DNA WORKSHOP
Welcome to the DNA Workshop! In every issue DNA adviser Karen Evans will be unpacking a reader case study, and giving step-by-step tips for you to use in your DNA research at home. It’s time to unlock the story of your genes…
DNA Advisor Karen Evans helps a reader struggling to make sense of a Thrulines challenge
ASSESSING DNA MATCHES
Reader Gina got in touch to ask for Karen's help in overcoming a Thrulines challenge she had met when assessing her Bailey surname matches at Ancestry DNA
GINA: Dear Karen, I am hoping that you can find a little time to give me some guidance to overcome my DNA frustrations.
My brother and I have both taken an Ancestry DNA test to try to establish who our father Arnold’s father was. For both sets of results, the strongest match is for a Howard (293cm to my brother and 235cm to myself.). He has an unlinked tree of 24 people, so I proved what was there and looked at common and matches. All pointed towards Baileys of Biddulph, Staffordshire, and proving their trees also, led back to an Emanuel Bailey (1803) married to Judith Barnet (1807).
The good news is, that Biddulph is about six miles from where my brother, myself and our father were born. The bad news is that Bailey is by far the most common surname in Biddulph and, when we first moved here, was renowned for its supposed intermarrying with other Baileys and three or four other local families.
I then used DNA Painter using the above few matches plus Howard tree. I parented my father where Howard would be my 2nd-3rd cousin, and no matter how I varied things, the highest scoring position was with an additional generation between my father and the common ancestor, an impossibility to fit within the dates and putting my father, who was born in 1912 in the same generation as Howard who was born in the 1940s.
Expanding Howard’s tree shows that he also has a link via his greatgrandfather Thomas Armitt (1821) who married Ann Bailey (1828). Ann Bailey is a distant relation of Emanuel Bailey. I have been trying to find out how a multiple link affects the value of a match, but have not been successful. Is it the simple sum of what each link would be? Is it this which is inflating the match figures for Howard, and skewing the DNA Painter result, trying to place our father as Howard’s 2nd-3rd cousin, rather than myself?
I decided that the best way forward was use Thrulines results to find the best position. Whichever Bailey grandfather I chose, the Thrulines results did not vary much, with matches below Emanuel Bailey and his wife Judith Barnet. I placed us under John Bailey (1889). This John is the son of Alfred Bailey (1868) and Emily Sherratt (Alfred is a grandson of Emanuel Bailey). It produced a few Sherratt matches, so taking the advice from the DNA video on family-tree. co.uk/videos, I expanded the Emily Sherratt leg, including the maternal lines. It produced a cluster of matches of all strengths, with some to both my brother and me, some to him only, and some to me only.
The results throw up some queries which I cannot find an answer to.
Initially, the strength of the Howard match led me to parent us under Joseph Bailey (1861) and Olive Simcock. This produced the same result for the Bailey matches and unfortunately no matches at all for Simcocks, even going back three generations from Olive. A few more
choices of parenting position also produced zero maternal matches.
Now a series of questions which I hope you can shed some light on, to help to point me in the right direction.
I cannot work out why, when my brother and myself both match, sometimes the figures are similar and sometimes very diverse. Also, for some matches, it is just one of us and not the other.
When I first expanded Howard’s tree, I used all the surnames plus Biddulph as filters against my brother’s and my matches, and by far Baileys of Biddulph gave the most results. Later, when I checked these against the Thrulines results, there were quite a lot of reasonable scoring matches not included. I decided to check their trees and it quickly became obvious that there were some common factors. None of them had many recorded sources other than Ancestry Trees and appear to have used one 6,000-person tree as their source. My checking found some discrepancies which linked some of them into our Bailey tree. Thrulines obviously sees the 6,000-person tree plus everyone else’s copies as the norm and ignores my ‘lone furrow’.
One other thing that these matches have, is a high value for so far back in the tree, some exceedingly high. Is this simply a sign of another link somewhere earlier in their trees, which I cannot easily find?
Stumped by the lack of matches on Emanuel Bailey’s paternal side, I asked my brother to do a Y-chromosome test with Familytree DNA.
I suppose with the number of matches for us both on Ancestry DNA, I expected a reasonable number of matches on the paternal line, but there were just nine. This surprised me and I still cannot find what is the norm?
I contacted the top match, Neil, and he confirmed that he had used the Y-DNA to trace his illegitimate father’s paternal line through a grandson of Simon Beech.
