Question about re-coding SNPs in PLINK

Question

[vickygarfield87]

Hi there,

I'm working on a blood pressure genetics project and have got a bit stuck. I am trying to analyse data to see if there are genetic associations between systolic and diastolic bp and several outcomes. I have external weights from the published literature to make my genetic profile scores in Plink. However, some of the effects/betas are associated with an increase in bp, whilst others are associated with a decrease in bp. Someone has suggested that I re-code alleles in the SNP data, so that they are 0,1,2 increasing blood pressure, rather than leaving them coded as the minor allele. However, I'm not sure exactly what this means, in terms of actually re-coding in any of my Plink files. I don't know which Plink file I'd be recoding the 0,1,2 in and what exactly to...

Any tips/advice would be much appreciated!

Thank you!

Vicky

Answer 1

[chrchang523]

While this is technically possible with plink, it's a pain because you'll have to constantly specify --keep-allele-order or --reference-allele whenever you want to use the recoding; otherwise plink will keep forcing A1 = minor allele and A2 = major allele.

Meanwhile, the allelic scoring function lets you specify which allele each score is associated with, and negative scores are permitted, so I'm confused re: what the purpose of recoding is supposed to be here.

Answer 2

[vickygarfield87]

Thanks for the reply. I think I've managed to sort it out by basically changing the sign of the effect and also flipping the alleles...it's a bit confusing, but the idea is to make a genetic score/instrument that accounts for both increase/decrease in blood pressure, as there is a mixture of effects associated with both...I've tried it in different ways: external weights, internal and unweighted, and now have to compare the results to see whether flipping alleles/signs makes a difference or not.