Harmonisation and strand flip by PRSice

[jianvhuang]

Hi

I wonder how PRSice deals with inconsistent strand and allele coding between the base GWAS and the target sample. I found a statement from the webpage that PRSsice performs strand flip automatically. Does it also deal with complimentary allele coding (e.g., EffectAllele.base="C", NonEffectAllele.base="T", A1.target="G", A2.target="A")?
In this example, I suppose no flipping of beta is needed.

It will be very helpful if you could elaborate what PRSice checks for this QC step.
Thank you very much.

Regards,
Jian

[choishingwan]

Yes. We check for complement too. We flip by changing the allele dosage. 0 1 2 to 2 1 0

…

On Wed, May 3, 2023, 3:42 AM jianhuang ***@***.***> wrote: Hi I wonder how PRSice deals with inconsistent strand and allele coding between the base GWAS and the target sample. I found a statement from the webpage that PRSsice performs strand flip automatically. Does it also deal with complimentary allele coding (e.g., EffectAllele.base="C", NonEffectAllele.base="T", A1.target="G", A2.target="A")? In this example, I suppose no flipping of beta is needed. It will be very helpful if you could elaborate what PRSice checks for this QC step. Thank you very much. Regards, Jian — Reply to this email directly, view it on GitHub <#318>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJTRYVTREU36U3NYENNBVLXEID7XANCNFSM6AAAAAAXUA3QUQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[jianvhuang]

Do you think this could be very different from flipping the effect allele and non effect allele (along with the +/- sign of beta) in the base GWAS?
Could you point out which part of the code in the below file deal with the harmonisation issue? I would like to learn more about this. Thank you very much.
https://github.com/choishingwan/PRSice/blob/master/PRSice.R