How Do You Explain The Correlation Over A Gene Set And Pathways And Other Ideas To Deal With The Gene Set?

Question

[J.F.Jiang]

Hello,

I am doing a job of post-GWAS analysis, and i have obtained an important gene set containing about 200 genes.

i want to give some detailed analysis to explain the relationship between genes and disease from pathway level.

traditianal analysis always use gene enrichment to look those important genes overlaped in KEGG database, then elucidate the importance of these enriched pathways. (Because i am not very good at pathway level analysis, i think it is too superficial to deal with the gene set i collected. Maybe there is any other better or deeper method to deal with these genes from pathway level? (not to construct a method.)

1. could you offer me some suggestions? or some related references to deal with?
2. And beside pathway level analysis, any ideas for dealing with this gene set in order to connect genes to disease? it will be very appreciate of you for offering the ideas!

And another question is i get some pathways from enrichment is KEGG. But i do not know whether the pathways i get is really harmful to the disease. So how can i confirm or validate the primary results? my idea is to look at pubmed using those pathway names as critical words?

Thank you !@@

[Larry_Parnell]

Your actual question "While i am interesting..." could be better formulated, but I gave it try to offer some ideas.

[Giovanni M Dall'Olio]

Sorry, I don't understand the question..

[J.F.Jiang]

Thank you for your reply, i edit the question, hope to make it clear.

Answer 1

[Larry_Parnell]

I suggest that you take a look at the latest paper from Suhre et al on GWAS of metabolite levels. This is Suhre Gieger 2011 Nature 477:54. The table they present on the biological role or implication of the findings will give you some ideas - such as, for example, the GWAS hit is in the enzyme responsible for the rate-limiting step in pathway X. This type of explanation is readily understandable. In other words, I think it would help to know the bottlenecks and branch points of your pathway. When GWAS signals (association with biomarker phenotypes or eQTL type associations) fall into those genes, then information flow is either disrupted or enhanced (bottleneck) or shunted in a different direction (branch point). This is how I as a biologist would think of interpreting the pathway hits.

[J.F.Jiang]

thank you! i will carefully read this paper and continue my study with your suggestions! thank you again!

Answer 2

[ff.cc.cc]

Good reference Larry, I found it useful (+1).

I suggest to check www.babelomics.org too, the online server is sometime slow, but it is more oriented to functional analysis with respect to GSEA, DAVID...

Besides the related publication - http://nar.oxfordjournals.org/content/38/suppl_2/W210.short - could give you suggestions on how to setup a wokflow for annotating your results

[Larry_Parnell]

There are too many deadlines at work for me to go into detail. Basically, I let the biology lead me, then carefully check alternate possibilities to rule them in or out. Maybe I can add more later...

[J.F.Jiang]

yes, great comment. I just want some workflow of how to analyse my gene set to disease. Because i do not have any datsets of SNP genotypes or gene expression, (and WTCCC do not offer the raw data of this disease), so i use snp location map to find those genes with some published method.

But after obtaining the gene set containing 200 genes, i have no ideas how to give an extra and sufficient explanation.

Any way, thank you guys, great suggestion! And hope can communicate more~~