How to select SNPs for haploblock identification after GWAS analysis?

Question

[rimgubaev]

Hello everyone! I made a GWAS analysis and found several SNPs that are located close to each other (less than 0.5 megabases). Then I want to find haploblocks in order to make haploblock-trait association. In case If I want to find haploblocks that are associated with the trait of interest should I use only significantly associated SNPs in order to find haploblocks (in Haploview) or the other nearly located SNPs should also be included in the analysis?

Answer 1

[Kevin Blighe]

Hey,

No, for HaploView, use all of your SNPs that have passed quality control checks. HaploView will then determine haploblocks and haplotypes based on whatever measure / metric that you select. You can note the haploblocks in which your statistically significant SNPs lie.

As far as I am aware, HaploView provides functions that also then permit you to perform association tests based on the defined haploblocks / types.

Kevin

[rimgubaev]

So If I understood it right I should first identify haploblocks and only then look if the haploblocks identified by Haploview contain my significantly associated SNPs?

[Kevin Blighe]

That is what I am saying - yes. If you just use the statistically significant SNPs, then the data would likely be too sparse and you would be biasing the haploblock analysis. This does not, however, prevent you from pre-filtering the input SNPs based on quality control measures.