counting number of individuals the possess a mutation in 2 sites

[Caterina]

Hi I wanted to do something like this:

../plink --bfile sample --extract mutations.txt --recode A --out filtered_sample

where mutations.txt are just 2 SNPs. 

I would like to obtain the number of individuals with the 2 mutations at the same time, but would also like to know if they're homozygous or heterozygous.

Unfortunately I get this error:

Error: --recode does not support multipass recoding of very large files.

I imagine because it shows me the data of each individual. Is there anyway I can just get the count or frequency instead, but with the homozygous or heterozygous information?

Thank you!

[Christopher Chang]

1. Please post a full .log file, rather than just the error message, when asking for troubleshooting help.

2. plink 1.9 --freqx and plink 2.0 --geno-counts provide this information.

[Caterina]

All right. Thank you!

And how could I measure the number of people that have 2 mutations at the same time?

[Caterina]

Never mind I guess it would be just multiplying the frequencies of mutation in SNP1 and SNP2.

[Christopher Chang]

Actually, it would not, especially if the two genomic positions are very close to each other.

plink 1.x's --twolocus command provides one simple way to get these SNP1 x SNP2 counts.

[Caterina]

I see, thanks again! And how could I count SNP1xNOT_SNP2? Basically frequency of people with SNP1 but NOT SNP2? Or would twolocus give me that information as well?

[Christopher Chang]

--twolocus would also give you that information.

[Caterina]

What if I don't have the variant ID, only the position? What would be the syntax? Or it isn't possible to do something like that?

[Christopher Chang]

In that case, you can first use plink 2.0's --set-all-var-ids flag to change all the IDs to be position/allele-based.

[Caterina]

Thank you! Last question, I promise. What do I do if my loci are in 2 different chromosomes and my PLINK files (.bim, .fam, .bed) are separated by chromosome. Do I have to merge them or is there another way?

[Christopher Chang]

You have to merge them to use --twolocus (or extract just those SNPs first and merge them, etc.).