Hello, 

Some of my significant DE gene lists obtained by "Limma" contain far more than 3,000 genes. I tried to use the RDavidWebService through  "BACA" or "clusterProfiler" to retrieve functional annotation (e.g. KEGG).  Unfortunately,  I received error messages due to the size of my genelists as indicated on Davis website : "A job with more than 3000 genes to generate gene or term cluster report will not be handled by DAVID due to resource limit" ). 

I  am wondering what R packages or web resources would be a good alternative for "RDavidWebService".

Your advice is highly appreciated.

Thank you,

Anita 

Do you have Entrez Gene IDs for your genes? If so, you can simply use the kegga() function in limma to do the KEGG functional analysis.

Suppose that 'fit' is your fitted model object and you are testing for the second coefficient. Then you can call kegga() from the fit object using something like:

k <- kegga(fit, coef=2, geneid="EntrezID", species="Hs")
topKEGG(k)

This assumes you have human data and that the Entrez Gene IDs are contained in fit$genes$EntrezID.

You will need the very latest version of limma for this to work, so make sure you are using R 3.2.0 or 3.2.1 and type biocLite("limma") at the R prompt. You will also need to install the KEGGREST package.

As an aside, given the very large number of DE genes, you might consider using the treat() method in limma to prune the list down a bit to those with strongest effects.

Hi Anita,

Unfortunatelly, DAVID's webservice api has the 3000 genes limit and 200 jobs per day. So, if you really need DAVIS's results, you can upload the genes manually in the website page and download the results into a file. Then, you can import them using RDavidWebService and continue with your pipeline.

Regards,

Cristobal

PS: let me know if other functionality needs to be develop and I'll update the package.

Thank you for advice. I was concerned about the large number of genes.

I will try both functions (treat and kegga).

Thanks once again,

Anita 

hi Alakatos  , just additional tip ... 
   Once you get the KEGG  P.value scores from limma or others ...  check the dynamics & visualising network  in the top X KEGG  pathways by  using  great libs:
1) http://www.bioconductor.org/packages/release/bioc/html/pathview.html


2) http://www.bioconductor.org/packages/release/bioc/html/KEGGprofile.html

For me one important property is that you can visualize all samples on 1 Pathway.

Best,
Branko
 

Thank you for all the great suggestions and help!  

Anita