How --genome calculates its results

837 views
Skip to first unread message

Rocio Mariana

unread,
Mar 26, 2019, 3:06:53 PM3/26/19
to plink2-users
Hi,

I was wondering if someone could help me to learn how does plink calculate some values for the --genome IBD relations.

    FID1    IID1    FID2    IID2   RT     EZ      Z0         Z1        Z2    PI_HAT   PHE      DST      PPC     RATIO
    0001    0001   0002   0002  UN    NA  0.8157  0.0180  0.1663  0.1753    -1     0.867248  0.3245  1.9714


This is a header and one example of a result of --genome in a relatedness.genome file.

I notice that PI_HAT, that is the "Proportion IBD" is calculated as: P(IBD=2) + 0.5*P(IBD=1), then it would be 0.1663 + 0.5*0.0180 that results on 0.1753.
Since, Z0 P(IBD=0), Z1 P(IBD=1) and Z2 P(IBD=2), my question is, how this values are calculated, and what does IBD=0 means?

Thanks!

Christopher Chang

unread,
Mar 26, 2019, 11:48:12 PM3/26/19
to plink2-users
This is explained in detail in the original PLINK paper (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1950838/ ); start reading from "We use a method-of-moments approach"...

Baraka Mkumbe

unread,
Jul 21, 2022, 4:27:53 AM7/21/22
to plink2-users
How can I identify identical twins in my samples? I have already checked for relatedness using --genome cmd in plink and obtained the z0, z1,z2 and pihat value

Christopher Chang

unread,
Jul 21, 2022, 10:32:44 AM7/21/22
to plink2-users
The currently-recommended method is to use plink 2.0's --make-king-table command, which uses a formula that is more reliable than --genome's at identifying close relations.  E.g.

plink2 \
  --bfile <your fileset prefix> \
  --make-king-table \
  --king-table-filter 0.354 \
  --out identical_twins

(See the third paragraph of the linked documentation for an explanation of the 0.354 cutoff.)

Baraka Mkumbe

unread,
Jul 21, 2022, 11:39:09 PM7/21/22
to Christopher Chang, plink2-users
Thanks Christopher

--
You received this message because you are subscribed to a topic in the Google Groups "plink2-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/plink2-users/3_5QNEe9kac/unsubscribe.
To unsubscribe from this group and all its topics, send an email to plink2-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/plink2-users/a539a2d6-f7d1-4570-811e-ceb65c680ba0n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages