GWAS results and linkage disequilibrium issue

Question

[seta]

Dear all,

Regarding the GWAS results, assuming there are 10 significant associated SNPs with a given trait that are in high LD. If all significant associated SNPs should be reported or just those are not in LD? I saw both in the various papers, what's your idea about it?

Also, regarding the risk score calculation, if all significant associated SNPs should be considered for risk score calculation or just those are not in LD?

Many thanks in advance

[Nicolas Rosewick]

IMO You should try to use a fine-mapping tool e.g. FINEMAP in order to get the causal SNP. Then use this SNP as reference SNP and report all SNPs in the neighborhood that are significant and their associated LD R2 with the reference SNP. Also I would use only the putative causal SNP for the risk score.

[seta]

Thanks for your comment; could you please let me know how successful is finding the causal SNP for complex diseases? Regarding risk score calculation, however, the other risk SNPs (not necessarily causal, but associated or in high LD) existed and have their own effects, so why we don't consider them for risk calculation?