 So you've got this here, and you just go like that and then back as James has already introduced himself I won't bother introducing him anymore. I'll just say ladies and gentlemen James Irvine. Thank you So as I said a bit of background on some main projects why DNA a little bit about my project is background I'm not going to go into detail and then the three main points I want to take up going right back to the beginning SDR results tables. I think that's some grind that needs going over again matching and grouping very basic But I've got some very fresh ideas on this perhaps controversial, and I want to go into TM RCA is the dating of DNA data, which has always been something nice to have But I've done some more work on it, and I hope we will find interesting as well So let's crack on basically why DNA is about the male inheritance like the surnames This is very basic, but the males only get it females can get a father son or uncle to chip in We get mutations that happen slowly and surnames change as well a lot of parallels the spelling changes But occasionally you can get a radical change and NPE which has been introduced already As genealogists of most of us are were originally and still are we don't have the Biologists here at this sort of a meeting we want to get the DNA into a family tree And this is the sort of family tree we're looking at Mike Walsh develop this right the way back to Adam And it comes down very higgledy-piggledy and the snips and the SDRs on the left Help us do it SDRs are more useful fairly recently snips a bit older But in fact they completely overlap and you can put dates to them at the end. I'll come back to the stating business There's all sorts of things you can use why DNA for including deep ancestry, which I discuss But the three points that I want to pick out today are this business of matching How they identify genetic branches and how we date these branches. So the other aspects. I'm not going to be addressing a Little bit about my project. It's it's a Scottish name, but it appears in England Ireland and in America As I said, we go back to 2005. We've got nearly 500 members now Which is medium-sized? It's got a strong US bias because we have a clan association there So 80 odd percent of the data is American, but it comes back to Ireland and Scotland as I'll show 92% in fact of the of the ancestry of the of the study come back to British Isles We were brought up in the family of the surname if you meet any urban He will tell you that he was brought up. We all have a common ancestor DNA has shown that to be complete rubbish and this isn't unique to us by any means We actually have about 40 branches all quite unrelated to each other within the surname Europe each has its own story to tell But we've you almost uniquely one of our branches is two-thirds of the total project. It is huge It's the biggest I think of any surname project in the world Candles are slightly bigger, but they include in that a much older grouping So it's much more amorphous that this l55 branch is specific to the surname Almost find one exception, but it's virtually specific and we're well now into big y and snip packed tests Identifying subbranches. We've got nearly a hundred Either big y or packed tests, which gives us a lot more confidence in what we're doing And I've even been able to explain this time last year Integrate the DNA into the conventional genealogy. We've got an overlap for a couple of families Thought it was going to be a big breakthrough We've only now a year later was still struggling to get a third but it can be done if you've got decent genealogies Right now the first thing I want to get rid of and this is not peculiar to us the surname the spelling is Misleading it's not completely irrelevant, but it's misleading you can see in our project on the right Centages of the different spellings and if you look at the world population data There's a very good correlation to the study is just picking up what the real world has It can be spelled all sorts of ways Geographically and historically there were some focuses on the spelling But now it's all got mixed up and indeed I found in Aberdeenshire for example as a branch is traditionally spelled with an e Some of the people from the Scottish borders where it's spelled with a g have moved up there And they change the spelling and that's in the last century people tend to spell as a native spell the name Very clear evidence on that You can play this game. I picked out one here. This one shows very clearly where most of the others came from It's very clearly a border Scottish border family often the opening in Shetland also should register quite strongly I happen to be of the Auckland branch, but I'll be speaking today about the main group And in Ireland a different kind of map more up-to-date You can see the scattering very much a Scottish name, but a significant number in the south and I'll pick that up in a minute The Dublin grouping is is typical of Wherever there's an urban place people might tend to migrate in the night and essentially to get work So you see a bit around Dublin For Manichur's as well, but Tipper area is coming up fairly strongly. I'll come back to that Those maps I showed were the whole population for our project I've plotted where the earliest known paternal ancestors were for By the third of the project they can identify this mostly Americans and you'll see the distribution is very similar to Population map the surname possibly comes from the sign of Irving up in Ayrshire No wrong