I am following the quality control steps in this tutorial for GWAS:

• https://github.com/JoniColeman/gwas_scripts/blob/master/README.md.

My dataset is from a family study design with:

VCS[:number:] is the family ID
FID_1 is the ID of the child
FID_2 is the ID of the mother
FID_3 is the ID of the father

In the Pairwise identical-by-descent (IBD) check, I used plink --genome command some of my result are as follow:

FID1    IID1    FID2    IID2    PI_HAT
VCS043  VCS043_1    VCS101  VCS101_1    0.5
VCS043  VCS043_2    VCS101  VCS101_1    0.5
VCS043  VCS043_1    VCS043  VCS043_2    0.5
VCS043  VCS043_1    VCS043  VCS043_3    0.5
VCS043  VCS043_2    VCS043  VCS043_3    0.5

Which problem can make the pihat to be exactly 0.5?

When I filter the result with PI_HAT = 0.5. Every pairwise PI_HAT of either VCS043_1 or VCS043_2 are exactly 0.5. Meanwhile, other PI_HAT in the result are good: pairwise PI_HAT of parents <0.05; pairwise PI_HAT of child-parent ~ 0.4-0.6 (1st degree relative)

To be more clear, an example of a family with identical twin (_4); result of plink --genome looks like this

FID1    IID1    FID2    IID2    PI_HAT
VCS027  VCS027_1    VCS027  VCS027_2    0.4803
VCS027  VCS027_1    VCS027  VCS027_3    0.4816
VCS027  VCS027_1    VCS027  VCS027_4    0.9698
VCS027  VCS027_2    VCS027  VCS027_3    0
VCS027  VCS027_2    VCS027  VCS027_4    0.4794
VCS027  VCS027_3    VCS027  VCS027_4    0.4835