AncestryDNA’s new BETA

[jorghes]

On the original dashboard there is a link entitled 'Still want to use DNA Circles', which I presume would flag your choice not to use ThruLines.

Not sure what you mean here. I just click on the 'back' button on my browser and it takes me to the page I was on previously. I notice that after you have been looking at links for a particular ancestor, if you use the 'Back to all ThruLines' button it takes you back to the top of the page (i.e. your parents) but if you use the browser back button it takes you back to the place you were in the list, which I've found much more convenient.

I frequently use the back button, but that is not really my issue. I often try and navigate between several different DNA results and skipping between them is reasonably easy in everything except ThruLines, as once I get back to one of the main pages, I have a drop down menu that allows me to switch tests. I have been clicking the DNA button to go back, and that always takes me back to my own DNA results and then I have to switch, (and that is fully my own usage), while a button the page theoretically would take me back to the DNA test I have been viewing.

NOTE:

I just noticed something interesting - as I have had my entire family DNA tested with Ancestry, the results of my siblings and I are now flagged either "Mother's Side" or "Father's Side", which is rather convenient if I was using them predominantly (which I don't all that much).

However I was just scanning through and noticed that a few results are not flagged with either. What are the odds these people may not be related to us at all?
The top one on my brother's list (who shares DNA with me, and a relative I tentatively flagged as a relative of my father and grandmother) shares about 32 cM across 2 segments, but does not show any relationship with my father or grandmother. Given that's where our DNA for that side of the family is coming from, is this person an anomaly?

 

[peter]

However I was just scanning through and noticed that a few results are not flagged with either. What are the odds these people may not be related to us at all?
The top one on my brother's list (who shares DNA with me, and a relative I tentatively flagged as a relative of my father and grandmother) shares about 32 cM across 2 segments, but does not show any relationship with my father or grandmother. Given that's where our DNA for that side of the family is coming from, is this person an anomaly?

There will be matches that don't show as matches with either your father or your mother - I remember reading a blog article a year or two back which suggested it could be as many as half. The obvious assumption is that they are false matches, but I'm not so sure - I suspect it's a lot more complicated.

 

[Jacqueline]

ThruLines has all the problems you have all identified. I have been on it non stop since I discovered it, working out what is and what isn’t useful. It does not seem to recognise illegitimate births, for example. It likes parents to be married to each other. I gave feedback frequently; Ancestry needs to know where the problems are. I don't like, for example, the fact thatsome ancestors in my trees are not included in their "tiles", nor that other people's ancestors seem to have replaced some of my more remote ones. But I was delighted to discover new matches with credible looking trees at very low levels of cM; I would not bother to follow up many of them now there are so many matches, increasing daily apparently, where the matches are much lower than around 20cM unless there was a good other reason to do so.

But I am still reeling in shock that I think I am now at last able to be certain which ancestor of mine was a N.E.P cuckoo after 3 years of collecting so called 3rd and 4th cousins who were Cunninghams from South Shields. I am a very thorough and careful researcher and am confident that my well resourced paper trail tree is correct and there are no Cunninghams in it. Although my maternal illegitimate grandmother Isabella Proudlock Hepple was born in Hebburn in 1887, I was pretty sure of her otherwise childless father’s identity and I have dna matches to his siblings. She had ½ brothers Glass for whom I also have dna matches. The common ancestors in my new cousins’ trees were John Cunningham born in 1803 and his son James Fairweather Cunningham born in 1824. From the mid 1850s to the late 1860s James Fairweather Cunningham and his growing family lived away from South Shields in Deptford in Kent. This was where my great grandmother Henrietta Hurrion was born in Strood, Medway, in 1866 to William A Hurrion and Henrietta Holloway, an unmarried couple. She was the 7th of their 9 children. Henrietta Holloway was a busy lady. Before taking up with William Hurrion she had 2 sons by different fathers, in 1853 and 1855. William “adopted” John Henry, the second, and John Henry’s many descendants do not accept that they are technically Holloways, and not genetically Hurrions; but they are the only Hurrion dna matches that I have. William had plenty of brothers and sisters and uncles and aunts but I could not find a match to one of them; I pestered those descending from them, messaging them to see if they had tested, but still got nowhere. William and Henrietta married in the summer of 1870 and in December, he died. A week later his last and only legitimate son was born, and died. 3 years later Henrietta married Samuel Ixer and gave him a child, her last.