wrong way the chief of the borders branch the tire on short are and in the plantation period Relatively affluent Irving's migrated castle Irving But the vast majority of course were far less affluent and they were scattered all over the north Would expect now one of the things that's come out of the study very interesting 76% of our study nearly 500 members of the study 76% of them live in the USA 38% of them say they think they had Irish ancestry That's as far back as they can get but the study shows that 86% actually came from Scotland So quite an interesting picture there going back in time The project that's grown steadily over the years. I used to wonder why I think I know why You'll see along the bottom the green line is the number of branches we have in the surname It started off just zero because I wasn't able to split them up in branches and then as I've got more members Along here you can see the number of branches increasing fairly steeply And then it flattens out so great slightly and the number of people I can't put into branches started off at 50% You see the brown line is half the blue line and now it's less than 10% And most of the branches I can identify where they came from geographically so this is the statistics of the SDR work 40 odd branches one of them is two-thirds and we are very lucky to identify fairly on a very specific snip fairly young snip but the age of our surname that characterizes these these Border Owens we have a lot of NPEs 19 groups in fact you see that They're 19 different branches nearly half of branches are border branches, but there are other surnames The DNA is other neighboring surnames Bells Armstrong's Graham's Obviously border surnames obviously something was happening in the 1415 centuries and there's a large number of groups, but but not of people and Then Abedin should I mention for show or me Shetland Island? This is very interesting two branches of these ones are even though they've got the Scottish surname But they're very clearly nothing to do with Scott's Irish the Catholic In living memory they spoke gaelic. They know the old gaelic surname and they all came from a particular bit of Tipperary So what happened was the gaelic name had become anglicized maybe as long as the 16th century So they fit into our project, but they're not Scott's Irish at all And of course as you can imagine get quite upset when I tend to label them as such the general blame The African one is very interesting. This is only a 12 marker one single test came in And he was this e group e Here we are e Yeah, they're e completely different everything else African We're very proud, you know, I'm African afro-american. I'm black My great-grandfather was liberated as a slave in 1865 Anticipated, but I don't know how he got his surname and I was able to say well probably the owner of the plantation was the Scott's Irish who migrated to Carolina, Florida in the 18th century and you probably got the surname from him So he was thrilled with that and I was thrilled show that our surname actually has that social spectrum Right from right wing Americans. I don't know. They're any present Lady Right through to the other end of the spectrum. So we're we're broad church Now going on a bit to the happen to the snitch and the haplo trees Alice to show you a bit of this. This is my Walsh's picture of R1B and you'll see we're in Here's us down bottom bottom right the border Owens were L5 5 Literally at the bottoms near the Royal Stewart's that was mentioned, but we're quite young sort of the bottom Why are out in the limb? I don't know Johnny come lately or something like that We're quite small nothing like these big ones, but we figure on on that map Here's Alex Williamson's big tree. There's L555 up there. You'll see it's in a huge block So that whilst the bottom of that block is about a thousand years old the top of the block is probably about 3,000 years old Eventually when enough people test that block will break up and we'll be able to split it possibly But our bit is only one in this according to Alex Williamson. This is not up to date There's only two two branches. Those are all Irvins and these are all Irvins Part from some NPAs I'll show you in a minute that I've got more detail on that So that puts us on that map This is the way FDA DNA do it now FT DNA when they first brought up big wide the haplotry was a way behind now I believe it's better than Alex Williamson Alex and Williamson's work is more of a higher quality It's got more potential but at present the best haplotry for as I'm concerned the most detailed one is the FT DNA one There's L555 and you see they split it up into an awful lot and when it says more That's all those ones that were in that big block that Alex Williamson showed and if you click on more you get all the other ones So you'll see there's a lot of detail there Now I believe in this is one of the things I want to get over today do your own haplotry because most of us are pretty Introverted egocentric we're looking at our surname. So this is the haplotry going back to Adam of our surname And you'll see there's the African one. I was talking about there's one of them at the Gaelic Irish ones These are little codes. I have these two-letter things And right down here at the border of in CL 5 5 5 the other Irish one is is done there I for I island M for master I For Island L for Lester is the way I've done a sort of coding The brown areas the bit that Alex Williamson does and the astra the hatches signs Are where you can get a pack