updating .bim with new rsids

Question

[hi.there]

Hello,

I am a novice with plink and working with a .bim, .fam and .bed data with outdated rsids for chrom position. With help with the UCSC genome browser, I have a new .txt file in hopes of updating the rsids by chrom position:

system("./plink --bim oldBim --update-name onlySNPs.uniqLocAndld.txt 2 4 --make-just-bim -out newBim")

.... Unfortunately, I am not getting any updated rsids (it runs through with zero updates). Would anybody know what I may be doing wrong?

The new onlySNPs.uniqLocAndld.txt was generated following these commands to get uniqueness:

curl -O https://hgdownload.gi.ucsc.edu/goldenPath/hg38/database/snp151Common.txt.gz
gunzip -c snp151Common.txt.gz | cut -f 2,4,5,12,17 | grep single.exact | cut -f 1-3 > onlySNPs.tsv
sort -k3 -u onlySNPs.tsv | sort -k1,2 -u > onlySNPs.uniqLocAndId.tsv 

Any help would be greatly appreciated.

Answer 1

[chrchang523]

--update-name requires both old and new IDs; I'm guessing the file you provided did not include the old IDs, so plink didn't know what to update.

plink 2.0's --set-all-var-ids flag is useful here for forcing the old IDs into a position-based format that's easy to generate in the --update-name file.

[hi.there]

I see. Thank you. I was under the impression that the end chromosome position in the .txt file was all that was needed to search to replace the rsid with --update-name with correct fields "2, 4" specified. I suppose that isn't correct.

I have updated to plink 2.0 but the documentation for --set-all-var-ids is unclear. How would one use it in this instance? Also, for future reference of "--update-name", how would I reformat my onlySNPs.uniqLocAndld.txt file in order for the --update-name command to work?

[chrchang523]

You should take a step back and learn how to use some Unix text-processing tools. The most relevant ones are "cut", "paste", "head"/"tail", and to a lesser degree "sed" and "awk" ("sed" and "awk" are somewhat complicated, but you only need to be aware of their existence and learn a few simple usage patterns for now). I can see from your question that you should already be aware of "cut".

After you've done this, you should be able to come up with a coherent way to use --set-all-var-ids in this context.

Answer 2

[hi.there]

Thank you again. I will review my my unix commands. Are there any additional resources besides the plink documentation (link below) where I can get a better feel for how --set-all-var-ids is used?

https://www.cog-genomics.org/plink/2.0/data

There are plenty of biostar posts where --set-all-var-ids is recommended but I could not find any committed lines and the documentation explanation of how to use "\$r,\$a" and "@:#" is unclear to me in this usage case.

[chrchang523]

Just try a few things out on a small dataset? This isn't a complicated command. (Technically, you don't even need it; you can use a Unix one-liner to do the same thing.)

[hi.there]

Okay. Thanks. Perhaps tinkering with the command will help me learn it.