Genotype changes in GWAS are ambiguous

[olivierlabayle]

The genotypes in a GWAS are ambiguous because they encode the number of A2 allele in the BED file (see https://openmendel.github.io/SnpArrays.jl/stable/) which could be either the minor or major allele. I see two options to remove any ambiguity:

• Evaluate and arbitrate the impact of string representation on memory footprint (This is preferred because this would make the output of TarGene consistent). I suspect a 6 folds increase based on local example.
• Make the 0, 1, 2 consistent with a minor allele count (use theminorallele function in the above package to provide the proper encoding).