test which is a cheap way of getting your snips you won't find any new snips But if they exist and your pack test is designed well You can get onto this tree fairly quickly, but I recommend that everybody be their project leader or just a member of The surname group make their own haplotry of their surnames so you can see how closely the branches connect Because we're all related back to Adam, but the surname era Starts away down here at the bottom Roughly, so anything above the bottom line is is before surname So the surname bit is very recent Now if we I want to go on to the next bit the next tree sort of takes the L 5 5 bit and Expands it and look what we've got There's L 5 5 up there But it's earlier ones. I just put in on the other tree and then you see we've got one two three four five Like cut six six levels and we've got 43 subgroups All the ones in boxes are big Y The ones in green down here at the kick numbers that have done pack tests And I've been able to tag on a few family finder and conventional and genealogy ones at the bottom So there's about a hundred kicks on that about a third of my project. I can get onto this one haplotry There's only one or two I can connect up with convention genealogy the Bond show one there The castle urban one I've got and there's one coming But that's a family tree, but without names, but that's all Irving's except for it's one of the extreme end There's a Wilson, so I think we now have to recognize that L 5 5 5 is not surname specific It's the next one down that is surname specific That one is just at the beginning where the Wilson's some Wilson spit off Okay, that's the background now the first point I want to make is this business about results tables and it's got it's changed a bit in the last Year in fact, we've lost the world families tables, which were very interesting very nice And we now FT DNA do four different results tables. I'm going to show you three of them It depends whether it's classic colorized and whether the opt-in Version of all the not opt-in version is is what you're looking at Here's the classic one without the colorization And you'll see that it's got two groups That's the avidine show branch and then they go into the borders branch So this is the beginning of 300 people that belong in that branch and there they are listed out. This is the way FT DNA do it Now you can get the colorized version which FT DNA do and you'll see that these colored bits here are where the the mode is shown For each group and then differences from the mode are picked out in color That's very useful, but it's very limited because they've only used two colors Blue is is less than the modal value for each marker and pink is more than the modal value So the modal value there is 29 and that in pink is 30 and you'll notice down here the three with that marker there When I go on to this one There's four Now this is the private version and the previous one was the public version You only get the private version now if you're signed in You don't know looking at the screen you're looking at a different version But if you're logged in as a your personal pages FT DNA, this is what you'll see I shouldn't show that one because it's a private one But in fact I can show it because he's since signed in so he's not public so I can show you That he's of the other one is out of date. That's why I can show you But you'll see it's still the color eyes is still only two two two colors and all you've got is your STR markers now This is my own private project website which is on the web And this is the borders group at the top And the first page top left was there's two hundred and fifty columns and this In this website on this street and it's five hundred long They've got five hundred members, so I have to show it bit by bit, but I've been able to put in the hap Put in the hapler tree. See there's the hapler tree very much as per FT DNA I've got a column here for family finder results I've got more genealogical data here which test they took how many markers they tested and all this sort of it And that's just for the first 30 and just top left top center as well as more familiar You'll see here are the Here are the one to sixty seven markers going on top right is for the hundred and eleven markers I use a more sophisticated coloring thing. So if it's one more, it's it's It's a yellowy color a brown color red if it's three more. There's no reds there and similarly if it's less And I've been able to identify blocks because these are all if you remember on the left-hand side in the hapler tree So they're all closely related And you'll see some significant SDRs coming out as a theme There is a line there showing that that one was private, so I can't show it to you Got 96% of my project and I was showing in public. There's only 18 left out of 500 that aren't showing On the public website. I haven't had a single one who said he doesn't want to it's just that some have been defaulted out And I haven't got them back in yet and then the middle of the thing is a more familiar picture, but you'll see the The different branches the smaller branches as the Abadinsha one the color here is different from the Model of the whole project in other words effectively the borders one and you'll see how they each Project has got blocks that are quite unique to that that particular branch But quite separate the others there for example is somebody that's private I've had to a raisin from the screen, so I can't show his results in public And I consider this a public forum