 That looks okay. And it's still recording, which is always a good sign. I will unmute you so you are now live. Great. Okay, ladies and gentlemen. We'll move on to our second presentation from today. And it gives me great pleasure to introduce to Robert Casey. Now, Robert Casey comes to us from America. He's been an active genealogist for 40 years, publishing nine 600-page family histories on his website. Ten years ago, he started concentrating on genetic genealogy and created the L21 YSNIP predictor tool in 2011. He has extensive websites on a variety of different surname projects, including Casey, Brooks, and Kiersey. He also runs the L226, or has an L226 private haplogroup project, and an L21 private haplogroup project. Now, due to the explosion of YSNIP testing, he continues to concentrate on determining the relationship between STOR signatures and SNIP testing results, and is now developing new methodologies to produce genetic descendant charts based on both why STOR and why SNIP testing. And the impact of the explosion of YSNIP STOR discovered is covered in depth in Robert's presentation. So, ladies and gentlemen, it gives me great pleasure to welcome Robert Casey. Well, this is a distant cousin of mine. We found a confederacy who's a lieutenant. His name is Joseph Henry Brooks. He survived the war. This is a kind of overview. The first section I'm going to talk about were the nice things for project administrators, certain project administrators. You've got to sort out all these groups of STRs. It's really difficult to do without SNPs. And so, with SNPs, you can quickly separate all the groups of people who are not related closely. The second one is, well, STRs are very noisy. They're like water flowing down the river. If you go out to the edge of the river and you're unique, that's great, but you can flow back and forth and then end up in the middle river again where you were 5,000 years ago. That doesn't happen too often, but it's five to 10% of the people have a problem in that area, and that's why you get a false hit rate with just normal genetic distance matching. SNPs will actually help you filter out those false hits. Also, you need to understand what signatures are. Signatures are better than genetic distance. Genetic distance is how much different you are. Signatures are what do you share with other people. What you share is really a much better methodology to determine relatedness. And also, once you develop a pattern between STR marker values and particular SNPs, you can actually sort your spreadsheets and predict very accurately. I mean, we're talking at over 99% on the vast majority of SNPs that are at a certain time range. So you don't have to do the R1B or the L21. You can just go directly, look at your signature, and test down to the L26 or M22 directly and save a couple hundred dollars. Also, if you look at YSNPs, YSNPs, YSDRs, surnames, always combined together make a really good combination. You really need to have as much information as you can. If you just do SDRs, that's like depending on census records only. You know, SNPs can be probate records or tax records or whatever. So the more information you have, the more ways you look at it, the more likely you'll be able to figure out what's going on. And something that's very exciting right now is there's now getting to be a critical mass of actual genetic information that you can create pretty good charts now of how everybody in your group are actually related. It's not quite there yet, but it's getting very close. And also it's very tedious to create these charts, but there are people that can do it, and I'm trying to figure out ways to automate it. Now I discuss a little bit about, it's a really bright future. The full genomes, which is the biggest test, is about 35% more coverage. They're getting one step per three generations when they do tests. So that means if everybody were to test, that means every third generation, just based on SNPs, you could put a marker on your great-grandfather. Oops, did we lose the signal there? We're also getting with all these NGS tests in big Y. We're getting a lot of SDRs, and we're not really looking at them yet. We've got hundreds of submissions that are that way. And so eventually, if you use 500 SDRs versus 67, I'll say that's a lot more content. Now we need to sort out which ones are unreliable and which ones are mutating way too fast. But between these two combinations, there's lots of big Y tests in the future and lots of SDRs. We'll be able to assign actual certain mutations for everybody on our pedigree chart. And there may be a few unlucky ones that don't have one or skip a generation, but they're going to average two or three mutations per ancestor eventually. And that's 10 years out, but that's going to happen someday. Most projects do organize their groups. I'm talking about certaining projects. And what they do, it's really hard to sort them out if you're a certaining project administrator and you're looking at all the SDR patterns and saying which ones are in common. It's a very tedious type of job. You'll have a lot of smaller groups like E, I, and J, which are pretty common in Europe, but Europe is dominating our own B. You probably heard that 80 times the