DNA Shared Matches

[Bryman]

I think that I must be beginning to lose it, or am just expecting too much from the data. Perhaps someone can suggest why I am getting confused.

I understand that A matching B and A matching C does not necessarily mean that B matches C, because the matches may be in different segments.
ie if A=123 and B=145 then the match can be in segment 1,
and if C=267 then A and C can match in segment 2 but there is no match between B and C.

However, if A and B were to share a match with C in any segment, then why wouldn't A and C share a reverse match with B in that same segment? When I try to extract similar information from my results in ThruLines, I find that the reverse match is not identified in several instances, and I don't understand why.

Unfortunately, my laziness is now catching up with me where I have not fully researched all branches of my tree coming forward since about 1800 ( ie all 2G Grandparents and their descendants). I am trying to use this shared match information to help identify where in my tree these other individuals might be found. When I contact some of my potential DNA relatives with perhaps 30-40 cM of DNA shared with me, they are convinced that we are not related because they can find no shared matches and the tree linked to their results, if any, only goes back 1 or 2 generations (and even that data is private because those individuals are still alive).

 

[peter]

Autosomes come in pairs and so it is quite possible to match two unrelated people in the same apparent position.

It's quite possible, even likely, that you will have many matches with people who share 30-40cM, but are such distant cousins that you'll never find out how you are related. See the coloured chart in my Masterclass. However, in the cases where your cousins have minimal trees it's more likely that you are related.

 

[Pauline]

I’m not quite sure if it’s what you are asking, but Ancestry doesn’t list people as shared matches if they are in the Distant cousin list - which can account for people not showing up as shared matches even though it seems they should.

 

[Helen7]

However, if A and B were to share a match with C in any segment, then why wouldn't A and C share a reverse match with B in that same segment? When I try to extract similar information from my results in ThruLines, I find that the reverse match is not identified in several instances, and I don't understand why.

I think it is a case of your potential cousins' matches (with each other) falling below the 20 cM threshold, as Pauline indicates. So their shared match list doesn't reflect yours. For example, I have a match of 43 cM with one potential cousin (B), but my son (C) only shares 16 cM with him. So whilst I appear as a shared match with B on my son's list, the reverse is not true, i.e. my shared match list with my son does not include B, as the B-C match is < 20 cM.

 

[Bryman]

See the coloured chart in my Masterclass.

Yes, I even have this chart printed next to my screen for easy reference although that may not be as frequent as perhaps it should be.

 

[Bryman]

I’m not quite sure if it’s what you are asking, but Ancestry doesn’t list people as shared matches if they are in the Distant cousin list - which can account for people not showing up as shared matches even though it seems they should.

That was not my main concern but is certainly something else to consider. Although I understand Peter's insistence that most of our matches will be with distant cousins, I am still more interested in finding 2nd-4th cousins rather than 6th-8th cousins, and so far, that is what I am achieving.

 

[Bryman]

I think it is a case of your potential cousins' matches (with each other) falling below the 20 cM threshold, as Pauline indicates. So their shared match list doesn't reflect yours. For example, I have a match of 43 cM with one potential cousin (B), but my son (C) only shares 16 cM with him. So whilst I appear as a shared match with B on my son's list, the reverse is not true, i.e. my shared match list with my son does not include B, as the B-C match is < 20 cM.

Thanks for further explanation. I had not realized that there was a 20 cM threshold. That seems rather high. The vast majority of my matches are below that figure.

 

[peter]

Having shared matches doesn't necessarily increase the chance of a match being genuine/useful. Many single segment m

That was not my main concern but is certainly something else to consider. Although I understand Peter's insistence that most of our matches will be with distant cousins, I am still more interested in finding 2nd-4th cousins rather than 6th-8th cousins, and so far, that is what I am achieving.

We all have far more distant cousins than we do close cousins. I would be surprised if even 1% of your Ancestry matches are in the 4th cousin or closer category, and many of them will turn out to be more distant than 4th cousin (indeed, Ancestry will show many as 4th to 6th cousin).

