Hi,
I hope you can provide me with some insight into these 2 questions.
Context:
I have a GWAS of a relatively isolated East Asian population. GWAS shows a significant SNP1 associated with this cellular phenotype (E-12). SNP2 is the second top signal (E-07) and is in relatively high LD with SNP1 (LD~ R2=0.78, D'=1). Both SNP1 and SNP2 have positive beta associated with this cellular phenotype.
Conditioning analysis on SNP1 shows an independent second signal (SNP3-SNP7, with high/perfect LD between them) and SNP2 pvalue is E-02 with *negative beta*. Then, I conditioned on SNP2, so pvalue of SNP1 is ~E02 and positive pvalue. This suggests that SNP1 & SNP2 are most likely transmitted together but I can not explain the change in the direction of effect.
- Can someone tell me if they have encountered something similar? How can this be explained?
From all SNPs in that gene, I found SNP8 whose pvalue is E-03 with positive beta. Conditioning on SNP1 or SNP2 leads to an increase of the SNP3 signal (pvalue=E-06) and a change in the direction of effect (negative). Conditioning on the SNP8 leads to a huge increase of the SNP1-SNP2 pvalue (~E-15) as well for SNP3-7 (~E15-E-12). SNP1-SNP2 are in strong LD with SNP3, ~r2=0.8.
- Is this SNP8 an independent signal? is SNP8 a repressor of SNP1-2 and SNP3-7? how can the change of the effect be interpreted?
Thanks a lot for any advice.
Note: Conditioning of SNP3 seems to increase the signal of SNP1-2 or SNP8. So these SNP3-7 seem to be an independent signal.
