Re: how clump works in plink

[Chris Chang]

r^2 is undefined if one of the variants is monomorphic.  In the .clumped file you posted, 14:30847595:C:T is clumped with nothing at all, so I'm guessing it is monomorphic in the clump-reference dataset you are using (despite not being monomorphic in the GWAS dataset).

On Wed, Nov 15, 2023 at 7:03 AM landscape95 <trminhth...@gmail.com> wrote:

Hi, I am doing clumping with the follow command:

        --clump-snp-field SNP \
        --clump-field P \
        --clump-p1 5e-8 \
        --clump-r2 0 \
        --clump-kb 250 \

I set radius = 250 and r2=0,. To what I understand, pilnk will harvest all SNPs around the index SNPs at radius 250 but from the result, I have 2 SNPs with distance < 250kb, i.e. 14:30847595:C:T and 14:30864928:T:A (17kb)

My plink version: PLINK v1.90b7.1 64-bit (18 Oct 2023)

Can I confirm what I understand is correct and how could this result happen?

Thank you very much!

--
You received this message because you are subscribed to the Google Groups "plink2-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plink2-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/plink2-users/dac0bcb4-ce82-4072-94c3-d6fb932dfbf6n%40googlegroups.com.

[landscape95]

Thanks so much Chris,

How about 14:30818761:C:T and 14:30877360:G:C?

Each of them are not monomorphic and they are clumped with several SNPs.

Thank you!

[Christopher Chang]

You can use --ld to look more closely at the r^2 computation --clump performs between two specific SNPs.