Incorrect allele ordering in plink #1315
Description
See sgkit-dev/bio2zarr#391 for details.
It's less obviously incorrect here than in bio2zarr where we're striving for compatability, but it's still a surprising choice which should be documented if there's a good reason for it.