How to do gene enrichment analysis?

Question

[michealsmith]

Very dump question: I'm wondering how to do gene enrichment analysis? I'm simply doing in a very very simple way: copy-paste my gene list into gene-ontology website, which is now powered by PANTHER. ( http://geneontology.org/) Then just try three different categories: biological processes, molecular functions, cellular components.

I mean just that simple? Anything else should I do? More tricks?

Also any other recommendations for other gene enrichment tools in addition to gene-ontology?

================================ Thanks. Another confuse is there are many different tools and websites, so we tried different tools and report the one that is most statistically significant? Kind of cheating?

[russhh]

Most of these tools work using hypergeometric statistics. If you have any biases in your 'gene-list-generating' workflow (eg, due to gene length or expression level : hello somatic mutation callers) these gene enrichment tools will give you another level of bias into your analysis. Hopefully that isn't an issue, but I can't quite tell from your question

[michealsmith]

Thanks. This is also another question. I'm working on brain genes, which are usually large. I don't know if GO take account of gene size?

[WouterDeCoster]

No, it doesn't. But depending on your analysis prior to enrichment analysis you will probably have normalized your data.

Answer 1

[WouterDeCoster]

Have a look at enrichr: http://amp.pharm.mssm.edu/Enrichr/

Same principle, copy paste a list, but enrichment versus diverse definitions of pathways and sets.

Answer 2

[####]

Try this GSEA http://software.broadinstitute.org/gsea/index.jsp

Answer 3

[biostart]

There was a good discussion recently here: GO analysis: DAVID vs GREAT vs GOrilla