Ah I missed that aspect in my skim. The consistency excess copies seems to imply that losing them entirely would be extra bad for survival enough so that redun

Ah I missed that aspect in my skim. The consistency excess copies seems to imply that losing them entirely would be extra bad for survival enough so that redundant copies would boost offspring survival and dominate for human adjacents but not their more wild counterparts. The domestification might even imply reverse order - that the ones with the starch genes were more likely to wind up in "dog" pools - say because they were attracted to starch and got closer to humans - and thus plenty of extra copies among their descendants. Meanwhile ones without were more likely to stay away and their descendants more likely to wind up in "wolf" pools where prevalence was lower.