None of this proved anything about the paternity of my cuckoo: clearly, James Cunningham and Henrietta Holloway/ Hurrion were in the vicinity at the time of the conception of my great grandmother (who pushed my push chair from Hounslow to Romiley in Cheshire when the second V2 arrived too close to home). And Henrietta Holloway was “no better than she ought to be” as they used to say in my youth. But I needed evidence. Another dna cousin advised me to use an X chromosome sheet to enter my ancestors back 5 or 6 generations and highlight those for whom I had dna matches; this would show me where my cuckoo was not to be found. But there were too many gaps; I didn’t have matches for half my families.

Enter ThruLines; using a Lost Cousins ancestor chart this time I entered my ancestors and then got stuck in to ThruLines. Some of the matches confirmed by family trees (where I believed them) were as low as 7 - 15 cM, but family by family, matches appeared and the yellow highlighted names increased. When I had worked through all my mothers’ 8 great grandparents and 16 2 x great grandparents I was stunned. There was one white patch in the middle of my ancestors - for one great grandfather, William Hurrion, and 2 white patches for a 2 x great grandfather and 2 x great grandmother, Williams’ parents. Otherwise the sheet was yellow.

I am 97% convinced; I don’t expect a Hurrion match to one of William’s uncles or aunts to appear now, and after knowing for 76 years that my great grandmother was called Henrietta Hurrion, I think I must get used to knowing that she should have been called Cunningham. I have amended all my trees (except 1, for insurance - 3% doubt!) and will go back to ThruLines to collect the low level matches for my Cunningham cousins.

 

[Kane133]

ThruLines has all the problems you have all identified. I have been on it non stop since I discovered it, working out what is and what isn’t useful. It does not seem to recognise illegitimate births, for example. It likes parents to be married to each other.

Yes, I think Ancestry has got us all busy with ThruLines and/or New & Improved DNA Matches 'BETA'! Although the two improvements are obviously linked they are separate algorithms with different approaches to the same ‘problem’. AND like all computer software, they have their limitations and can never be 100% correct all the time.

ThruLines searches for connections using “names” (and possibly location) and with surnames not being a reliable indicator of relatedness many false positives and incorrect connections are made as has been attested in a number of posts. As the main criteria for a ThruLine search is a surname it will not and cannot be expected to find a name change or NPE.

Whereas New & Improved DNA Matches’ first focus is DNA then looking for overlapping trees that include the same ancestor(s). This results in matching that jumps over brick walls and see around name changes and provide many more new potential ‘green leaf’ Common Ancestors which are increasing almost daily. My old ‘green leaf’ count is 24, then it jumped to 39 when I first looked at BETA 3 days ago and is now 51! Having at least briefly investigated each new ‘green leaf’ it appears that New & Improved DNA Matches is at least 90% correct compared to my ThruLines experience which indicates the potential ancestor found using the surname approach are incorrect about 90% of the time. So I have been ignoring ThruLines and just looking at New & Improved DNA Matches with the ‘Common ancestor’ filter turned on.

As to evidence as to which works best, I have a 19th century case of identity change (if not theft) in my tree … Benjamin Major changed his name to William Wilson. ThruLines provides a lot of incorrect potential ancestors with ‘Wilson’ connections whereas by following the DNA, New Matches BETA has found known plus previously unknown DNA connections that are true matches. So while ThruLines has been underwhelming for me New & Improved DNA Matches 'BETA' is most powerful and a real winner. Congratulations to Ancestry.