However, for most researchers testing atDNA isn't about finding cousins, it is about knocking down 'brick walls' - and matches with distant cousins are more likely to provide a vital clue (since they're more likely to be on the other side of the wall). The strategies in my Masterclass are designed to knock down 'brick walls': if your main interest is finding close relatives then conventional research will be far more productive than using DNA.

 

[Bob Spiers]

if your main interest is finding close relatives then conventional research will be far more productive than using DNA.

Agreed

 

[Bryman]

We all have far more distant cousins than we do close cousins. I would be surprised if even 1% of your Ancestry matches are in the 4th cousin or closer category, and many of them will turn out to be more distant than 4th cousin (indeed, Ancestry will show many as 4th to 6th cousin).

You are correct. Approximately 0.5% of my matches are shown as 4th cousin or closer and approximately 0.5% of them have common ancestors identified from the linked trees. Most that I have investigated so far have turned out to be slightly closer related than the Ancestry prediction but that might just be my good fortune.

if your main interest is finding close relatives then conventional research will be far more productive than using DNA.

Generally I agree but what I was trying to explain is that I have more interest in breaking down a brick wall to expose a 2nd or 3rd cousin than a 6th-8th cousin. That might not be my intention when I start my investigation but finding a previously unknown 'close' relative is more satisfying than learning about some that are much more distant. Having said that, I did get much help from a 7th cousin that was identified through LC and seemed to be much closer related when communicating. I have probably done more conventional research than most LC members, but certainly not all, and get pleasure from identifying family members, however close/distant, using whatever tools are available. DNA matches do have the advantage of enabling a channel of communication to be opened which is not generally available otherwise, except through something like LostCousins.

As I mentioned earlier in this discussion, I have not traced all descendants of my 2G Grandparents and may never accomplish that due to hard to find/confirm certain records but when my DNA matches lead to 2 previously unknown and unrelated 2nd cousins now living here in NZ, I find that more interesting than almost all other discoveries. I only ever had two 1st cousins, in UK, and one of them has already passed away. From the results of my DNA test, I have now learned about over 100 2nd and 3rd cousins living just in NZ.

 

[peter]

I have more interest in breaking down a brick wall to expose a 2nd or 3rd cousin

I suspect you're using a different definition of 'brick wall'. Most researchers regard a 'brick wall' as something that prevents them researching further back on one of their ancestral lines (see for example this article from an American genealogy magazine) - and inevitably most of these 'brick walls' are many generations back.

I appreciate that it can be more difficult for researchers outside the UK to track living cousins using online resources. I don't think I've yet found through DNA any 3rd cousins in the UK who weren't already on my tree, but I have have found quite a few overseas.

 

[jorghes]

I have found that using the colour coding system, even to just mark family groups, helps enormously. I found two more close relatives of my father's appear on his results when I was considering what to write here, and while I can immediately place one (I think I had a conversation with someone about her grandparent and our link not long ago), the other I don't know where they fit - and they're a "2nd cousin" - but I do know what family group they belong to, as every shared match appeared with a little yellow dot, and most of them with a note marking exactly where they fit.
And this is my Australian side of the family - emigrated in 1839/1840 which can be challenging to trace. It may be helped by the branch which headed to the US, but the majority I have been in contact with are still in Australia.

I have a colour for each main family group that someone attaches to and I often also use colours to identify groups of DNA results which match, but I don't know where they fit - they are entitled "Unknown Family Group".

Admittedly, I am very lucky in the results I can access - and all of the results have "2nd cousin", "3rd cousin", "4th cousin" and "distant cousin" results (namely those for my parents and paternal grandmother).

 

[Bob Spiers]

I have found that using the colour coding system, even to just mark family groups, helps enormously

At the risk of asking a silly question, especially if covered previously and missed or forgotten by me , how does one colour code family groups in Ancestry? I know how to do so in (say) FTM but - apart from assigning 'groups' to certain Ancestry matches (and then applying a colour code) - I do not think this is what jorghes means. Can someone kindly enlighten please.

