Brick walls & DNA

[Pauline]

Because they didn't get the answers they wanted.

Peter, I accept that my reply to you yesterday evening was probably terser than it should have been, but if you review our exchange above you may realise that it was not what you said but the way that you said it that got under my skin - the all too familiar assumption that yours is the only right way of doing things, and that I don't have the intelligence or experience to have worked out the best way for me to keep in touch with my research cousins. Not only are round robins not my thing - I prefer a more individual and personal approach - I have amassed too many correspondents over the years for this to be viable.

Moreover, I don't think the way I manage my correspondence has any bearing on how I feel about offering to give DNA kits to distant cousins and asking them to let me view or manage their results in return. It may be true that I am being overly sensitive about other people's feelings, but I would much rather be that than risk riding roughshod over them.

So as I suggested before, let's agree to disagree and move on. I was interested to know how you approach distant cousins to ask them to do a DNA test for you, and you've already answered that for me, thank you.

 

[jorghes]

A lot depends on what you regard as a 'brick wall'. I had two ancestors whose baptisms were missing but who fitted neatly into families that were in the right place at the right time - DNA proved the connection. And the birth certificate for my illegitimate great grandmother gave the name of her father - an error by the registrar in the early days of registration - but I couldn't find anyone of that name and occupation (lime-burner) who fitted. DNA proved that her father was an apparently happily-married farmer who was had been described as a lime-burner when his sons were baptised.

I do have the unlikely story of a brick wall which was moved (a generation or so back) through DNA, but without discovering any birth records or other documentation which places my ancestress with her family in the earliest census she would have appeared (1841). Baptismal records are unlikely in this case as my ancestress was Jewish.

But, with help from the wonderful people of the forum, I was able to locate a possible family and then using DNA results, was able to confirm DNA links to descendants of that family to confirm they were actually my ancestors too. I still have not located documentation for my ancestress or her older brother, but at least I know where they fit.

Have the links with your DNA matches been that far back? My most frustrating brick wall lies with one pair of 6 x great-parents. I have found documents relating to them both individually and as a couple, but no baptism for either. I have constructed extended family trees going back several generations for both parties, based mostly on parish records and wills, and can argue a strong case for where each of them fits in the family tree, but without a baptism to name at least a father I can't be sure my reasoned hunches are correct. Would DNA really help that far back? Plodding through my matches looking for surnames and places hasn't thrown up anything useful so far.

My grandmother's results have seemingly found results for what would be my 8th great grandparents. But I place a waiver on it - they are Ashkenazi Jewish (which is a fairly insular community) and there is some pedigree collapse in that side of the tree (my 4x great grandparents included a pair of sisters) as well as other instances of intermarriage within the same families, which could have effected the amount of DNA passed down.

 

[peter]

which could have effected the amount of DNA passed down

Just to clarify for anyone who is new to DNA, the amount of DNA passed down from parent to child is always the same, but whilst half comes from each set of grandparents, the proportion inherited from each grandparent varies.

Pedigree collapse means that the same ancestor(s) sit in multiple places in your tree, and their ancestors are also duplicated. For example, if you have two direct ancestors who were sisters, all of their ancestors will appear twice in your tree. This means that, on average, the amount of DNA inherited from those ancestors will be twice as much as for other ancestors of the same generation, so you're more likely to have matches with distant cousins who share them. However in itself this doesn't make matches appear closer than they really are once you take into account the fact that you are (say) double cousins.

An additional complicating factor is that in any group which tends to intermarry - not just religious groups, but also isolated colonists (eg the American colonies), or people who speak a different language - certain stretches of DNA will be more commonly found than in the general population. This will almost always make matches appear closer than they really are. For example, whereas most people have 200-400 close matches at Ancestry, people whose ancestors belonged to one of these groups could have thousands or even tens of thousands of close matches.

 

[Pauline]

I have found very little pedigree collapse in my tree, and where I have found some, it is probably too far back to have much effect on my DNA. However, in some parts of my tree there is plenty of intermarriage involving people who are closely related to my ancestors but not in my direct line. This presumably could affect the amount of DNA I share with their descendants.