will go on the way. That's why I have to do There's one books One null value there Some reds or quite a few days, but you can see the amount of extra detail I've got the different Atlantic west of then take a nail 21 and the rare markers up at the top and the Mutation rates you can put in all sorts of detail Now what I want to do is go over the pros and cons of doing these do-it-yourself Exhale sheets. There's some very significant different disadvantages. It's very labor-intensive. That's been built up over 15 years. It doesn't happen overnight. It's huge. It's less up-to-date obviously time somebody Gives a new test result. I'm not necessarily catching that and putting it up to date In fact, I update it in public every six months It's not fairly stable and I have to be very very careful not to have transcription errors or to breach privacy Which is not quite sensitive, but the advantages that are to me And you point I want to get over you don't have to be a project administrative of this you can do it yourself You can add in the the results of the Of the panels above 111 markers you can put in SNP big-wide data family finder Non-FT DNA results can go in I can put in the mobile genetic distances genealogical data The group the groups don't have to be an alphabetical order anymore within the groups I can put them in the order. I want rather than the all of FTA DNA gives us The flexibility and aura suit much better inside and it's a much better advertisement for the project Which I think is why we have grown the way we have I could never understand why we grew Or anything like that this is on the web And I think people like what they see and the bulk of what's good about our website is that it's got all this extra data on it on one one spreadsheet But you're doing self-table is a supplement to the FDB and a rather than Substitutes that's important. The next thing I want to go on to is matching and grouping We talked about this several of us a different different presentations over the years I want to go over it again because I Wasn't happy that I understood it myself. There's no better way of learning a subject than Preparing this lecture sort of rusted in matching is answering the question of why the two results match It's answering a question grouping is putting several together Similar processes, but actually there's significant differences And We use the tool primarily genetic distance for those of you not familiar. This is how you do it The four different kinds of genetic distance. It's pretty academic stepwise infinite levels, but it's not completely irrelevant and the special rules are some of the markers these These these ones here A and B they have to treat them differently and one and two you have to treat a bit differently And all these genetic distances don't take account of the fact that the different markers mutate it Very different rates of factor of 400 between the slurs and the fastest and yet with genetic distance They all get given the same way so you looking at a very crude measure, but still it's it's useful Now for the FT DNA matches pages There's a matches page. I'll come back to that They assume 15 generations. Well, that's that's all right There was Morris and Alistair would point out and several others quite a few of our surnames go back further than that and it's relevant They assume a hybrid definition of genetic distance. Well, that's all right. It's not a big issue But they ignore the surname evidence So anything that's not it in for 37 markers, which I'll talk about less than that's than the genetic distance of four It's in as a match now that is rubbish because some of the ones with this similar surnames or NPEs and most of the ones with the similar surnames will be false positives caused by convergence and Also, there's a rule the 10% rule this 4 by 37 is an arbitrary cut off and there are a lot of our show that Are excluded by this crude to it's a useful rough and ready start But I'll show you I'll go through each of these points in a bit more detail to show why the matches pages are a good start But they are very limited in what they try to do. First of all NPEs I was touched on this they're caused by remarriages to me. That's the most frequent cause of an NPE a Woman remarries and her son by a first marriage takes the name of his stepfather quite innocent nothing immoral about it And that's why I think a lot of the NPEs happen There can be more sinister reasons and more interesting reasons Historically NPEs are said to be one or two percent per generation But if you're looking at a surname that's 24 generations old and it was having two percent per generation That's 50% of us are NPEs. And if you think today 50% of the population is getting a surname different to their father Compared to two percent historically the NPEs are going to get vastly a number of the others But that's a problem for the future I hope and you can get two types you can have a surname surname a But the DNA of surname B or you can be surname B But have the DNA of surname a quite confusing this and you have to be careful how you handle it So that's the theory of NPEs now When you look at the matches page you can get NPEs and false positives And true matches. So this is an Elliot There's an Elliot that's Christian name for privacy reasons and you'd expect all these matches to be Elliot's when you see the ones in green four of them Or Elliot's but then we've got one two three four five Irwin's their NPEs And then you've got the blue ones with Donald Armstrong and Snowden