The only fly in the ointment is that although Neil is number one in the Y-DNA matches, he is not a match to either my brother or myself on Ancestry, and even more puzzling Neil does not appear in the autosomal matches for my brother’s FTDNA test. Any ideas? Surely paternal DNA should reflect in the autosomal results also?
Match number two was a lot easier as John Bailey has a skeleton tree on FTDNA which confirmed the Bailey surname link. Neil’s match is Genetic Distance 1 and John Bailey’s match is Genetic Distance 2. What is the significance of this?
I do hope that you can answer at least some of my queries, and if possible, point me in the direction of a possible Bailey grandfather. Thanks in anticipation.
Gina Jervis (née Nixon, possibly Bailey)
KAREN: Gina really has been making her DNA work hard! I was keen to answer some of her questions so that she could move forward with her research.
Why do siblings have different matches or matches with different amounts?
One of the fascinating things about DNA is how DNA between you and your matches is shared. The DNA each sibling gets is completely random so, although you get 50% of your DNA from your mum and 50% from your dad, ‘which’ segments they give you is due to recombination. So, mum may give you lots of DNA from her mother but not so much from her dad, meanwhile your sibling gets more maternal grandfather. This is why siblings can have very different amounts with a match. As an example, my mom had a new match on Ancestry with whom she shares 28cms. I was not initially excited but then saw the match shared 50cms with my aunt and 118cms with my uncle – this match was worth investigating! Mom has many matches which she only shares with one sibling and they all have matches the other hasn’t got. If you look at part of my spreadsheet above, which I use to track matches, you can see the vast range between my mom and her two full siblings. Second cousins and closer always match, after that, it’s all down to the joy of recombination!
DNA amounts & relationships
You have a match, Ancestry tells you they are in the 4th-6th cousin range, so that is where they fit, right? Not necessarily. All the testing companies give you a possible relationship but there are often multiple places that match can fit in your tree. A match of 55cms is in that range of 4-6th cousin but looking at the relationship chart you can see that there are multiple possibilities. However, intermarriage can affect those positions so a match can appear more closely related than they are. In Gina’s case Howard is related in two separate ways (creating pedigree collapse) but the one relationship is quite distant so will have little effect on his overall relationship. If you have a tree which shows your direct ancestors were interrelated, then more caution needs to be taken when looking at amounts of DNA and cousin ranges.
This is a complex and fascinating subject which requires expert
explanation – if you have endogamy (multiple ancestors in your tree sharing common ancestors) then I found The DNA Geek’s blog extremely useful https://familytr.ee/ endogamy. She also did a fantastic talk on her own endogamous family on Rootstech which is still available.
Thrulines-user beware!
One of my ancestors was Mary Fletcher. I know she was illegitimate as I have her birth and marriage certificates. She is the daughter of Eliza Fletcher and granddaughter of William Fletcher. Mary is mistaken in many trees as being the daughter of William so I often get hints through common ancestor hints and Thrulines suggesting I put William as dad. Thrulines is not telling me I'm wrong, it is merely showing me how other trees see Mary – I can take the hint or ignore it. The hint won’t go away if a match has it on the tree. If another DNA match has a tree with another father for Mary, that will also appear as a possible ancestor. Thrulines can be wrong, right or somewhere in the middle. It doesn’t say a Thruline is more ‘right’ if lots of matches have the same information on the tree and you are the only one who believes differently.
Another problem in relying on Thrulines is that they only exist if matches have trees which Ancestry can then use to find possible links. A lack of Thrulines along a particular line doesn’t prove the line is incorrect and finding a Thruline doesn’t prove the genealogical link. Gina shared her Ancestry results and I could see many of the Thrulines in the Sherratt line were below 20cms and even single figure matches. At this level matches could be false. Thrulines can be wonderful for hints, but it is the traditional paper research which validates or disproves them. For me the biggest tool to identifying matches is through clustering – I go where the matches take me and how the clusters of matches relate back to a common ancestor.
The wonders of WATO
What Are The Odds? or WATO is a tool at DNA Painter that allows you to create a tree, add matches and then, using the amount of shared DNA, place the tester in various places as a hypothesis. WATO works out if the place is possible and also how probable the hypothesis is. You can put in various hypotheses and the higher the number, the more likely the position is correct compared to each other.