 

[jorghes]

At the risk of asking a silly question, especially if covered previously and missed or forgotten by me , how does one colour code family groups in Ancestry? I know how to do so in (say) FTM but - apart from assigning 'groups' to certain Ancestry matches (and then applying a colour code) - I do not think this is what jorghes means. Can someone kindly enlighten please.

Part of the new DNA results page includes an "add/edit groups" on the right hand panel when looking at your matches. Clicking on the plus sign opens a new colour and a labelling system. I'll grab some screen shots of how I use it for you.

 

[peter]

I've colour-coded my matches, but so far it hasn't helped me in any way (since I always add a note saying who the common ancestors are). I'm beginning to think that this feature might be more useful for highlighting clusters of genetic cousins whose connection is unknown.

What do others think?

 

[jorghes]

So, these pictures come from my father's results (so that's why there's a "Mother's Side" and it mentions Parent/Child relationships).

This is the DNA results page, the new addition with arrows:

When you click, you get this:

I have a lot of groups already - if you add a custom group, this is what happens: (You can change the automatically chosen colour.)

The other ways you can use them are like this in the main list in the "Groups" dropdown menu. The numbers beside the groups are how many results you have added to that group. (You can see the "Mother's Side" marker since I had to be at the top of the list... the last one without the marker is my father's mother!)

You can also add people to a colour coded list when checking in the comparison match:

I tend to click o the "Shared Matches" here, if I know exactly where someone fits - like my half cousin Diane, so that I can add those we share to the list, as it also allows this. I do have a screenshot for that too, but I've gone over my allowed limit.

I hope that clarifies how you can use the colour coding Bob? Unfortunately it's only usable in the DNA results - I presume TreeTags can be used in this way on your tree itself since it has a "create custom" addition.

EDIT:
I think it is useful for adding people to a grouping where you already have some people identified - like the difference between Joan and J. on the first screenshot - I don't know exactly where Joan fits, but I know she's from my paternal grandfather's family (the yellow dot!) - the three yellow dots you can see are all descended from the same great-great grandfather, the one I've mentioned before who married at 16 - and he was born in Australia and never emigrated.

I also use an "Unknown Family Group" one, but not with my father's results as of yet.

 

[Pauline]

What do others think?

I do a bit of both. I have coloured coded all my identifiable matches according to which grandparent they connect to, with a second 'possible' category for those who are shared matches of the above but can't otherwise be identified. I find this useful for quick identification of matches, and for viewing particular groups separate from the rest.

I also have a few further categories for larger 'shared match' groups where the connection is unknown, which can help in sorting the different clusters, and also in identifying any overlaps between them. I've only done this for larger clusters as the number of groups you can create is limited.

 

[Bob Spiers]

I hope that clarifies how you can use the colour coding Bob?

Thanks jorghes, well explained and as it happens, you are referring to the colour coding I meant when I said "apart from assigning 'groups' to certain Ancestry matches (and then applying a colour code)" but so far I have only created one custom one which I call 'Potential' (intriguing, nothing proven, but worth following up). I had not given thought to assigning color codes to separate paternal/maternal/lines, and not absolutely sure this will help as I can place most anyway, and those I can't likely deserve a 'Potential' follow-up category. However, now the matter has been aired I will see what others have to say and see if more extensive colour coding will help.

 

[canadianbeth]

I have done this as well. All my known Joyce ancestors have a specific colour, as do the Riches, Barratt's and Bowyer's on my mother's side. I also have a colour for unknown. If I have known shared matches they go into that specific section. I also have one for "no shared matches" I just wish I had the high numbers that jorghes has I only have a few above 100 and all but one are confirmed as either Joyce or on my mother's side. The third highest - 198/9 segments - is unknown.

 

[Helen7]

I also use colour coding for organising DNA matches into groups. I find it useful for seeing at a glance how they are connected, and for displaying a selected group. I use notes too, but the colour code gives an instant visual indication. I have broken down the coloured groups further than just the 4 grandparent lines as I found that too crude a grouping, and I organise the groups using other common factors I identify along the way. For example, I have one group all descended from a relative who emigrated to Canada, and another descended from ancestors from the same town. So far I haven't used it for 'unknown' clusters of shared matches, but I can see that might be worthwhile too.