Eqtl Data For Post-Gwas Analysis

Question

[J.F.Jiang]

Hello, everyone;

i am recently collect some sources of eqtl data from published papers.

when i came into detail look about the results, i am doubt that why do results differ so much among those papers?

dixon (nature 2007 mRNA browser）: lots of pages including more than 200000 snp-gene relations;

while stranger (nature genetics 2007): less than 8000 snp-gene relations;

and much more like this,

since almost all results do not contain much information, i do not find possible results i want to collect.

my question is since the genotyped ids are not much differed, why results of these papers differed so much?

To my results, ony 3 papers offer lots of snp-gene information!

Answer 1

[Sean Davis]

You might take a look here:

http://eqtl.uchicago.edu/Home.html

But, in general, you will need to read the papers you named in your question to determine why they have different numbers of putative eQTLs.

[J.F.Jiang]

thank you for reply, i have checked the link you offered before. In order to finish my object, i have to collect much more eqtls for disease-snps overlap. During data collecting, i found those problems above. And found few overlap between database by references and snps i collected.

[Shicheng Guo]

What's the difference between eqtl.uchicago.edu and GTEx etql? Thanks.

Answer 2

[Larry_Parnell]

There are many reasons why the results of eQTL analyses differ. As Sean writes, you will need to read the papers to understand this (+1). Some general reasons are:

1) Different cell types tested; 
2) Different conditions in which the cells grow or are treated; 
3) Cells from different genetic backgrounds; 
4) Different statistical methods employed to determine significant SNP-mRNA level associations; 
5) Use of different genotyping and/or gene expression plpatforms could point to different SNPs or different mRNA isoforms as involved in the SNP-mRNA level association.

There are many other papers offering SNP-mRNA association data. See those by Schadt et al for human liver and adipose. See the UChicago server Sean mentions. See GenVar at the Wellcome Trust. See papers by Spielman,Cheung et al (Nat Genet. 2007 39:226-31) and Zhang, Dolan, et al (Am J Hum Genet. 2008 82:631-40). Dig deeper - there are more!

[Larry_Parnell]

In terms of connecting your SNPs to the eQTL SNPs, do not forget that your SNP may be in strong LD with the eQTL SNP. Thus, you need to look at your SNP + all the proxy SNPs (in LD) to see if any support eQTL associations.

[J.F.Jiang]

that's a great suggestion, the reasons you offered is very helpful. And i will check it later. i have read much more papers than you offered. But since these papers do not offer much snp-mrna associations, i have failed to collect those genes associated with my snps.