Replies to post #15252 on DNAprint Genomics (DNAG)

[Mike Tiernan]

Yes. The thing is, the Witness update didn't make clear whether they could just distinguish primary ancestry or if they could have percentages as well. It is one thing to know if a person is major ancestry European _or_ Middle Eastern, which is of interest to forensics. It is much more difficult to quantitate this and say that such and such a person is 55% European and 45% Middle Eastern, which would be of interest to genealogy.

Likewise, a person may be majority Northeast Asian _or_ Southeast Asian, but what about mixtures?

I think that Witness 2.5 is capable of making the crude yes/no decisions on majority ancestry, but maybe they are not quite there yet about the percentages.

Another point is that Gabriel's comments didn't really fully distinguish between ABD and ABD Witness.

I don't know what's going on there. If they are having trouble getting enough samples for a wide range of parentals, they could ask for volunteers from, for example, ex-2.0/2.5 customers. In exchange for some cheek swab dna, they could promise a discount when the product is ready.

Of course, they'd need to somehow confirm that a person's reported ancestry is actually what it really is, but that's another story.

Population genetics studies even from 10 years ago (e.g., Cavalli-Sforza's book) clearly show that you can distinguish between many, many ethnic groups and sub-groups using autosomal markers. The "catch" is those studies are done on large sample sizes, and the ABD test needs to be accurate on single individuals.

That's a big challenge, and I do admire DNAP for trying to get it done.

I just wish they'd be a little bit more, let us say, innovative and enterprising about how they are getting it done. DNAP is located in a nation (the USA) that has many people of just about every ethnic group on Earth. All the possible parentals are there. There are many people of mixed heritage - a perfect sample population to test the accuracy of the test to detect admixture. The USA is also home to many universities, with many potential collaborators.

Making a test that can distinguish a Swede from a German from a Greek from an Iraqi from a Pakisani from a Japanese from a Chinese from an Ethiopian from a Nigerian, etc etc is possible. The just need to work hard AND work smart.

The second is more important.