and their false positives Now the proportion of those three groups will vary from person to person Some of them will be all green all between matches some will be only one surname Your surname and the others are all mixed up. So the mostly false positives It's very difficult to tell the only you only way you can really tell is by looking at the hapler group That will with a snip test, but you'll see even here They do list the hapler group FT DNA, but they they're fairly old one So you'll see the P312 is not helping sort out the wheat from the chaff. You've got to get a fairly young Snip to help sort out this ambiguity And then convergence is a very simple illustration of convergence I'm assuming very arbitrarily a single mutation every five generations. So after five generations, there can be three different readings no mutation plus one or minus one and then another five generations you can have five different answers but two of them so two of them the two that are Dashed that one and that one they've come back to where it started So you think the history is the straight line, but it might be that line or it might be that line And that's what convergence is and why we get mixed up. We find people the same surname Identical SDR kinds, but in fact we've got a very different history over time It's just that we can't see the history and in that particular example You've got a 22% probability of convergence happening as I say it can be even more than that very confusing very difficult to understand 10 years ago, none of us in this room knew what convergence was and I can say that safely Now we're beginning to realize that it's the it's the elephant in the room so far as STR is concerned Now this is a very interesting bit. I've taken 20 of my L5 5 Irwins and Compared them now you can compare them two ways you can compare them with the mode and so happens But the mode is this this fellow here in one two six three three seven He happens to be the mode and you'll see if you can make it out that his Gds with one exception the five are less than four so he fits nicely that one's a bit freakish But when you do the matrix you'll find that there's a lot more a lot that exceed for I put them in black There are a lot of sixes there and look at this line five five six eight nine None of them are below four. So you think he definitely isn't a match If you look at the matches page, it doesn't appear and when you do this work He's a way of a five even a way up here and yet he is deeply embedded. It's not just fringe L5 five He's right in the middle of of this. So he is a bona fide a real dinky-dye Irwin But very different from his STR because STRs occur randomly and this crude match of four Doesn't necessarily pick it pick it up So what I did that was just an example of 20 I've got 91 bona fide L5 five wins taking away the the NPEs There's no gray area and that means that if you multiply 90 by 90 you get about 8,000 So a huge matrix of 8,000 and I've done the statistics from them if you do the mode kit to modal Genetic distances it's in green and you see most of them are within four But there are a couple that's a five and a six, but if you do the kids kid ones, that's the whole matrix You'll see that a significant number are beyond the four and unless you look at the surname group and start playing with it You won't pick up those matches. They're real matches, but they're not being picked up on the matches page So when you find a matches B and B matches C, but C doesn't match a This may be the reason why because it's it's outside the outside the threshold of four So what do we get out of this? It's a bit complicated there There is the first of all kit to kit, which if you're doing matching is what you want to look at You'll see the biggest Spread I got was six by 37 of them for the Sorry modal. Yes six thirty seven nine percent were more than four by 37 Now Morris Gleason gave me his data for the his glistens Which is slightly older and he actually gets up to eight by 37 So we're not not the worst if you I think if you have a younger surname This problem is less, but it's a very old surname This problem is more acute and then if you look at the matrix Morris gets up to 11 by 37. I've actually found 13 by 37 true matches and 18% or 9% for me 27% are over the four and when we come to TMRCA, which I'll come to in a minute. This is what you want to look at The the mobile comparison for useful for grouping, but it's the kit to kit ones you need to use for your TMRCA Sorry, it's so complicated, but it'll be on the way of and you can digest it Right now another way of doing grouping is Rather than looking at your just your DD's this fellow chase actually has just come out the year or so ago with a It's surprising easy to use very very easy to use you can run through three or four of them in literally five minute Well, but what takes about a minute to do what I'm doing now once you get the hang of it very simply you just copy the public page of your surname Plug it into his app and you get you get something like this Now what he's done is the second column. I want you to look at Those are his groupings now the brown lines of what I've done myself Historically deciding there's the Irish Lensley group the Irish Munster group and the Netherlands group Going to detail on that and then there's an ungrouped one So you'll see in my ungrouped ones where I couldn't group them. He struggled as well but for my Netherlands one my Diagnosis of what belongs and what doesn't belong is exactly the same as his for