 'Users' and 'Tags' links at the top right to navigate users and 
tags.

Poor search on the support site is a long-standing and well-known issue; 
some reasons (are they still valid?) for not simply using google are at

 
https://github.com/Bioconductor/support.bioconductor.org/issues/4#issuecomment-56036398

where one can also access the code and propose pull requests.

Martin
This email message may contain legally privileged and/or...{{dropped:2}}