last few days. And also, many projects have very big clusters, along with a lot of really small clusters. And then they have all people who haven't tested enough R in 269 is pretty close to R1B. And that's like 90% or 85% of all Irish people are following that category. And this is like four or 5,000 years old, so having a common ancestor of four or 5,000 years is not recent enough. There's no need to test it down to more than 1,000 in a 1,500-year range. Also, a lot of people get isolated, and so you'll know once you get people separating groups, you know you don't have to look at the surname for all those other groupings, because your small group in the 34, the only ones that are really related, all the other 150 are not going to be related to you. So that's a very good thing, a process of elimination of people who are not related to you. Here is my KC project. E thereby says Africa, but they've tested some ancient remains in Europe recently and found E in Europe after the Ice Age. And so they're really part of the original Europeans. All these R1Bs, including all those Irish people, were just invaders from the Bronze Age who took over Europe. E is actual, well, a few surviving original Europeans after the Ice Age. I, of course, I, means Viking. And J, there's a lot of controversy about J got here. J is, if you look at all the academics that said that, you know, in Syria and that kind of stuff in the Middle East, but they may have come with the R1Bs as a minority. They could have come in with the farming inflection of that technology. And so they're probably, you know, were a tag along with one of the other incursions. DS-27 is really kind of, is an anomaly here because that one kind of has sources in Spain and France. So, you know, we've got an Irish person who, you know, has worsens actually in Europe pretty recently. Here's the first big cluster under KC. And this is probably most of the cases. I'd say from the clan basis, KC had two large clans, according to clan legend. And they're both from, you know, the Minster area. And so, you know, we have two big clusters. But this one is a lot more genetically diverse and a lot older. So this is probably the largest clan cluster. So it matches up with the clan history books. Here's the second one. This is a much larger cluster. But I really think this is a much smaller cluster. This one, actually, there were Protestants in 1750 in Western South Carolina. And Western South Carolina was very sparse back then. Now they may have converted to Protestants because there may have been just one or two men there. And they were in Catholic Church. There probably weren't too many churches at all. So they probably, you know, one or two churches to pick from. And they were probably Protestant. This is mine, by the way. And also, I've been researching this for 40 years. And this is the reason I got ended to Delegate Geology because we're trying to break up this cluster. We have around 50 to 75 known men that we cannot connect in the 1700s. And, you know, and just all the documentation, the probate records were really poor in this county. So we could make any progress. Then you have the other, the biggest Irish cluster up to two, which is the O'Neill's. Well, there was a couple of cases in there. And if you look at the clan lore, they says that there was a Casey group in that area. So this, and it wasn't a very large one either. So this could be them. And then you get out to, this group is actually, all these people actually share a several markers together. So I had them grouped together just based on why is the air markers at first. And then people started testing. Now these people have tested very much. But so the people they matched has tested L21. You know, very close matches. So I kind of grouped them into the L21 group. And L21 is a big chunk of R1B. So, you know, three quarters of this room. Well, actually, L21 is probably only 50% of Ireland, some of their ranks. And then this one was genetically different from this one. So I separated it. And you got Dn21. That's another major branch around 2,500 years ago. Z251 is another major branch. Then you have to do something with the new cases. This guy, he knows he's a Murphy. That was his grandfather. And he has proved that, you know, he was adopted. Or his dad was adopted. But he's a Casey. So you get a clue. He's actually a Murphy. He should go in the Murphy database more. But he's a Casey now. So you have to respect that when people fairly recently adoptions come in, you want to have all the genetic new lines of Casey's. Because, you know, 500 years now, that's going to be a major Casey line. Right now, it's limited to like seven people. And then this is a grouping. This is a grouping that really bothers me the most. They haven't tested. They're just all over the map. I don't know where to put them. I mean, this is just a grouping. These are not related because it's way over 4,000 years ago. So having to where else to put them. The second topic is when are they false met matches? And so convergence is a pretty common scenario. It only happens about five or 10% of the time. So you say, well, who cares? You know, that's a very little small amount. But if it happens to you, you can have false hits 20 to 