my Then so one's exactly the same for the monster one's exactly the same So the way I'm grouping is the way that he groups I'm not saying that he's right, and I'm right There are other ways of doing it But the way I do it happens to be the way he does it and I think most people would roughly find that Now what I've then done is look at about 20 other projects to see how consistent that is We're there are the Irwins about Third of the way down That's the listed in sizes in size of projects. So the browns very common surname They've got 1200 right down to the outskirts. We've only got 38. This is a random selection If you're not your surname doesn't hear and don't worry. There's burn I've got some burns in the audience just above the Irving's and what I've done is there's the size of the of the project The number of groups that are I've identified you see I identified for once they're 40 I got 41 on another slide, but about that actually identified 44 So the ratio of 40 to 44 is 0.9 So if you're above one you're doing it about the same way, but the ones highlighted are obviously doing it differently Now Davis and Miller they're getting far fewer Ones themselves and actually did so that's a fairly superficial Grouping that they're doing and when it's a red one burn is doing these find 81 Whereas Ashley only find 44. So he's doing it much in much more detail There's nothing right or wrong about that. He may have some good evidence But it's just an interesting Variation and you will see some elsewhere that may raise some eyebrows, but I won't go into that The largest group you see we've got this huge one here of 286 a third of our two-thirds of my project in one group and that if you could read it would be 60% And it's easily the highest and this is why I'm worried about are we a freak? Is there something funny about this? But you'll see there's some others up over 50 so we're not way off and the sum that are very low And I've had a look to see if that's due to founder effect It wasn't there's no correlation to founder effect at all even though I expected it would be It seems to be a selfish and Irish thing But I wouldn't like to hang my hat on that and then I did another thing comparing the number of groups that Ashley found to the number of kits and there's a bit more correlation here We come out at 9% we are a plural name P for plural and the red ones which are much higher seem to be But with the multiple names and we've got one single name source single source surname down here Which is lower there's some correlation between size of group and Type of surname This is all done last week and I don't ask you to To hang a hat on it. I certainly don't but it's way with this tool now We can compare projects on a like-for-like basis. I think there's lots of potential here to to explore all sorts of possibilities So here are the different ways of different different people do matches 37 markers There's FTDA cutting off at 4 they have another one that's a bit woolly there's Ashley and here's mine and I take it up to 7 Now this is a side that I borrowed from Morris really when you're doing grouping you've got to think much wider than just genetic distance There are all sorts of other factors some are nothing to do with DNA at all surname Features and most distant ancestor some you can get from SDR data And these be people who thought the RAM authors were very significant. I Discount that I used to be keen on tip. That's got the base because they fiddle with it. There's Ashley's group Ashley's grouping app, which is easy to use, but it's not infallible. The hapler group is the one that is good But it's got to be fairly mid-level and really absolutely sure you've got to get into big y and Pack tests and or SNP tests So it's a grouping is iterative and it's subjective we can quantify it much better than we used to be able to but it still needs a bit of bit of art as opposed to science and That's one of the things I want to get a lot of a lot more work can be done on this Now the last thing I want to talk about this team RCA is time for the most recent ancestor and You can do this to two individuals which is the way most of us have done it You find two kids you find the genetic distance I'll go through it and you can find out when time back to the common ancestor Or you can do it for a branch and that's more complicated But that's what we need because the branch is going to give you a much better picture than just two individuals And the three tools we can use documentary evidence SDR estimates on the snitch and if you want a bit more background a fellow called Ian McDonald has done a very good web page It's just about intelligible for the layman Academic and he's got the old statistics up his sleeve and a lot of that But you can just about understand it which is refreshing for for some cases and I'm going to use again our project to illustrate some of the points So the documentary evidence for the border Irving's We've got one pedigree that we can back take back to 1500 quite good But other other branch the family I can take back further But certainly in 1500 there's several well established branches. So we know the common ancestor was was before that we've got odd references in the century the previous century of Irving's living there and the earliest goes back to 1376. That's the earliest record. We've got reliable Evidence we've got at 1296 Scotland have the ragman role which isn't an exclusive list of every surname in Scotland But it's not far off it and the no Irving's there So