- Joined
- Sep 8, 2006
- Messages
- 9,262
- Reaction score
- 3,969
I know... I was hoping it would be a little more official since it is from VIN and all. I did find it helpful that there were some known breed mixes being tested which isn't something you run across very often.
FWIW, my lab is developing a similar panel for cats, only it's more focused on 'regional' population origins rather than breed origins since 99.9% of 'breed unknown' cats are randombred DSH/DMH/DLH with absolutely no purebred origins - though we also have some markers that we use for breed. We did decide, however, that in addition to markers for people asking about specific breeds, we would have a "secondary" panel that also looked at breed functional mutations, IF the primary panel identified a major breed influence in the mix.
One of the criticisms of the breed testing is that it forces decisions to be made about breed ancestry based on the markers - it makes assumptions that a purebred animal existed in the animal's bloodline in the last 3 generations, which may not even be true for many shelter mutts. That was kind of the impetus for doing the 'regional' thing for us. The clustering algorithms really, really want to cluster, it's what they're meant for. So we tried to give some 'catch-all' sort of categories to cluster into that weren't necessarily breeds, but say 'European randombred' or "African randombred' or whatnot.
It was a fun thing to work on for us, we may or may not actually market it but do plan on being honest about the limitations and such as we should be. Anyway, we did test it on our colony cats of known mixed ancestry and it actually came out pretty good for us.