90%. So you have to be aware that this can happen. And if somebody says you had convergence, you need to be really careful about why it's the R matches. Because they may be a lot more questionable in nature. What causes this is you can look at R1B. There's a bunch of charts out there that says, here's all the common values of R1B. Okay. And, you know, 5,000 years later, what are yours? You have a genetic distance of two from something 5,000 years old. You're in trouble. Because that means you probably mutated back and forth all the place, but you ended up about where you started 5,000 years ago. And that's what causes convergence. So you can have very common SDR marker values, but they can be false hits. So as somebody said, it's not a match. Trust them. If they have enough skills and you trust them, it's not a perfect system. So you have to depend on a lot of leadership. There's a lot of really good volunteers out there that have been doing this for quite a few years. Then here's a new one that I just discovered. Since we have all these nice new information today and we're getting them to be able to look at charts, I started to look at sorting along just the signature within L226, and I'm finding I had actually a 90% error rate on the first person who was on the left. These were matches of 1, 2, and 3 genetic distance. Now, this is even a smaller portion. It's only for about 50 people out of 500, which is still a big, major thing. I'll go to that in quite a bit more detail. And so the only thing that you're going to be able to do in that timeframe or one of that scenario is you're going to have to test naps to break them all apart because your wise diaries are just not going to be dependable for that particular grouping of people. Now, this is a big, busy chart, and I'm going to blow it up, but we're going to pay attention to this column, the genetic distance column, and the next to last column, which is the most recent SNP or the one found in Treaty A, called Termistip, which people don't like that term anymore for some reason. And I blown this up. Now, this particular guy pulled up is one person and this is all his matches, and this is what he sees, but this is very condensed down. So he tested SRY-2627, and his first match is this, all right. He said, well, that's a match, right? That's good, but that SNP is 2,000 years old. So they are a match at 2,000. That's just a little too far back for surnames and good genetic genealogical connections. For Irish people and Scottish people, that's around 1,000 years for when surnames actually started being used. And then down here at 6, which Famicrine calls a match, and this is just not case a group. It's probably much worse than other groups. You have BY-2628. He's part of the L21, but that's 4,500 years ago. So this guy can't be related more than 4,500 years ago because he has a genetic distance of 6. And not to go more detailed, there's more examples, but these are all false hits, and this is really good at it. So he has an 80% false hit rate. But if you look at this little common sense, it's like, well, 4, it's still good, but at 6, they're all bad. So if you do a little filtering of some of the higher ones, then you can get back to pretty good matches. Here's the new thing. I just took the fingerprint of L226, which is 9 wide. So this column here means all these people match all 9 signatures, so they're a very clean match with L226. And then this is the genetic distance from the actual 226, actual haplotrite. And so I picked the very first one, ran and asked how I was going to take the first one, and everything that's in light green, they're actual matches. So we have 1, 2, 3 matches. Everything in the light burgundy, they're non-matches. That's a 90% error rate of SDR matches. Now there again, this only happens 4, 5% of the time, but it happens with your guy. And it happens to all these people here. And they're saying, I've got matches of 2 and 3. What's going on here? This is a different branch of 1,000 years. This is kind of a new phenomena, and it's something to contend with, because this is not a perfect system. So genetic distance is a good way to determine, I mean 90% of the time it's very accurate, so it's a good way to determine your lateness. But also you should look at the diversity of the surnames for your matches. If you have a whole bunch of matches, and you have 25 matches, and there's 25 different surnames, that looks like you're not really having very good matches, because, you know, FPEs are not going to happen at that high of a ratio. But if you only have 80% one surname on all your matches, then you've got a pretty good thing going there, because now you have a genetics confirming that everybody with this genetic pattern are actually coming back with the same surname. And not all SDR mutations are equal. There are some little ways that you can kind of tease that out relatedness. For instance, like some are very slow mutating markers. I mean some mutate every 50 generations, and some mutate every 500 generations. So if a SDR marker mutates only once every 500, that would be much better, because it doesn't happen as often. And so there's going to be fewer redundant mutations of that. Also if you go to a very rare marker