PI_HAT stands for “proportion IBD hat.” It is a statistical measure used in genetics to estimate the proportion of alleles shared identical by descent (IBD) between pairs of individuals. The PI_HAT value ranges from 0 to 1, where 0 indicates no alleles shared IBD and 1 indicates complete allele sharing IBD. It provides an estimate of the genetic relatedness between individuals and is commonly used in studies of population genetics, pedigree analysis, and genetic association studies. In the context of paternity testing, the PI_HAT value is often used to assess the probability of a father-child relationship by estimating the degree of genetic similarity between them. Higher PI_HAT values indicate a greater likelihood of a biological relationship, although additional factors and analyses are typically considered for conclusive paternity determination.
Using PLINK
We can determine the genetic relatedness of two dogs by processing their Embark Vet lab results using a free tool called PLINK. The PLINK commands we’ll use look like this:
plink --dog --noweb --tfile alleged_father --make-bed --out alleged_father_data
plink --dog --noweb --tfile son --make-bed --out son_data
plink --dog --noweb --bfile alleged_father_data --bmerge son_data.bed son_data.bim son_data.fam --make-bed --out merged_data
plink --noweb --bfile merged_data –genomeThe Results
The resultant plink.genome file created by the final PLINK command will contain several columns of data. For our purposes, the PI_HAT, Z0, Z1 and Z2 are of interest. A PI_HAT value above 0.7 is a likely father-child relationship, whereas a PI_SCORE in the mid-range of 0.5 is unlikely. In a paternity scenario between a father and a child, you would typically expect to see the following ranges of values for the Z0, Z1, and Z2 columns in the PLINK .genome file:
Z0: The Z0 column represents the probability of sharing zero alleles IBD (identical by descent) between the father and son. In a father-child relationship, you would expect a relatively low value for Z0, indicating a low probability of not sharing any alleles IBD. This value is usually close to 0 or small compared to Z1 and Z2.
Z1: The Z1 column represents the probability of sharing one allele IBD between the father and son. In a father-child relationship, you would typically expect a moderate value for Z1, indicating a moderate probability of sharing one allele IBD. This value could vary but is generally lower than Z2.
Z2: The Z2 column represents the probability of sharing two alleles IBD between the father and son. In a father-child relationship, you would expect a relatively high value for Z2, indicating a high probability of sharing both alleles IBD. This value is typically the highest among Z0, Z1, and Z2.
Setting-up Your Computer
Here are the steps I use to run PLINK against Embark Vet data.
Start by creating three folders:
c:\usr\bin
c:\usr\var
c:\usr\var\paternityNext, download the PLINK 1.9 binaries and unzip them to c:\usr\bin.
Now, download the RAW DATA from Embark Vet for the alleged father and the son. Unzip them in c:\usr\var. Each .zip file will contain a .tped and a .tfam file. For the alleged father, rename the files to alleged_father.tped and alleged_father.tfam. For the son, name them son.tfam and son.tped. Copy the 4 files to c:\usr\var\paternity.
From a command prompt, change to the c:\usr\var\paternity directory. You’ll have to prefix each PLINK command with c:\usr\bin\ or add that location to your PATH environment variable. Here again are the commands.
c:\usr\bin\plink --dog --noweb --tfile alleged_father --make-bed --out alleged_father_data
c:\usr\bin\plink --dog --noweb --tfile son --make-bed --out son_data
c:\usr\bin\plink --dog --noweb --bfile alleged_father_data --bmerge son_data.bed son_data.bim son_data.fam --make-bed --out merged_data
c:\usr\bin\plink --noweb --bfile merged_data –genomeFinally, analyze the plink.genome file with a text editor or import it to Excel.
