I am currently working on a project using GENESIS. I have used the PLINK .bim, .fam and .bed files to create the .gds file as suggested by the software. I have also run fitNullModel and assocTestSingle successfully. It is stated in GENESIS that the effect size estimate (“Est”) is for each copy of the alternate allele.

My question is: which one should be considered the alternate allele in the output? Is it column 5 in the .bim file in accordance to the PLINK2 format of “When no header line is present, the columns are assumed to be in .bim file order (CHROM, ID, CM, POS, ALT, REF)”, or is it different for GENESIS?

Thank you for your time, Artemis

The documentation should say, "the effect allele is the alternate allele for SeqVarIterator objects and the 'A' allele for GenotypeIterator objects." I will update that soon.

If you are using SNPRelate::snpgdsBED2GDS to convert your PLINK files, and then creating a GenotypeIterator object using the GWASTools package, then assocTestSingle is testing the "A" allele, which will be column 5 in your .bim file. snpgdsBED2GDS reads the first allele as column "A" and the second allele as column "B", and stores genotypes as a matrix of counts of the "A" allele. These functions were written for array data in the days before PLINK files had any notion of "ref" or "alt". You can check the character values of alleles with functions getAlleleA and getAlleleB.

If you are using SeqArray::seqBED2GDS to convert your PLINK files, and then creating a SeqVarIterator object using the SeqVarTools package, the ref and alt alleles are currently reversed. I just opened an issue for SeqArray, so hopefully this will be fixed in the future (https://github.com/zhengxwen/SeqArray/issues/64). Currently, you can work around this problem by following the note in the PLINK manual about recoding your .bim file before conversion (https://www.cog-genomics.org/plink/2.0/formats#bim). You can check the character values of alleles with seqGetData(gds, "$ref") and seqGetData(gds, "$alt").