When using biallelic markers, the GS models estimate an allele subsitution effect for each marker ("alpha", which is the difference between the average effect of the 2 alleles).
When using multiallelic markers, what is the value estimated for an allele. Is it its average effect ?