Hi Doug, You have really put some hours into analyzing the data I personally am a big fan of Gedmatch I think it is clear that Family Finder has some shortcomings and you have effectively illustrated that Gedmatch is a very useful tool as well. I think in some ways Gedmatch is superior to Family Finder.
Antonio On Friday, September 12, 2014 2:13:15 PM UTC-7, Doug da Rocha Holmes wrote: > > Thanks to several private exchanges, I thought it might be interesting to > illustrate one example of the use of Gedmatch. > > Fellow list member Antonio Faria, whose ancestry stems from Lajes do Pico > on both sides of his family, long ago tested for Y-DNA and he has written > about it countless times on this list how it leads straight to Baltazar > Vieira born in the middle 1500s and from Sao Mateus, Pico. Decades before I > met Antonio, I knew Baltazar Vieira was also my own direct ancestor and the > fact Y-DNA testing revealed his J haplogroup which means he has direct male > line Jewish ancestry was simply fascinating to me. But for years I wondered > exactly how Antonio descended from Baltazar, because my ancestry is from my > paternal grandmother whose parents were both born in Sao Roque do Pico and > I had already traced all families there and still had no idea how Antonio > connected to Baltazar. So last year we finally made contact and found that > it was his grandfather or great-grandfather from Sao Roque who moved over > to Lajes in the late 1800s. Then I was able to know who our most recent > common ancestor on that line is. It turns out that on that line it takes 8 > generations back to reach that ancestor and 9 generations for me to reach > the same ancestor, making us 7th cousins, once removed. > > This means that his father and my father are 6th cousins, once removed > from this particular ancestor. > But there is another line which takes 6 generations for his father and 6 > for my father to reach, making our fathers 5th cousins. > > These two proven lines (with many more remote connections I won't name) > are not enough for FTDNA to show that our fathers are related. Sure, one > could say that it's simply because one or both of them didn't get the DNA > from that ancestor. So that's the basis of my liking Gedmatch so much to > look at DNA matches from all angles. > > Just now I ran a comparison between my father's kit and his father's kit. > I set the cM level down as low as it goes which is 1 and here are the > results: > > There are 6 separate segments of common cM on 6 different chromosomes (5, > 6 11, 14, 16, 17). Two matches are quite small in the 1 range, three are in > the 2 range and the largest is a 9.5 cM segment on Chr 17. The total of all > segments equals 20.5 cM. > > Genealogy research has proven we're related. Then DNA testing has > confirmed it as above. Yet they don't show as a match on FTDNA. > > So now we compare Antonio's father to one of my cousins. My cousin's > grandmother was my father's mother, so my father is an uncle here. And all > this DNA is coming from my father's mother with all Pico ancestry. If she > were still alive to test, she would surely match about double the number on > her side of the family that my father matches, just like my father matches > about double the number on his side that I match. > > First of all, Antonio's father shows as a match on FTDNA to my cousin, > unlike my father. So the matching DNA my grandmother passed down to the > parent of my cousin was greater than the matching DNA my father got from > her. This illustrates why it's best to get all the older generations tested > first (today if you can), but do not neglect to get all your cousins > tested, as well as all your siblings. > > I ran the two kits (Antonio's father and my cousin) through the one to one > DNA comparison with Gedmatch and found the following common DNA: > > Chromosomes 1, 3, 6, 10, and 14 all had at least 1cM of shared DNA > segments of various sizes. The total came to equal 20.9 cM. Notice how the > total for my father was 20.5. The smallest segment was in the 1 range, two > segments are in the 2 range, one in the 3 range and the last and biggest > segment is in the 11.4cM. That larger amount was enough for FTDNA to list > it as a match whereas the 9.5 segment for my father was too low to reach > the minimum threshold. > > I'm certainly no scientist, but I think that the two totals of 20.5 and > 20.9 cM are significant. It's telling me that my grandmother had no less > than 20.9 cM of this same matching DNA and could be way more, like 40 cM. > But it's impossible to test that. > > The way I interpret it is that the matching DNA from my grandmother was > pretty much equally passed down to my father and his sibling, then passed > to my cousin. Too bad we can't test my father's sibling to know exactly how > much matching DNA they had, but no matter what, my cousin got a tiny bit > more of it and chances are pretty good there was more of that matching DNA > in the parent. > > But the reason FTDNA called a match in one case and not the other is the > size difference (11.4 verses 9.5). It is my understanding that this is only > part of the equation when deciding who is a match and who is not. The total > cM is the other part of the equation and since that is about equal in both > kits, the only reason one is a match is the size is just a bit higher for > one segment. > > But it's the randomness of DNA passing from generation to generation that > accounts for the reason someone in a later generation shows as related when > the closer generation (my father) doesn't show. By chance my cousin got a > segment that recombined to form a larger segment that pushed them over the > threshold and my father didn't, but got one additional segment showing up > on some other chromosome. > > This is not the case, but let's assume we had just the