But perhaps more importantly, one of my illegitimate great great grandmothers is in a part of my tree where I've found quite a few cousin marriages and so on. So I feel I cannot rule out the possibility that my GGGM was somehow related to her husband, and this could muddy the waters somewhat in trying to identify her unknown father.

 

[peter]

I have found very little pedigree collapse in my tree, and where I have found some, it is probably too far back to have much effect on my DNA. However, in some parts of my tree there is plenty of intermarriage involving people who are closely related to my ancestors but not in my direct line. This presumably could affect the amount of DNA I share with their descendants.

That's an example where the pedigree collapse is in your cousins' trees. With DNA there are always two sides to every story.

But perhaps more importantly, one of my illegitimate great great grandmothers is in a part of my tree where I've found quite a few cousin marriages and so on. So I feel I cannot rule out the possibility that my GGGM was somehow related to her husband, and this could muddy the waters somewhat in trying to identify her unknown father.

I don't think it's something you need to worry about at this stage - it's unlikely, particularly in these circumstances, and even if it were the case it wouldn't necessarily have any direct impact on your analysis.

 

[Pauline]

That's an example where the pedigree collapse is in your cousins' trees. With DNA there are always two sides to every story.

Very true, though I wasn't thinking only about cousin marriages. I've found plenty of those but also marriages between closely related lines in my tree. So, for example, with the illegitimate GGGM I was referring to, her aunt married her husband's 1st cousin, so descendants of that marriage would be related to me on both sides, but it wouldn't result in pedigree collapse in their tree. But it could potentially affect how much DNA we share, couldn't it?

 

[peter]

Very true, though I wasn't thinking only about cousin marriages. I've found plenty of those but also marriages between closely related lines in my tree. So, for example, with the illegitimate GGGM I was referring to, her aunt married her husband's 1st cousin, so descendants of that marriage would be related to me on both sides, but it wouldn't result in pedigree collapse in their tree. But it could potentially affect how much DNA we share, couldn't it?

Yes, there's always the possibility that we are related to someone through more than one line - indeed, it's a certainty if you go back far enough. But if I've understood you correctly, what you're describing isn't a marriage between cousins (because the two people who married had no common ancestors).

There are several examples in my tree where two brothers married two sisters, some of them in my direct line. For example, my maternal grandfather's second wife - my grandmother - was his younger brother's wife's younger sister. As in your case there is no pedigree collapse - because they weren't cousins - but the descendants of the other couple are double cousins of mine. (If I hadn't already tested my only living 1st cousins they would have made a good substitute.)

 

[Pauline]

Thinking more about unknown fathers of illegitimate ancestors, how likely is it that you will be able to identify a particular person from DNA evidence alone?

Of course it may vary depending on how far back the unknown father is in your ancestry and so on, but I've tended to think that DNA evidence is more likely to highlight a particular family or a range of possibilities, which are worthy of further investigation.

It seems to me you would need quite a few reasonably close DNA matches to be able to home in any one particular person, and even then it might be more of a probability than a certainty. Or is this being overly pessimistic?

 

[peter]

Thinking more about unknown fathers of illegitimate ancestors, how likely is it that you will be able to identify a particular person from DNA evidence alone?

Of course it may vary depending on how far back the unknown father is in your ancestry and so on, but I've tended to think that DNA evidence is more likely to highlight a particular family or a range of possibilities, which are worthy of further investigation.

It seems to me you would need quite a few reasonably close DNA matches to be able to home in any one particular person, and even then it might be more of a probability than a certainty. Or is this being overly pessimistic?

In practice it's round the other way - often DNA evidence will be the only hard evidence that you will ever have. Anything else will be purely circumstantial: crime writers talk about motive, means, and opportunity, but when it comes to a topic like this the only one that usually matters is opportunity.

Once you've identified a particular family using DNA the challenge is to identify the precise culprit, something that is more difficult when some of the candidates are brothers.

But usually that's the easy bit. It's the initial identification of the family that's most difficult, because you generally start off with no information at all. It's rather like trying to do a jigsaw puzzle when you haven't seen the picture on the box, there are several pieces missing, and most of the pieces you've got come from different puzzles, some of which are vaguely familiar because you've done them in the past, but some of which you've never seen before.

