I have GWAS data for analysis. I have two groups, control and case, with 80 and 100 in each group respectively. I have done all the required QC steps, after which I am remained with around 300,000 SNPs out of 1,100,000. Now, I am plotting QQ plot for my data where I am in dilemma whether to keep df=1 or df=2. If I use df=1, the graph shows distribution line of P-values very much near to the middle line with few points towards end scattered substantially, so I am considering this as ideal. But with df=2, the P-value distribution line has significant deviation from middle line towards Y-axis, with again, few points towards end scattered substantially.

Can anybody suggest me what degree of freedom should I use? And, what my both of the graphs, i.e. with df=1 and df=2 are suggesting?