Allele 1 reference vs effect allele

[brde...@colorado.edu]

Hello,

Please excuse me if this is a rather naive question, but I just wanted to clarify the specification of Allele 1 in the summary statistic file.  For my GWAS analyses the minor allele was used as the effect allele (PLINK analyses).  My sumstats file includes signed z-scores.  For my sumstats file I would then assume that Allele 1 would be specified as the minor allele, which was tested as the effect allele in PLINK.

However on this page of the LD score GitHub   I see that A1 is assumed to be the reference allele. Since my effect allele is the minor allele should I still specify it as A1?

I am hoping to use the hap map 3 SNPs to filter down my sumstats file with munge_sumstats.py as suggested by the tutorial. In addition I am hoping to use the reference European LD scores already provided for my LD scores. To use these, I want to make sure that coding my minor-allele as the effect allele A1 will not cause any problems.

Thank you so much!

Britt

[Raymond Walters]

Hi Britt,

It’s probably clearer to say that ldsc assumes A1 is the effect allele (the “reference allele” label has unfortunately acquired multiple meanings, and is used inconsistently even within the plink documentation), so plink’s output A1 is indeed the correct A1 to input for munge_sumstats.py.

Thankfully the impact of getting this wrong would be fairly manageable. For "munge_sumstats.py --merge-alleles”, the order of A1/A2 doesn’t matter as long as the pair is consistent with the HapMap alleles. For heritability estimation, the A1/A2 alleles have no impact. For genetic correlation, reversing your A1/A2 will flip the sign of the correlation, but the magnitude of the correlation and it’s p-value will be unchanged. (And depending on the strength of your prior expectations for the correlation the sign flip may be a sufficient red flag to diagnose that A1/A2 were incorrectly reversed).

Cheers,

Raymond

--
You received this message because you are subscribed to the Google Groups "ldsc_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ldsc_users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ldsc_users/f364614c-0317-4b98-9a87-75bdef682d10%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[brde...@colorado.edu]

That makes sense, thank you so much for your thorough explanation!

Britt