bvilhjal / ldpred

MIT License
96 stars 57 forks source link

what does effalt in STANDARD really refers to? #133

Open mathieu-lemire opened 4 years ago

mathieu-lemire commented 4 years ago

Could you confirm what the column effalt in the STANDARD format really refers to?

STANDARD format: chr pos ref alt reffrq info rs pval effalt

Just from the choice of the label that you used, I would think it is the effect size of the alt allele, but looking at the code in sum_stats_parsers.py , you defined in there A1 as ref, A2 as alt and eff as effalt.

But your --help option mentions the following (if one wants CUSTOM):

--A1 A1 Column header containing the effective allele. There isn't any standardized label for the effective allele, therefore extra care must be taken to ensure the correct label is provided, otherwise, the effect will be flipped. --eff EFF Column header containing effect size information

Despite your choice of the word "effective", I feel confident you meant that A1 is the effect allele (meaning that EFF is the effect size of the A1 allele)? I that what you meant?

If so, then is effalt from STANDARD really the effect size of the ref allele?

rkarlssonlinner commented 4 years ago

This is further confused by the explanation of this flag: --reffreq REFFREQ Column header containing the reference MAF

As far as I know, it can be important to differentiate between the AF of the reference allele (which is not necessarily the minor allele), or the MAF of the SNP. I am not yet sure whether it is the AF of the reference allele or the MAF of the SNP that is sought as input here, and whether this difference actually matters for this particular analysis?