I run gwas for 7 traits of plant leaves, with 886930 snps, after that I got result, in which p-values and FDR adjusted P-values are given. Now i want set a threshold. Let suppose 0.05, to declare significant snps. I did it by bonferroni method, which is too conservative, and the threshold would be 0.05/886930 =5.63742E-08. And non of my snps are passing this threshold. I ranked the p-values for smallest to largest to calculate q-value. But after doing all this, i still have confusion about, how to select a threshold. Can anyone elaborate, how can i select my threshold? i want to select a threshold, above which i can declare significant snps. Another question is, what if we select a threshold randomly. Let suppose, my threshold by using statistical methods, such as bonferrroni and FDR is 10-8, and i didn't find any snps above this threshold. Now let's suppose, i declared a random threshold 10-5, and took all those snps which have a threshold of 10-5, and peformed variant annotation. I found some downstream gene variants, and extracted the genes, in which those snps are present. I selected those genes and search for homologs. I found some homologs in other crops, also I checked expression of those genes in RNA seq data of those crops, which are highly expressed in leaves, and my l traits are also related to leaves. Then i checked, the expression of those genes in my crop, and found those genes highly expressed in leaves. As i declared a threshold randomly and by this way i found the genes which are expressed in leaves, and my traits are related to leaves. Will it be a wrong way? Can i declare a thresold this way?