rate, I'm real lucky in my group. I went from 460, 11 to 12, and 12 to 13. I think under L226 that's unique. There's nobody else in the group that has that value. And under L21 there's only like 16 out of 12,000. So that's a pretty unique marker all by itself. And also multi-step mutations, those are the ones who, the value has changed more than once. Sometimes they can do it all at once. But those are another really good powerful combination because they don't happen that often. Multi-step mutations happen maybe 5% of the time at most. So if you have one of those, they're a lot more unique in nature. And so the most powerful way to figure out if you relate is sharing common mutations of a signature. And so you say, what is a signature? Now this is a busy spreadsheet here. This is a big old Excel spreadsheet. Ignore everything to the left here. And can you see any pattern of color? Well, yeah. Doesn't the yellow just jump out and scream at you? Well, that's what a signature is. At the very top there are, I put the L21 modal and the L743 modal. This is for L743 inches. It's a branch that, I'm not sure if it's in, no, it's an English branch. But it just jumps out. And once you find out what the signature print and then sort the spreadsheet, you can just visually see it jumping out at you. Now there's one really interesting one here. I'm going to zoom in again because this is obviously way too much information. If I zoom in, I've now deleted about three-quarters of the columns, about 90% of the rows. And you see there's one that kind of just jumps out at you. There's one that's green and red and white and what's going on there. Well, that one, this step turns out is pretty young. And what is actually happening is the value of 16 at this one particular marker guy, everybody who has 16 is testing positive for this particular branch. And, you know, even when I found the guy down here, he's in row 107. And he tested positive and he had all these guys that are a lot closer matches, but they don't have 16 for a value. So it looks like the value of 16 of this marker and this branch, this net branch are almost one to one. But it could, one could be a little earlier than the other. They don't have to be exact. They just are tracking that way. So you might find a couple of 16 that are negative or a couple of 17s that are positive. But today it's, you know, all tracking. But, you know, signatures is really good, but signatures need to be filtered by what? Genetic distance. So you just, I just said, well, genetic use is no good, but it is a good filter. Because here under, this is for all of our, the whole entire R1, not R1, all of our half a grid, which is probably, well, it's 50,000 people in this database and 67 markers. And he matches at, he's got an eight out of nine matches, a real strong, but he's got a genetic distance of 30. So obviously he is not related. So you can't just depend on signatures. You also have to look at genetic distance to filter out really bizarre matches. Anyway, and R1A is really old. It's probably 6,000 years old or something like that. Yet they just, they have an eight out of nine match on the signature. So that can't happen. So how do you do signatures? What it is, a signature, you take the modal, which is a mathematical term for the most common value. And then you take some huge SNP database, in my case, for L225, that's L21. So you look at the most common values in L21, you say, well, that's where I started, you know, four and a half thousand years ago. And you say, who cares about four and a half thousand? You care, because it helps develop a signature. So now we're down to the 1,500 year range. It's really over 1,500 years, but so few people survived that 1,000 year bottleneck. L226, we have nine very unique markers or mutations. And so I find the modal of L226, and there's nine markers that stand out as being different. And so those nine markers that are different between this average most common value, that's what a signature is. And then you can take the signature and you can go find good matches and build hapo trees and descendant charts like all the genealogists want to do. Then you can do it next. Once you get an older haplot group like L226, you say, well, 1,500 years, you know, that's way too old. Well, then you can take that signature and compare it to your surname cluster. For instance, my South Carolina cluster. And I'm really lucky because my guys wrote that Dyson got 12s of mutations when they mutated all. My wife says that she knew that, but they really had a lot of mutations. The next biggest cluster had like three. Mine has seven. And so I can real easily predict anybody in my South Carolina Casey group by looking at that signature on L226 real quick. So finding close genetic matches, such as the family tree DNA matching is great, but really shared SDR mutations are better matches. And if you find both, how close you are, how much you share, what kind of mutations they are, they really rare marker days. There's a lot of hidden genetic information that you can use to analyze. And then also the more information you have is just like, you know, you can't stop a census record and family histories. You know, as you dig into the tax record, you get more