Search
Question: How to carry out a gene set analysis using custom gene sets?
0
gravatar for SamGG
2.2 years ago by
SamGG90
France
SamGG90 wrote:

Hi,

I would like to carry out a gene set enrichment analysis using gene sets designed by an expert (ie lists of carefully chosen genes). The tools are numerous and the turnover is fast. So I appreciate any clue.

Best.

ADD COMMENTlink modified 2.2 years ago by Gordon Smyth32k • written 2.2 years ago by SamGG90

Your comment about "numerous tools" makes me wonder whether you are looking for an initial suggestion or whether you already know of a number of tools and want to choose between them? If the latter, then we'd need to know more about what tools you are considering and what you want to achieve in your study.

ADD REPLYlink modified 2.2 years ago • written 2.2 years ago by Gordon Smyth32k
3
gravatar for Gordon Smyth
2.2 years ago by
Gordon Smyth32k
Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia
Gordon Smyth32k wrote:

We designed and use roast and barcodeplot (part of the limma package) for this specific purpose.

Here are some examples of use. Each of these papers includes at least one barcodeplot and accompanying roast test for a custom gene set:

 http://www.ncbi.nlm.nih.gov/pubmed/25973911 (figures 2 and 4)

 http://www.ncbi.nlm.nih.gov/pubmed/24939936 (figure 5)

 http://www.ncbi.nlm.nih.gov/pubmed/25605372 (figure 6)

 http://www.ncbi.nlm.nih.gov/pubmed/25655195 (figure 2)

ADD COMMENTlink modified 2.2 years ago • written 2.2 years ago by Gordon Smyth32k

Thanks for the links.

ADD REPLYlink written 2.2 years ago by SamGG90
0
gravatar for svlachavas
2.2 years ago by
svlachavas570
Greece/Athens/National Hellenic Research Foundation
svlachavas570 wrote:

Dear Samuel,

the term "gene set" enrichment" is very general and it varies depending on the specific experimental design and the goals of your study, and also of the "kind" of the test you want to perform(i.e. hypergeometric, topology based, self-contained null linear models by mroast/roast etc.).

You can check the following review (http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1002375) and also the course material from bioconductor (http://bioconductor.org/help/course-materials/2013/EMBOBGI/GeneSetEnrichment.pdf), which adress in details the specific modules. Just to pinpoint, from personal experience and from literature, the Molecular Signatures Database from the Broad Institute has excellent and curated gene sets regarding various field(TFs, cancer etc)(http://www.broadinstitute.org/gsea/msigdb/collections.jsp#C6), as also the web-tool Enrichr (http://amp.pharm.mssm.edu/Enrichr/).

Best,

Efstathios

ADD COMMENTlink modified 2.2 years ago • written 2.2 years ago by svlachavas570
1

I chose a general term purposely in order to get as many information as possible. Your given links are relevant. The MSIGDB is a good resource, but in order to reduce the multiple test correction, targeted or ad hoc geneset collection is valuable. The C6 might be an interesting compromise. C6 contains some curated signatures from publications. Nevertheless, IMHO, cancer signatures have a lot of common genes that puzzle the interpretation. Thanks for your comment and the links, especially Enrichr which is very rich.

ADD REPLYlink written 2.2 years ago by SamGG90

Well, it depends on your experimental analysis and study goals. Anyway, from MSIGDB you could subset a general category to include only some releveant parts(for instance for C2, you could keep only to test for the KEGG pathways).

ADD REPLYlink written 2.2 years ago by svlachavas570
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 192 users visited in the last hour