Nevertheless, difficult though it might be, the answer is there in the data. The only thing that can prevent us finding the answer is a lack of time, resources, or determination - and in my experience it's the latter that is the key, because if we don't believe that a problem is soluble we're unlikely to solve it.

 

[Pauline]

I certainly believe it is possible to find an unknown ancestor through DNA, as I've already done it with the unknown father of a 3 x great grandmother, although in the end it was documentary evidence that enabled me to narrow it down to a specific individual from a range of possibilities. But in this case I had a cluster of related matches in the 20-35cM range, and although it was only a relatively small cluster, it was enough to enable me to focus on a specific family. I think I probably did get a bit lucky with this one, and sometimes things may be rather more complicated.

And I would definitely agree that identifying the right family can be the real challenge, particularly if all you have to go on are a number of unrelated distant matches, all of whom have numerous relatives who had opportunity. And at present, despite investing lots of time and determination, that is where I am well and truly stuck.

 

[peter]

And I would definitely agree that identifying the right family can be the real challenge, particularly if all you have to go on are a number of unrelated distant matches, all of whom have numerous relatives who had opportunity. And at present, despite investing lots of time and determination, that is where I am well and truly stuck.

If, like me, you have multiple illegitimate ancestors and 'brick walls' in the early 19th century the challenge is to figure out which genetic cousins are in which part of your tree - and that's where shared matches and known cousins who have tested can help.

I'm as stuck as you are on my outstanding illegitimate ancestors - there are other 'brick walls' which are likely to be easier to break down, so it makes sense for me to focus on them for now, and wait for more DNA evidence to turn up.

 

[canadianbeth]

DNA has not helped me at all in finding my paternal grandfather. I do not know the name of the family where my grandmother was sent from the workhouse. I do have a few high cM matches that do not appear to belong to the other three branches, but none are willing to reply to my messages. My Dad is in the 1911 census with a foster family and grandmother is working elsewhere. Any cousins would be half at most anyway, since I do not believe that the same man fathered my aunts. I have no DNA matches at all to any other paternal cousins related to my aunts - Merpaw or Lalonde. I have however found a lot of ancestors from the other three branches.

 

[Tim]

Pauline/Peter, do you have any X DNA matches with matches you can't place in your tree? This helps to narrow down which part of the tree the match is in.

Unfortunately Beth, X DNA won't help with a paternal grandfather.

 

[peter]

DNA has not helped me at all in finding my paternal grandfather. I do not know the name of the family where my grandmother was sent from the workhouse. I do have a few high cM matches that do not appear to belong to the other three branches, but none are willing to reply to my messages. My Dad is in the 1911 census with a foster family and grandmother is working elsewhere.

Finding a grandfather is relatively easy, because roughly a quarter of your matches will be connected to you via each of your grandparents, and even if as few as 10% of them are connected via your unknown grandfather that should still provide you with upwards of 20 close matches and hundreds of thousands of more distant matches.

Any cousins would be half at most anyway, since I do not believe that the same man fathered my aunts. I have no DNA matches at all to any other paternal cousins related to my aunts - Merpaw or Lalonde. I have however found a lot of ancestors from the other three branches.

Excellent - you can eliminate the matches who are connected to you via the three identified grandparents, and focus on the matches that remain.

 

[peter]

Pauline/Peter, do you have any X DNA matches with matches you can't place in your tree? This helps to narrow down which part of the tree the match is in.

Ancestry only use autosomal DNA. If for some reason I was particular intrigued by an Ancestry match I could ask them to transfer their raw DNA to GEDmatch or FTDNA, but it's a long shot since the chance of a meaningful X-DNA match is very low. There are certain ancestors from whom I'm more likely to have inherited X-DNA than others (see this chart) but the same is unlikely to apply to my match.

 

[peter]

Ancestry only use autosomal DNA. If for some reason I was particular intrigued by an Ancestry match I could ask them to transfer their raw DNA to GEDmatch or FTDNA, but it's a long shot since the chance of a meaningful X-DNA match is very low. There are certain ancestors from whom I'm more likely to have inherited X-DNA than others (see this chart) but the same is unlikely to apply to my match.