information. And that's what this kind of stuff is talking about. There's a lot of tax records in that area. Here's another signature, but this is the one I was talking about for my Casey now. I'll zoom in later on this one, but the blue, you know, it stops pretty quick. It's like it falls off the edge. So that's because a certain cluster just isn't going to be as big and pervasive as a big branch with a very large group of people and ministers. So I'll zoom in here. And so these are all Casey's. And this handy person, he wrote a 900-page book on the Handleys, and he came and started doing DNA testing. And guess what, you know, and in the 1700s they all matched, and all of a sudden in the 1830s, he imagined all these different cases. What's going on? Well, he called me up and it turns out, I was always wondering why I received probate pre-seeds from some Casey's. Well, now I think we know that this Casey died and this handy was a nice person and adopted some young Casey males. He is really petrified. This is the one you pull up a can of worms. He printed out this beautiful book for like $20,000. I mean, it was a beautiful book. And now, you know, that half book isn't really part of the handy line. They're really Casey's, so he's keeping a low profile. Now, that only jumped out. It says we had one kind of outlier. He doesn't match everything. And he's cursey. And guess where he lives? He's a good Irish person that we're expecting to know. He lives in Oxford, England in 1600. Now, what's going on? Well, he lives in Oxford in 1926. He's definitely Irish. So we suspected that he may have moved to England because English people don't like Irish people that much back in the 1600s. He probably changed his name to cursey to avoid the prejudice at the time. L21 is about 40% of the United Kingdom and Ireland. I don't like to use the term British Isles, but it's been a long, for a long time. Around 50% are scientists, newer SNPs. And what you can do with SDRs is you can actually develop these patterns and I have a tool called the L21 Predictor. And it will take 50% of the 15,000 people and I'm going to tell you L226 or M222 or L193. You'll just go directly to it and it says you got 90% off. And this is all based on statistics. It's called binary logistic regression. It gets perfectly for genealogy and genetics. It's a perfect match. And in the timeframe of a SNP that's a thousand to 2,000 years old, it works beautifully. Now when you get down the recent times, there's a lot more parameters involved. Here's how it works. There's an 11 marker signature for M222. Everybody who has tested a 6 in a 1, that's about 500 people have tested positive. Everybody who tested a 5 in a 1, all 15,000 has tested negative. It's called a perfect curve. Now one time we had a guy up here who tested negative. And this guy was upset because he thought for sure he belonged to the needles. And so he actually went and tested the competitor and he came back in to 2 and they readdressed it and boom, they corrected it and we come back to a perfect curve with you. But actually statistics likes that better, that error because from a math, it makes a nice beautiful S curve with one error. And math doesn't like things that are so simple because it's just that simple, why are you using math? And so all the algorithms don't really work as well. And so thanks to Dennis O'Brien, I mean, and Dennis Wright, we got L226 on a tree but we had for a long time, is this it? We're going to have a SNP that has 300 surnames in it. And then in the last year or so, we've done 50 ingest tests, 50 SNP test and probably about 100 individual SNP tests. And now we were just really happy as it's happening. We're able to actually sign people, like there's several O'Brien lines. I mean, we thought they're all one and sometimes they have genetic distance of two and three but they belong to a different group over a thousand years ago. And the case is now, I described the third branch of cases under this guy and that SNP branch didn't exist five months ago. So the nice thing is there's the senate charts that are going to be generated now. They can't generate for everybody because it's still a signature-based type thing to make these charts and not everybody is tested for every signature yet. And so it's only like 50 to 75% complete but now we're in a regular chart about 50 to 75% of L226. When we use the earlier SDR only type things, we didn't know if that's accurate or not. It looked like good math and it put people together. Well, then when we started doing SNP testing in the early stage we found out man, SDRs by themselves do not work. It was like over a 90% error rate. And so everybody gave up on building trees based on mathematics and stuff. But now that we have a whole bunch of SNP branches and a whole bunch of testing of SNP branches, both, we have a lot more information. There was a real good promising tool called SAPP and it still does a really good job of collecting certain enclosures. You just run it and it'll collect everybody as a common pattern. And it'll collect about 10 or 20 of them and it beats the tedious method of going through a spreadsheet. So it still has a lot of value, but it's based on the wrong kind of math and the guy kind of gave up