probably the hereditary surname was not being used in 1296 at least by anybody who owned land We can go back another century where there are many in Scotland, but they weren't in Delfisha so out of guess Probably the ancestor that of Fritscher Irving's was born between 1250 and 1350 There's no science in that it's sort of gone field But this is how you can use documentary evidence to get a ballpark figure of where the DNA should be pointing Now if you use strs, you've got to look at the number of kits you're comparing it can be two It can be several the number of markers and how the mismatches are counted I think that's more important than I had appreciated the number of mismatches the average mutation rate and the Average years per generation all those are unknown variables So any answer you get out is dependent on the decision you've made on my input data The average mutation rate there's three that are in the literature to use McGee was 0.024 Most of 0.033 and Doug mcdonald. I'll show you his method. He's now recommending 0.42. You see there's nearly a factor of two This isn't through that of research. It's not one's right one's wrong Just the research is as good as it might be and the number of years per generation I'll show you why we need that but that can vary between 20 and 40 Typically three to 35 in McDonald's 35 So when you plug that all together you go through the or you go through the process this is Doug McDonald's tool You plug in the number of markers you want the number of mismatching markers And he's using infinite allows which is not the same way as GD kind of the same way as FTD and a kind of difference isn't big But it could be significant and you plug in whatever mutation where you want and you go out this graph and you'll see the mode The most common is about 12 generations And if you take the 95% limits that's point 0 to 5 either side that something between 5 and 20 odd generations So if you if you're looking at a mismatch of four at 37 markers on that mutation rate That's the answer you get 12 now here are four different Approaches at the four different applications on the first one. I've assumed 37 markers point oh for two 30 years per generation four by 37 the mode the most common is 1590 And it can be between If you go up to point down to down to point three three you'll see it's moved by 180 years And if you go to 111 markers should be better You get 15 30 and 14 60 now those tables apply to anybody They're only bits in the bits in green are the one specific to me But you that data you can use any surname and you don't need to look My mass is right. That's what you want to use Please check it. I'm fine. I'm wrong. I'd be delighted to know So you don't need to go through the boring process that gives you all the answers you want But you'll see for all of them as a huge the 95% limit is a very wide spread I've haven't when it's earlier than the surname. I've just stopped at 10,000 arbitrarily No, that's one way of doing it We can also do it with snips and the way to do it snips. It's very easy You take the number of snips within a group that aren't shared by Everybody that's the number of private snips Divided by the number of people that are sharing that snip and multiply by the average number of years for snip You don't need a graph and all rest of you can do it yourself The less kits and groups you have the more unreliable it's going to be we were looking at statistical process You want a big database two methods why fall and Ian McDonald Why fall and my thanks to to John Cleary for this Took me a bit of time to work this out But they work in years before present and they assume that the test that was born in 1960 Ian McDonald works in absolute years, but if you want it years before present he Back to 1950 so one rather you have to add or subtract 60 or 70 to get a comparison Why fall imply they use a hundred and forty four years, but in fact if you look at it, they they jiggle around with it And as I'll show you for us. They're using 156 Ian McDonald recommends 125 But small print up to 197 depending on the quality of the test now. This is beyond me It's very important. I'm not dismissing it, but we're getting into a very gray area here Or even the experts aren't aren't sure But still you can do the maths. Here's here's what you get when you go to y-full It's all there and you can decode it There's the 700 looking for 850. There's 850 down the bottom 850 you get out of it So if you go to y-full they did five kits of 150 eight years if you did one of them you get 1330 and if you use the Five five kits they looked at which is quite a small sample There's not many people in our project gone to y-full other people all the people gone to y-full Which makes the prediction more accurate, but they come up with 1170 If you do it by Ian McDonald's method, he's giving you slightly different things when you come up with 1310 with with the Hapla group 19 the old big y with the new big y I've gone up from five snips to seven snips under l55 as an average But I believe that Ian McDonald hasn't yet gotten years for generation for hg-38 So I'm saying results not available because by definition if I multiply something by seven I should be getting a similar answer Multiply something else by five. They should come out Similarly, I don't know the answer for that yet, but I don't think he's done the work on it So last slide of substance Documentary sources 1250 to 1350 37 markers strs 1410 111 markers 14 1460 and two different the snip ones so they're all in the