On Wednesday, August 30, 2017 at 3:37:10 PM UTC-7, taf wrote: > On Wednesday, August 30, 2017 at 2:28:15 PM UTC-7, wjhonson wrote: > > > But you can identify which parent was the source of a snippet of DNA > > Not sure what you're saying there. > > > > The entire point of Autosomal is that you can identify from where a > > snippet comes. > > Only if you have both parents' data already in hand, and even then population homozygosity means that for most of the DNA you may not be able to tell anyhow, not that it matters. > > I see little point in continuing this. Your grand plan for DNA-based medieval genealogy is both impractical and impossible. > > taf So you are claiming, against all actual data, that testing an *aunt* can never be any sort of subsitute whatsoever for testing a *mother*. That in other words, the DNA you share with an Aunt says nothing at all about your Mother? If that is what you are claiming, I'm at a loss for words.

On Wednesday, August 30, 2017 at 4:25:21 PM UTC-7, wjhonson wrote: > On Wednesday, August 30, 2017 at 3:37:10 PM UTC-7, taf wrote: > > On Wednesday, August 30, 2017 at 2:28:15 PM UTC-7, wjhonson wrote: > > > > > But you can identify which parent was the source of a snippet of DNA > > > Not sure what you're saying there. > > > > > > The entire point of Autosomal is that you can identify from where a > > > snippet comes. > > > > Only if you have both parents' data already in hand, and even then population homozygosity means that for most of the DNA you may not be able to tell anyhow, not that it matters. > > > > I see little point in continuing this. Your grand plan for DNA-based medieval genealogy is both impractical and impossible. > > > > taf > > So you are claiming, against all actual data, that testing an *aunt* can never be any sort of subsitute whatsoever for testing a *mother*. > > That in other words, the DNA you share with an Aunt says nothing at all about your Mother? > > If that is what you are claiming, I'm at a loss for words. The *entire basis* of Lazarus is to try to reconstruct the DNA of an ancestor, and now you're claiming that that approach is utterly without merit?

On Wednesday, August 30, 2017 at 4:25:21 PM UTC-7, wjhonson wrote: > > Only if you have both parents' data already in hand, and even then population homozygosity means that for most of the DNA you may not be able to tell anyhow, not that it matters. > > > > I see little point in continuing this. Your grand plan for DNA-based medieval genealogy is both impractical and impossible. > > So you are claiming, against all actual data, that testing an *aunt* > can never be any sort of subsitute whatsoever for testing a *mother*. > > That in other words, the DNA you share with an Aunt says nothing at > all about your Mother? > > If that is what you are claiming, I'm at a loss for words. You just insist on completely ignoring the effects of the information loss in each generation. To give a trivial example, if your aunt has blue eyes, and your mother had brown, you can deduce absolutely nothing about what was going on with your mother at that locus. True, you would say this is only one locus and there are a million more, but the same applies to 1/2 of the total of your mother's DNA, and the next generation in each line the same happens, so first cousins share only 1/8, and second cousins only 1/32. The loss of information is dramatic and progressive and debilitating to the analysis you propose, and all of the genomes in the world won't overcome this. taf

On Wednesday, August 30, 2017 at 4:25:21 PM UTC-7, wjhonson wrote: > So you are claiming, against all actual data, that testing an *aunt* can > never be any sort of subsitute whatsoever for testing a *mother*. > > That in other words, the DNA you share with an Aunt says nothing at all > about your Mother? > > If that is what you are claiming, I'm at a loss for words. Let's take a close look at this example. If there is a living person X, whose parents are both dead, and each parent has one living sibling available for testing, what can we learn by testing the the individual, their aunt on one side and their uncle on the other side. First the father's side. The aunt would have shared 50% of the DNA of her brother, and the child would only have 50% of the DNA of the father, so they would share 25% of the DNA of their aunt. That means that 25% of the DNA of their father would have been determined and the other 75% unknown. So lets say you then test a second cousin, the next closest relation to the father. They will share, on average, 3% of their DNA, but a quarter of that will be the same part shared wth the aunt, so we now have a little over 27% reconstructed. Even if there are 10 second cousins, that still doesn't even get you to 50%, and it gets pretty asymptotic if you go any more distant. So, in one generation we have lost 50% of the information. Not all generations will be so limited, with one or two siblings, but even losing only 20% per generation, you will be down to minuscule levels of identifiable DNA before you hit colonial times, and there will be nothing that can be traced from medieval times. taf