Phenotypic Variation Explained By Snps In Gwas

Question

[kumar.vinod81]

Thanks Genotypes group for really making it clear in a simple way. It was also interesting. I also want to ask one more question, how to calculate the variation explained by all the significant SNPs in combined, because tool I am using only calculating it for only individual SNP basis for a particular trait but some paper mentioning that all the significant SNPs contributed this % of variation for this trait. Can you please explain. Thanks,

Vinod

Answer 1

[Joey]

Which paper? Unless you mention the paper, it is very hard to figure out exactly what they did.

Just a guess: Do you know what the significant SNPs are for your analysis? If yes, why don't you make a dataset with the significant genes only and then use the GCTA s/w (http://www.complextraitgenomics.com/software/gcta/) to calculate phenotypic variation?

[kumar.vinod81]

HI Joey, I am not getting how to calculate contribution of all the significant SNPs to a particular trait. Below some papers mentioned such analysis http://www.nature.com/ng/journal/v42/n11/full/ng.695.html#/supplementary-information http://www.nature.com/ncomms/journal/v2/n9/full/ncomms1467.html
I've almost done all of the analysis like significant SNPs for all of the traits but I am totally confused at this step on SNP contribution to phenotypic variation. Can you please help? Thanks,

[Tm]

Hello vinod,

Were you able to find the answer to your question? if yes, can you please give detail of the tool and steps used for the same?

[kumar.vinod81]

you can fit multiple linear regression in R by fitting all the significant SNPs (vector in R) as explanatory variable while the trait of interest (vector in R) as a response variable.

[Tm]

Thank you for the reply. But, my main concern is obtaining the phenotypic variance (var %) using the gwaa.R (GenABLE) result which I have for each trait separately. It contains 5 columns as shown below:

• SNP
• Estimate
• Std.Error
• t-value
• p-value

i.e I got the list of significant SNPs aasociated with particular trait using gwaa.R which uses GLM approach but, now I am interested in knowing whether these significant loci responds to that particular trait or not. Thus want to get phenotypic variance (R2). So which package can I use further to reach my goal? Please correct me if you feel I am on wrong track.

[kumar.vinod81]

Really didn't understand, exactly what you want to do? In GenABEL, its really tough to get r2 directly. You can follow this link http://forum.genabel.org/search.php?keywords=phenotypic+variance+explained+by+SNP for more clarity. Or you can deleted the most significant SNP and then see the difference in h2 between whole set and deleted SNP set. Thanks,