If one of my matches was prepared to collaborate I'd prefer them to make me a Viewer - which I suspect they'd be more likely to agree to, since it would involve less hassle for them, as well as less perceived risk..

 

[canadianbeth]

Excellent - you can eliminate the matches who are connected to you via the three identified grandparents, and focus on the matches that remai

I have endeavoured to do that, with my second-closest match, who is a first cousin once removed on my mother's side, daughter of a male cousin. Anyone who comes up as a shared match with her goes into a probably not paternal group. My closest match is my sister's daughter and she would match with both sides. And of course, I have groups for Joyce, Barratt, Riches and Bowyer, all of whom, except Joyce are in my mother's two branches. Joyce is my maternal grandmother. But there are so many unknowns, and no shared matches, and no trees, etc. And as I said previously, the ones with high cM do not reply to my messages. The closest one has a tiny tree and the persons in it are from a different county altogether. One person, with a cM of 99/7 did reply but none of the names he gave me matched anything I have or could find, and I did not hear from him again after I replied. I only have one male relative, my son, who could take a DNA test, but half of his results would be connected to my husband. My brother is gone and other paternal male cousins are either Merpaw's or Lalonde's.

 

[peter]

Excellent - you've made a great start with that list of names from your 99cM match. There must be a good chance that one of those surnames is the surname of your unknown grandfather.

How many of your close matches have you been able to positively identify as coming from the three-quarters of your tree where you know your grandparents? Have you been able to positively identify any others (apart from the 99cM match) as connected via your unknown grandfather? How many matches does that leave which are currently indeterminate?

 

[Pauline]

Any cousins would be half at most anyway, since I do not believe that the same man fathered my aunts.

This would only be the case for cousins descended from your unknown grandfather - that is, a first cousin of some remove. More distant cousins are likely to be full cousins, since you and these matches will likely descend from both of your grandfather's parents, grandparents and so on.

You would be quite lucky to find a match who is a descendant of your unknown grandfather, and are more likely to find matches where your nearest common ancestors are further back.

The closest one has a tiny tree and the persons in it are from a different county altogether.

Being from a different county doesn't mean the match is not related to your grandfather. After all, if you don't know who your grandfather is, you don't know where he was born or where his parents and earlier family came from.

And while I am not suggesting it is what you are doing, a common mistake I have come across when people are looking for an unknown ancestor is the assumption that the unknown ancestor will appear as an ancestor in the tree of any relevant match. The chances are that they won't and that your nearest common ancestor(s) will be further back in their tree.

 

[JoyNor]

Well I can certainly agree with Peter in post #29 above

I got the most astonishing DNA brick wall demolition this week. A "cousin" by marriage only (her mother was married to my mother's brother, who was Jane's adopted Dad) recently took the test. She had absolutely no idea who her father was - her Mum had never told her anything and he was not on her birth certificate or adoption papers. So there was zilch paper trail to even consider. But with children and grandchildren now wanting to know their family history Jane decided to give it a go. After 3 months of only seeing small DNA matches and getting a bit despondent we suddenly last week got a new 358cM/13 seg match for Jane. Fortunately the match was keen to correspond, had good family tree information and records and with his help after just 6 days of checking all shared matches and narrowing down we had 99% certainty on the elusive father's name. The DNA match being her 1C1R. A case of eliminating all the impossibles left only 2 options of which one was felt by his family to be spot on as right man, right place, right opportunity for a brief romance with Jane's Mum at the latter end of WW2.

On the other hand I have spent 12 months trying to convince a 5th cousin, to whom I can prove my relationship with BMD records, censuses, poll registers, probate records etc, that we ARE related despite having no DNA match. Because up to 20 other trees on Ancestry have, like him, picked up (copied) the wrong husband in a marriage in early 1800s and I have the correct husband it has taken yards of correspondence to finally convince him I am right and 20 other people are wrong. But of course it was only one person who was wrong . . . 19 others copied the erroneous information.
Finally this weekend my 5C accepted any DNA was probably too remote and diluted and has agreed my correction to his tree! Phew!