Replies to post #63473 on DNAprint Genomics (DNAG)

[thesciguy]

Frog-

I think we are onto another topic here. My original question/comment was on the mechanisms that underlies the use of various forms of genetic markers (AIM's --I believe-- in the Ancestry tests) to determine placement into certain human population sub-groups (outwardly labeled bio-geographically). Those markers can only be used to increase or decrease the likelihood of actually sharing ancestry with a group (can be very high, but never 100%). Even DNAPrint's Retinome test can only boast a 0.9 probability of determining iris color. (Careful use of the word "only" in this case).

It's important to remember that these techniques are not actually reading sequenced genes. They are reading scattered letters throughout the genome to look for similarities that correlate with something of interest (drug response, eye color, ancestry, etc). Sequencing a genome takes more time, more money and more analysis.

It's similar to being told that the 25th, 495th 644th and 1889th letters of a book are a,v,f, and y and determining that the book is The Catcher In The Rye. Add to this analogy, that every copy of every book ever published contains random errors (not just spelling) and the limitations of the technique becomes apparent, and the need to find other, complementary tools, becomes urgent (especially when the cost becomes human life). Extend this analogy to include that books are descended from one another to add another layer of complexity.

If the susceptibility/responder issue in question can be causally associated with a known genetic variance there is no need for any other screening method. Simply screen for the variance and get on with it. Get on with what? It’s not that simple. It depends on the strength of the association. If an individual is shown to carry markers that 60% of negative responders also carry (via a single test), that patient, and doctor, would need to weigh the costs against the benefits of treatment. The results of this cost/benefit analysis, would be different for different diseases and different patients and therefore would make the strength of association a moving target.

I think we are splitting hairs as I think you know this... By extracting the bio-geographical trends from a population of responders, it is SOMETIMES possible to determine a general locus for the UNKNOWN variance.

Good Luck!