two above common > ancestors I mentioned where one line is a 5th cousin match and the other > line is 6th cousins, once removed. Let's further assume that DNA for both > sets of matches reached my grandmother. In fact, maybe DNA from one is > completely missing, but for now let's say she got about 15cM from each > ancestor and that the total cM she had equaled 30. She might have passed on > about 10cM from each common ancestor so that my father got just a little > over 20cM and my cousin got nearly the same. > > If there were two larger segments of about 10cM each representing these > two common ancestors, I think my father would have shown up as a match. But > by chance his 20cM was divided into smaller segments with only one larger > segment that fell below the threshold. Yet my cousin, by chance, had a > little bit larger segment that reached that threshold and ding, ding, ding, > we have a match. > > In reality, there are many more ancestors we share that are quite remote > from the 1600s and 1500s that probably account for some of these segments. > Maybe the ones in the 1cM range are from these more remote ancestors. That > throws a wrench into the above calculations, but that's just why it's more > complicated than my above explanation. > > And that is also why FTDNA wants to error on the side of less matches than > reality because it's hard to predict how segments all end up generation > after generation. Same for any scientist and that's why they operate on > hypotheses. They can't make any progress if they don't look at all the > possibilities. And with their reputations on the line, they don't want to > say something now that will be disproved next year, even if their hunch is > right. > > But the data I have to work with, like above, is giving me a different > conclusion than it's giving to FTDNA. They are not studying the same > families I am. And it takes someone with a very complete family tree plus > testing of many relatives to make these conclusions. I would be flattered > if someone at FTDNA were to take my observations and add their vast > additional expertise to come up with their own conclusions. It probably > won't happen, but that's the reason feedback is so important. They will > take a look at information and decide if it's worthy of their time. > > For now, the above is what I've observed. > > But before I end, there is one final comparison I can do that should help > support my above conclusions. > I can compare my own test to Antonio's father and this is what I get: > > Wow, this was not expected. We have a total of 22.2 cM in common. That's > more than both my cousin and father. I don't show as a match to Antonio's > father, either. None of my siblings do. And the reason is that I have mine > broken down into more segments with the largest being 7.2cM on Chr 6. For > my father, he had just 1.4cM on that chromosome. I had a second largest > segment of 5.6 on chr 17 (notice that it's the same chromosome where my > father's 9.5 segment lies). But I don't think it has much significance > where the matching DNA shows on the various 22 chromosomes. That's just the > random nature of it. > > Apparently my larger total cM than my father is coming from segments that > are less than 1cM. No doubt my father had the same DNA which I got from > him, but in my case, there was enough of it to randomly combine into more > 1cM and larger segments that gave me a higher total. > > Another of my cousins has the following compared to Antonio's father: > 15.3 cM total with the largest being 9.5cM which is exactly the size as my > father's largest matching segment > > A sibling of this person above has a total of 17.1cM in common with > another 9.5cM largest segment. > > Neither of these two above siblings show as matching Antonio's father. > > Yet another of my cousins, a sibling of the first who shows as matching > Antonio's father at FTDNA has the largest matching segment of 11.8cM, just > a little larger then the above sibling. The total amount of matching DNA is > 20.2, but that's only the DNA that is 1cM or greater. I noticed that FTDNA > says they have a total of 25.73 cM and that means about 5cM is smaller than > 1cM in size. This helps explain why I had a larger total than my father. In > fact, he must have at least the same amount of cM as I have, but some > happened to recombine into smaller segments below the 1cM level. > > Finally, a third sibling of the two who show as matching Antonio's father > does not show as a match through FTDNA. But when run through Gedmatch, > there is a total shared cM of just 2.5cM on just one chromosome. This is > clearly a case where they simply didn't get much of that DNA and instead > got DNA matching other people. Yet another example why testing all siblings > is important. > > I have been compiling this information for many hours, between other > tasks. I think I have made my point. > > Hopefully I have convinced people of two things, at least: > 1. get all oldest family members tested and also every possible cousin > 2. study the results as I have done above if you already have tested many > relatives and if you come to other conclusions than have I, please share > them. This is how we all learn. > > Doug da Rocha Holmes > Sacramento, California > Pico & Terceira Genealogist > 916-550-1618 > www.dholmes.com > > > -- For options, such as changing to List, Digest, Abridged, or No Mail (vacation) mode, log into your Google account and visit this group at http://groups.google.com/group/Azores. Click in the blue area on the right that says "Join this group" and it will take you to "Edit my membership." --- You received this message because you are subscribed to the Google Groups "Azores Genealogy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. Visit this group at http://groups.google.com/group/azores.