and he was having performance problems and he couldn't get above 150 and you really need, you know, 1,000, 2,000. We have 222 and I'll do it in 6% as well. So with the flurry of recent step packs we have gone up to 23% of our group that was tested out of 500. That's a pretty good penetration model. And it used to be, before the step packs I set that 50%. In the last two weeks I've been busy charting away and now it's up to almost 70% of the entire group is charted. So, and I'm sure since we're getting so many good results in the step packs for L226 I'm sure that more people are going to jump on the bag when I'm in that number that go up more. But then you're going to have those unlucky people who don't have any genetic distance. They're going to be the last ones that are going to be hard to figure out the last 10 or 15%. So we have 50 ADS tests. We're now at the 41 branches and I'm sure between the two dentists we're probably missing three or four because we've been working here for a vacation. And so we have 75 individual tests. They have revealed seven new branches and as well as six step packs tests. I mean step packs that have revealed six more branches. So the step pack test, this is a really unique and this right did an excellent job. We had 50 private steps inserted into our step pack and so those are just random private to one individual. And now we have six new ones that now we have three people that match. So it's pretty exciting. And 68% is like, you know, I didn't think we're going to get there that quick to be honest. Now how does charting work? As with this prediction, it's based on signatures. And it has a combination of signatures but it's got several more factors and this is fairly complex and if we can get this in a tool then you can just dial in what reliability do you accept in your charting. If you want to dial down to 20% or 10% I'll chart 95% of it or if you want to leave it at 40% you can get 60% of the chart. It's going to be up to then user. Of course there will be a lot of people giving the whole chart, you know. And it won't be very accurate. But I don't include ones that 40%, so I thought 40% probability would be worthy enough. That's a 50-50 chance but rather than just randomly test you got at least a 50-50 chance from this one. But something that has jumped out on me is that if you have three or four markers that match that's a really strong signature in this time frame and there's very few three signature matches that are overlapping which is a real surprise to me. And so those are intending to be pretty reliable because probably the 78% range. Once you go down to two signature boy it gets ugly quick because I found some that there are like six different copies of two signature matches and these are all the faster markers. So I'm just having to filter them all out. And genetic distance is another one because sometimes you get a pretty good match but everybody else has a genetic distance of two or three and also there's another guy eight or nine. That's just an anomaly. He doesn't really belong there. That's probably a small duplicate signature. Analyzing tract and 21 new branches it takes a lot of manual time because every time you get a new test result you have to go look at all of it. Its own signature for just that one test and then develop a branch based on that one that test results and so you're adding little chunks of the tree to the time for each person who has tested for SNFs. The price following this is I think this could be automated pretty easily and I'm trying to write a functional spec so if there's any programmers out there who really want to make an impact this tool wouldn't be that hard to get the 90% it wouldn't be that hard to get that last 10% where you look at unusual markers and values that would double the length of it to get your last 5%. So benefits of seed charting so you can graphically see your matches there's nothing better than having a little chart that shows your matches. It allows people to make better as an admin you can now say it looks like you belong to this branch and you have a very unique signature maybe you should do a big Y test or you're really close to this and just as you maybe should test some private steps of this I mean it's really going to really help us in giving advice to people with testing advice and it subrises all the connections in a pretty easy way it understands very graphic and it's automated a lot of administrators could use it and it's in a format it's a boxed to seed chart and here's the boxed to seed chart if you... these are all the branches now this is not a real good box this is just a portal into all the branches so if you branch down to this this is hypertext enable it it should be it's not done on my computer so you want to do this one? no this one here okay okay so this is my south carat final branch and it has a pretty good progression of YCRs and I actually put in the oldest known ancestor because I've been working on this line for 40 years I published a book in 1980 on this line so I'm a little familiar with this one so you know it's very easy to understand and we have a second ingest test pending on this guy but he's kind of an outlier he actually is interesting because you know we have