I created a simplified WATO with Gina in the hypothesis position. Arnold was placed as the grandson of Alfred, which was possible, but the least likely hypothesis based on others I tried. I placed Arnold as child of Joseph and a full and half sibling of Joseph’s other children, and both of these were popping up as more likely than the current place. Based on dates, It would seem very unlikely that Joseph could be Arnold’s greatgrandfather in 1912. I also imagined that Arnold was actually the grandson of John and Letitia through either Joseph or another son. They had come up very highly. Although Joseph and his brothers would be in their late 40s or 50s it doesn’t mean they couldn’t father Arnold.
Using Y-DNA testing
The human Y chromosome is the male sex chromosome which, like the paternal surname (in some but not all cultures), passes down virtually unchanged from father to son. A Y chromosome DNA test (Y-DNA test) is used to find a man’s patrilineal or direct father's-line ancestry. Occasionally and randomly, there are mutations (errors in the copying process), and these changes can be used to estimate the time frame in which two or more testers share a most recent common ancestor or MRCA. Basically, imagine testing a number of places or ‘markers’ on the Y chromosome looking for similarities and mutations. If test results are a perfect or nearly perfect match, men are related within a genealogical time frame.
If you are interested in understanding the science behind Y-DNA testing then there are several websites which explain in much more detail than I can here. I found this one really useful https://familytr.ee/demyst. The key thing to remember is that Y DNA by itself doesn’t tell you who you descend from. Mark’s DNA
is compared to the DNA of other men who’ve tested – the surname and the paper trail connected to that DNA is open to interpretation.
Gina’s brother Mark tested at Familytreedna and below is the Y-DNA test results page, which examined 37 markers, looking for similarities and mutations between the tester and everyone else on the database. He has nine matches: three share the Bailey surname but the other six have different surnames. Although some men have many matches, I’ve found this to be a fairly typical number. The genetic number column tells us how closely the matches maybe related by how many mutations there has been. If you click on the orange TIP button you get a probability of how closely that relationship actually is. Mutations are random so it is impossible to know exactly how many generations without traditional tree-building research. For example, John is a genetic distance of 2 but this could indicate a common ancestor as close as four generations away or as far as twelve generations. According to ISOGG (International Society of Genetic Genealogy), if the two tests match perfectly on 37 markers, there is a 50% probability that the MRCA was fewer than two to three generations ago, a 90% probability that the MRCA was fewer than five generations ago, and a 95% probability that the MRCA was fewer than seven generations ago. The higher the genetic number, the further the common ancestor is likely to be.
There could be several reasons why some of Mark's matches do not share the Bailey surname. For example, my great grandfather was Henry Overfield. There are plenty of male descendants who share his surname and could be tested to breakdown my Overfield brick wall from the 18th ccntury. Unfortunately, Henry’s grandfather William Overfield (born 1815) was the illegitimate son of Mary Overfield, so none of William’s male descendants are following the Y DNA line of an Overfield (I could test them to find what William’s father’s surname might be!). In my case I knew the surname Overfield came from a female, as does Neil, Mark’s highest match. But if any of the matches have a ‘misappropriated father’ then the surname (in Y DNA terms) is no longer the same. Other reasons might include adoption, taking mother’s surname to preserve the name, later hereditary surnames.
Gina knows that some of the Y-DNA matches have also taken Ancestry or Family Finder autosomal tests, so why don’t they appear on Mark’s match list? They are related but don’t share enough/any autosomal DNA. Autosomal DNA testing is most accurate up to five generations, so if these men are more distantly related then there is a high chance they won’t appear on the Ancestry DNA or Family Finder match list. Without the Y-DNA they wouldn’t know they were definitely from the same Bailey line and even then, the shared ancestor could be before any paper evidence.
What next?
Based on WATO, I would look again at the matches who share a lot of DNA but seem distantly related. If none of these show recent intermarriage then they are probably related to Gina in a different way to the one in which they have been placed.
If there is little/no DNA evidence pointing to Olive Simcock then it may be more likely (and indeed WATO believes it is) that Arnold is the grandson rather than great-grandson of James and Letitia by Joseph or one of their sons but Gina needs the higher level DNA matches to make it clearer. I have to say that, based on the matches available, she is presently at an impasse.
At this moment in time Gina can only carry on with what she is doing and look to disprove the possible scenarios given. Based on the matches, it may now be a case of hurry up and wait. It took nearly three years for a match to appear that showed me I was on the right track for my father's missing grandparent and then I was lucky to have a match which pointed (more or less) to the right man. While I were waiting for matches, I would try to gain a response from any shared matches of Howard that haven’t been identified, especially the ones above 30cms. I’d also have a look at Myheritage (the ones of 50cms) and see if contact with matches brought up anything. Nothing may come of it but nothing ventured...