Removing duplicate SNP based on allele

[R Stephanie L]

Hello,

I am trying to produce a .raw file with the following command. However, I get an error as there is a duplicate SNP with different alleles (C-G and C-T). The alleles that I require based on the GWAS I want to use are G-C.

How do I keep the SNP with G-C allleles only?

plink --bfile ukbb_edn_snps --remove ${HOME}/Data/UKBIOBANK/Exclusion_participants/data_removed_dec2020.txt --reference-allele "{HOME}/Data/Discovery_samples/Edn/ukb_education_reference_list.txt" --recode A include-alt --make-bed --out ${HOME}/Data/UK_Biobank/GeneticData/ukb_edn

Error: Duplicate ID 'rs7040995'.

TIA

[Christopher Chang]

You need to assign the SNPs different IDs, with e.g. plink 2.0's --set-all-var-ids flag.

[dodomba]

when we --set-all-var-ids with allele names , there's a problem because after merging theres many snps need to be flipped 

but now its snp name has changed what do you suggest to do ?

[Christopher Chang]

This question doesn't make any sense.

Please include a full .log file(s) illustrating exactly what you're talking about in all future posts.

[dodomba]

yeah , I fixed that issue , but I have last question , which is more general.
I wanted to merge two datasets but the first has SNPs ids that are not in rs ..i also downloaded its conversion file to rs ids.
but i don't know whether to update id names or to just rename all snp ids to chr and position , which would give me more overlapping between datasets when merging?
there's also the question whether to rename snp ids by chr and position  or chr/position and unique Allele like this 1:513251 vs 1:5235823,A,C

tell me your opinion on this and specially when do we choose to rename snp id by unique allele or just chr/pos ? in regards to overlapping and later issues like positions with same id etc...

[Christopher Chang]

plink is a tool.  You are responsible for knowing what you want to do with the tool.  You should not be using plink if you cannot reason out the answers to these questions yourself.