PLINK: filter for indels

Question

[bsmith030465]

Hi,

How do I filter my plink files so that only indels are kept? I used the following command to keep only snps:

plink --bfile myfile --snps-only --out myfile.snps

Which command/option should I use to retain only indels?

Thanks

Answer 1

[chrchang523]

If your variant IDs are unique, this can be done in two steps: use --write-snplist + --snps-only in the first step, and then run --exclude on the .snplist file in the second step.

[bsmith030465]

Since my variants IDs weren't unique, I did some file processing to my snps output file (myfile.snps above) and got the snp ids that I want to exclude ('snp.pos.txt'):

1 10101 10101 10101
1 10493 10493 10493
1 10560 10560 10560
1 11994 11994 11994
.
.

and then I ran

plink --bfile myfile --exclude snp.pos.txt --out myfile.indels

However, this doesn't seem to exclude the positions. Is the formatting wrong? Each field is separated by whitespace as recommended in PLINK documentation...

[chrchang523]

1. Variant IDs cannot contain spaces, so your snp.pos.txt really does not look like it does what you want it to do. plink 2.0's --set-all-var-ids flag provides a way to assign new position-based IDs to all your variants, if it's okay to throw away the old IDs.

2. You need --make-bed in your last command to generate a new filtered fileset.

[bsmith030465]

I am trying to exclude the 'range' of snps with this file format.

plink --bfile myfile --exclude snp.pos.txt --range --out myfile.indels --make-bed

but it says " Note: --range flag deprecated. Use e.g. '--extract range [filename]'."

It also gives an error if I use:

plink --bfile myfile --exclude range snp.pos.txt  --out myfile.indels --make-bed

I don't want to use plink2, because I can't seem to do other things (e.g. --mendel )

How should I set my options with plink 1.9?

Many thanks for all your help. I appreciate it!

[chrchang523]

1. You can use plink2 for just —set-all-var-ids, and return to plink 1.x for the remaining steps.

2. The problem with “—exclude range” here is that, when a SNP and an indel start from the same position, it’ll get rid of both; —set-all-var-ids at least makes it possible to get rid of only the indel in this case. With that said, don’t worry about the Note for now.

[bsmith030465]

Thanks for the help and explanation. That worked! Much appreciated.