same ballpark And you can start fiddling it and kid yourself if I disregard this one and tweak that one You can get them all to be equal each other But who are you kidding? There's so many uncertainties that we know it's roughly in that ballpark So I know that my surname is not as old as Morris's Soname, but it's younger than somebody else's sonate. It's that old of a magnitude So please don't hang a hat on the fact that you can work it out I mean, I've rounded them there for the nearest 10 years. They should be rounded for the nearest 50 years or 100 years So that's it What I try to get over today is that large surname projects and offer insights that you can't get the small surname projects Snips as everyone else to say it off a much more potential, but it costs more money Develop your own spreadsheets both for str results and haplitries when you're grouping do remember that four by thirty seven or six by sixty seven Is a very arbitrary limit and it includes some rubbish and it excludes some good stuff And when you're doing TMR CAs Just be careful Thank you very much Thanks very much James Now obviously all of this DNA work that you're doing is Being used in conjunction with a lot of the documentary evidence. So where did you get your documentary evidence for the Irwin's? Well, I started 60 years ago And in 2005 this American said he was going to do it all by DNA I could chuck all my documentary evidence out the window and I said not in my Nellie You're gonna tell me how I'm gonna ditch all the documentary work It's hard slog of public records office in Edinburgh just like the public records office here has got a lot family archives were very lucky Bonchamps got quite a bit of material a drum has got 20,000 documents, which I've helped index the actual raw data up in Albany I've looked at every Irwin reference there is and I couldn't do a one-name study on the Albany Irvins There's too many of them. It's just hard work and it's what we all do, but I enjoy doing the elder stuff I'm not awfully interested in finding my third cousin. I'm much more interested in finding what happened in 1460, but it's just many years of It's not hard work The reason I ask is because I guess a lot of the people in the audience will be doing or interested in doing their own surname study So that means going to the repositories the archives the libraries and looking at the old documents for that particular surname of interest But is there anything online? For example, do you know of any good online sources where people could at least start the study from the comfort of their own home? I know lots of Irwin online sources that are all I wouldn't even say tertiary the four layers of iteration and they've taken books and so forth some good some very good books published I could rattle them off and very good books being written on the family, but there are no details and they're dated and they're You know, they've been done best intentions and I'm writing a book on it all now And I know in 50 years time people are gonna say well, you know, he thought he knew everything But we now know much more than here when you that's the way of life But when you go on to the web and you look up Irwin family history You will find gushing things about Robert the Bruce and his armor bear and all the rest of it. I mean, it's gospel But it's not quite the same as that one It's not quite the same as that one and when you look at the original material which I found here in the public record Office in Belfast has got more on the Irvins of Scotland that you find in Edinburgh and I find the 16 80 document Just like a web, you know, tertiary web fellow today He was writing down what he'd been told it was a lovely flowery story and those days they weren't too worried about accuracy But because it's 1680 we all think it's gospel. It's crap. Excuse me Particularly when you have DNA and it shows the drum and bonchor which he said were all had the same ancestor We know another children cheese not very romantic and everybody was disappointed including myself, but you know the facts Come out and you want to bury them. You dirty apparel But I think most of us have been through this sort of thing in the family history one way or another Oh, yes, never let the truth get in the way of a good story Questions for James How many people are actually doing their own sir a surname study? The few people right? Okay Jeb, but Debbie you had a question Take into account It's purely STRs and he does make some rather subjective assumptions He talks about relative distance and we all have our hobby horses. I usually on about Tips and so forth. I don't really understand it, but to me its value is that it's one rigid former of the uses and Whether it's accurate or not is less important than the fact you can apply it to lots of different Sir names and you can make a comparison on the like-for-like basis Honestly, not too worried about how accurate it is. It happens to come out about the right way and perhaps more important It's not that I'm right, but he's about in the middle of everybody else So he's obviously hitting it about right. That's more important the fact that I guess the same roughly the same as me Right, but it's very easy to use. I used it a couple of times So once you get in the hang of it, you can rattle through half a dozen surnames in five minutes Great fantastic any other questions for James at all Okay, well it just remains for me to thank you very much for a fascinating presentation. You've done some great work. Thank you very much