SNP panel interrogated by KING for estimating kinship?

Question

[gaelgarcia]

Hi,

I'm trying to figure out what panel of SNPs KING interrogates to calculate kinship, but this doesn't seem to be explicitly stated anywhere.

Does anyone know what this panel is?

Thanks.

Answer 1

[chrchang523]

All of them. A great strength of the method is that it gets the job done almost regardless of what SNPs you give it, as long as they cover the genome pretty well. (Though you can certainly speed it up, and maybe improve accuracy a little bit, by choosing an appropriate subset.)

[gaelgarcia]

Thanks, @chrchang523. By < covering the entire genome very well > , do you mean as long as they're spread out across the genome fairly evenly? We have sequencing data of a [very] targeted panel of genes across ~15K samples for which we're trying to do pedigree/sex checks.

I'm trying to determine how many sites we actually have within this data that will be useful for this purpose, but I'm not sure what is the best way to go about asking that question. i.e., should we look at the overlap of our targeted regions with all registered SNPs?

[chrchang523]

Doesn't need to be that even, as long as they aren't heavily concentrated on just a few chromosomes.