GWAS result too significant

Question

[yehchan57]

Recently, I did some genome-wide association study employing PLINK1.9. However, when I analyze the result of association I found it too 'significant'. This must be false! My command is

 ~/Software/plink --vcf qz_beijing_wes.g.vcf.gz --snps-only --maf 0.001 --hwe 0.0001 --geno 0.1 --make-bed --out qz_beijing_wes

~/Software/plink --bfile qz_beijing_wes --assoc fisher --allow-no-sex --pheno pheno.txt --pheno-name CaseControl --out qz_beijing_wes_fisher

But the result shows too many p value less than 10^-20 !

[Nicolas Rosewick]

could you do a qq-plot ?

[yehchan57]

qqplot just like this

Answer 1

[chrchang523]

—assoc has been obsolete for many years. See https://www.cog-genomics.org/plink/2.0/assoc#glm for some discussion of how the “Eigenstrat” association analysis is performed with PLINK today.