Hello everyone,
I have a gene set of ~1600 gene symbols that I have retrieved from NCBI and would like to analyse for enrichment of immunological function/pathways.
I have tried using both DAVID and Panther, but the difficulty I am having is figuring out which of the gargantuan list of terms that is returned relate to immunological function (some are obvious i.e. explicitly to do with B-cell function, others are terms that I don't even know where to begin with).
Is there any way I can filter such an analysis so that it only looks for enrichment of a certain type (immunological in this case)?
And if so, is this something that can be done in DAVID or Panther? Or would another tool be more appropriate?
Apologies if the answer is something obvious (I am very new to this) - I have spent several days reading and aren't any closer to an answer.
I ended up using GOrilla which worked quite well. Worth a look.
Also, MSigDB has a set of immunologic signatures (c7), so you may want to try out GSEA:
That looks very interesting indeed - nice to see it also has an R interface, so I won't have to use any unfamiliar programming languages. I'll take a look, thank you.
On looking into this, it seems to want to compare array data to one of their gene sets - I can't seem to find any way to coerce it into comparing my gene set to one of theirs.
If you are working with HUGO gene symbols, you don't have to specify a chip design. You just have to set "Collapse dataset to gene symbols" to 'false'.
However, if you don't actually have any gene expression data, then that is indeed a problem. GSEA is designed to work with intensity / expression values, so that it doesn't depend on any upstream differential expression analysis.
Actually, GSEA is designed of users without any programming experience. So, I would typically recommend using the GUI over the R interface (that is really for users who need to design a pipeline to handle 100s of different projects). Hopefully, this can be useful for future analysis.
Yes, this is the case. I don't have expression data. I have a list of genes that were selected based on population differentiation statistics. The hypothesis is that population A and B should be strongly differentiated at immune genes, A and C should be moderately differentiated, and B and C should be differentiated relatively little. So I have obtained a list of genes that differentiate in the manner predicted, and am now testing these lists for enrichment for immune system related terms. I am having two main troubles. 1) Finding a test for enrichment that is relatively concise (i.e. lumps immune function together in a convenient way) and 2) Finding a test that can compare multiple gene lists against one another.
Log in to answer this question.
different GO levels have different level of details. for example GO-biological process level 5 for some gene might say B-cell activation whereas level 3 annotation for the same gene might say immune response.
So if you want a broad classification go for GO level 2 or 3
Oh, I see! Very helpful, thank you. I'll just have to figure out how to adjust which level I am viewing then. Yes, you are quite right, it is really on the broad strokes I am looking for, it is a fairly crude analysis.