this major branch which you probably cannot see I'm sure it's 460, 11 and 12 and this guy is 460, 11 and 12 so he belonged over here right but he had two people test so he's not 460 so he could belong here or he could belong here we don't really know it's either a parallel independent parallel mutation of these same marker values or this is a backwards mutation we don't really know and because of that he got really interested in the he ordered a Wiley 2.1 and I give him the opportunity so and as I showed earlier this group is really lucky because this is an SDR bait this is going to 12 to 13 is a very nice branch that splits up we know all these cases are closer related and we know all these cases are not as close to related and I knew my gut feeling I knew that this guy was related to me before he lived I knew this guy was really close to related to me but I thought this guy was close related because he lived in the same area but and he's not and I have this new guy who just tested recently and I gotta go see where he's from and research and I'm now actually interested in going looking at him and seeing where he fits in and then we have our Hanby and the really interesting thing is all these people in South Carolina except for this guy he's American he matches us he's from Virginia and all these cases are supposed to be from Virginia but all the people are supposed to match from Virginia guess what 7 tested and 7 are not matches so I guess the old family histories were not correct on that what are the colors here the colors the color coding if they're blue that's it and then if they're purple that's a SNP based branch which can have some SDRs if they're white box that's an SDR branch and then the other colors anywhere from being very green to more red the more green you are the higher probability this is a 70% and this is a 40% I'll do one more and my other favorite one which is another casey you've seen no brands and the rights a lot so and so here's a this one is very different this one is much more first of all it's a lot more reddish except for this one green so these people need to test to kind of verify this is probably the 50 to 60% right 50-50 type 60% of these people are cases which is a pretty good marker and this is DC69 is very unique because it's the only branch off of our oldest branch before we go down to everybody else so it's a and this tracks the clan lore that there are very early branch off the O'Brien's I'm not sure if they're kind of the O'Brien's just confederate with O'Brien's and there was there were an ally of that group I'll do one more okay and this one is I like this one because it's just it shows you the power of ingest testing before we had this a 606 97 and it was a proven mishmash but there were some canons over here and some canons over here and then they did another ingest test got DC19 now there's two cannons, three cannons a buck cannon and even over here you got a cannon and a cannon so we're now beginning to realize that the cannon surname probably happened here or in between here and the McCann is probably a sort of a variation of cannon or vice versa but that's a good wrong signal DC19 means you're this cannon group and over here you have you have Kennedy's videos which are quite different and 2e which is quite different so this is a really nice geological this that means you are that cannon and a particular cannon line great thank you very much that was really really interesting Robert and cutting edge this is very very interesting to see because I think a lot of us in the audience who are surname project administrators were going down the same path I think some of us have discovered it independently so we're kind of all converging oh yeah, not independently you've been doing it probably for a lot longer than most people yeah so this is very interesting because I've done this with my Gleason project as well I know James Irvine is doing it with his Irvine project and few other people in the audience are doing a similar type of exercises I think it's interesting the way that you're colour coding the probability of your predictions based on the SNPs and the STRs which is not something that I've done before one of the things that I think will be very interesting in the future is when we do get it automated will be to have also the dates for the branching points and that could be very difficult to estimate because it's usually a very large range it could only be anywhere from 500 AD to 1500 AD but then also putting in as well as the surnames the locations of the actual most distant common ancestor, most distant known ancestor for each of the participants in the project yeah and I just didn't have time to actually add the countdown and they're all in Munster okay so they're all in three or four counties except some are further up but that'd be nice to see that exactly if there's some kind of geographic clustering as well as serving clustering have you been able to get down to the townlands level no we're the bulk of our L226 and Dennis Cretme if I'm wrong is we have about 70% in the four counties and then the rest of them yeah the Seren memory Tipperary and the Cary Cork or Cork and Cary those are probably three quarters of everybody and the rest of them are just spread out over our and we have a few outliers in Germany and that kind of stuff I found matches in L226 in Lutuania so from Munster to Munster so we have time for a few questions I'm going to take one from Mike O'Connell over here Mike I'm going to ask you to move back to this chair because you're very close to that speaker and I'm just scared that you might do a lot of problems just a question you mentioned the bulk of the and so on you know and Marsman from townlands because townlands are retail you know in Tair between Innocent and Innocent in Roche there's a townland this is Casey which means the fort of the Casey's this is Fort and also for the rest near the town of Shannon there's a you pass through it now the motorway it's a place called Halley Casey which the town of the townland of Casey's Casey comes up in our blood in the end it's something to look at because whether close to a second they eventually became part of the name of the audience well unfortunately for my particular Casey Brats we have no connections even though we know we're arch you're just nobody's tested from there and we have they're all South Carolina or one Virginia wow we have a Casey we have a Casey okay we have to get some mail relatives if you give me one of those free DNA tests then you can prove your ancestor I'll give it to you oh well you are getting free DNA tests so do you want to give your free DNA tests away to a male Casey? well it doesn't have to be a male it has to be somebody who has a good pedigree chart of their Casey's good pedigree chart of the Casey's at least 3-4 generations back it's a great grandmother did you have the father of them do you want to have them memorized? it's a start and who is the other Casey? there are two Casey's over here okay I need to have a father because I can't tie it in the wide DNA without a father he thinks we may be related and I think the father was the major plan in Longford I'm a father but I've heard that there was kind of I don't really know the fards came from 10-14 from the bastard from TARF but the Casey's I think were in Longford before that time and there are a lot of Casey's in particular areas of Longford and I saw a scum on your chart but there was no sign of Longford I imagine nobody would be tested because I'd say down there we don't have a single person from Ireland of a Casey that's been tested they're all Americans who have tied back to Ireland no single person from Ireland you spoke about the modal STRs for L21 right is that something that sort of agreed across the community and are there similar modals for other happenings? yes all the big ones like you just go contact the administrator for U106 L21 go to see Mike Walsh's spreadsheet there's one controversial one on 449 is like 29 or 30 but it's so fast it doesn't make much difference but it's he and I have about 20,000 people at 67 markers it's pretty easy to figure out the most common value would there be one for L2? the L2s are pretty small branched you probably wouldn't have enough sample size to get a good marker but U106 the N27 all those big ones have very well established modals so Ireland Gail Matheson has a webpage which keeps track of all the modal happenings for R1B we could do is actually putting those in the wiki if someone can come up with a list which is in the wiki that would be really, really helpful because I don't think Diana maintains her page anymore they're all if you go to the half a group projects they're all advertised there as well but not consolidated, that's a problem quick question have you really thought about how 400 STRs if that's going to make any difference to this problem how do you actually study STRs how do you get rid of this problem I can't remember if it was Ian McDonnell who's collecting them what? U106 I think he's collecting all the STRs and trying to analyze which ones we need to filter out like the CDIs and also he's trying to figure out how do we put them on database which volunteer for that I wouldn't run a database another thing would be really nice would be an update of Leo Little's tool that looks at the frequency distribution of STR marker values half a group by half a group he's only done six of the major half a group to all in B or when A I think he's got I1, I in there as well but it would be nice to because I use that a lot for identifying the rare marker values in my R1B to members but we need something a little bit more extensive than that I think for a family tree DNA has all the data so it's just a question of somebody analyzing it I do know that there's at least two people that are thinking of animating and automating I should say everything that we've seen here today and I think that in the next 12 months or so we might get the first tool that actually automates the combination of SNPs, STRs and surnames and I think the take home message of the audience today is that come back in about a year's time and you'll actually be able to see these automatically generated charts for your particular branch of the human evolutionary tree so there are exciting times ahead and Robert thank you so much for waiting our appetite we hope you come back and tell us more in your course ladies and gentlemen Thanks Thanks Thanks so much