 

[jorghes]

Congratulations on finding the solution to your cuckoo problem, Jacqueline!
I know how difficult it can be to locate the parent of an illegitimate child - although yours seems to have been disguised! - I have a similar ancestor who had two illegitimate children (to, I'm fairly sure, two different men) before marrying a third and having another 12 or so. My great-grandfather took his mother's maiden name however, rather than the name of his stepfather, although he would list his step-father on his marriage documents, and that makes it a lot easier to spot his illegitimacy - unfortunately no signs yet of his father through DNA, although I have a few on the list I can't explain that may be relatives of his - not enough information on either side to say as yet.

I think that Ancestry has a bunch of kinks to work out - for example I think ThruLines should default to the user's tree rather than the perception of the "correct" ancestor from other trees.

However on the whole, I think this is going to be a good, massive step forward for the ability to use DNA easily on their platform, and shows signs of allowing everyone to access the benefits.

 

[Helen7]

ThruLines searches for connections using “names” (and possibly location) and with surnames not being a reliable indicator of relatedness many false positives and incorrect connections are made

I agree the focus is on names, and the 'potential ancestors' it displays are usually wrong, so I've found these a waste of time and filter them out of the display. If you ignore the potential ancestors and just focus on the actual ancestors, then I've found ThruLines very useful, not least that it displays groups of DNA matches on one screen and shows how they are related to each other. For example, I have 2 ancestors who each have 11 DNA matches descended from them which I can see at once, and see where these matches are close relatives of each other.

By contrast, the 'New & Improved DNA matches' only looks at one match at a time, and it too looks at names to match ancestors. I agree it is useful in discovering connections to specific DNA matches, but surely the 'common ancestors' (of the right names) must be there in your tree and others' trees to get a match. I must be missing something, but I can't really see how it can 'jump over brick walls and see around name changes.' As far as I can tell looking at my list, it will only report a 'common ancestor' if that ancestor has the same name in the different trees.

I also notice that, with the 'New & Improved DNA matches' disabled, I have only 11 'shared ancestor hints' (i.e. my old 'green leaf' count) compared to 42 'common ancestor' matches with the new BETA enabled. It is not surprising the numbers have increased, as Ancestry is drawing on its whole database of trees to find matches, but what I find surprising is that 3 of my old 'green leaf' hints do not appear in the new list, despite us having common ancestors in our trees. Must be a glitch somewhere.

Clearly they aren't perfect, but I find both ThruLinks and the New & Improved DNA matches Beta useful in different ways, and both can be powerful tools if used correctly. I've certainly found new leads using them.

 

[Helen7]

I think that Ancestry has a bunch of kinks to work out - for example I think ThruLines should default to the user's tree rather than the perception of the "correct" ancestor from other trees.

I agree, very annoying when your actual ancestor is replaced by an imposter from another tree! Fortunately, I don't have too many examples of this, but enough to be irritating. If we all send feedback to Ancestry about this, perhaps they'll do something about it.

 

[Helen7]

But I am still reeling in shock that I think I am now at last able to be certain which ancestor of mine was a N.E.P cuckoo after 3 years of collecting so called 3rd and 4th cousins who were Cunninghams from South Shields.

How satisfying to finally solve this conundrum, and well done on unravelling the mystery. I think many people will have illegitimacy issues in their trees, which DNA is finally resolving. I have an ancestor who was born 4 years before his mother married and, although he took his mother's husband's name, I was always unsure who his birth father was. However, I now have DNA matches to relatives of the (step)father so am reasonably confident that he was the father (having said that, there is a possible cousin marriage here, so I can't be 100% sure!)

 

[Bob Spiers]

which ancestor of mine was a N.E.P cuckoo

An excellent posting Jacqueline and very interesting and informative. But I must ask you to explain the acronym (or initialism a better term) 'N.P.E'. I dislike their use on principle unless very common or garden and in everyday use, or first spelled out in full. When encountered and unfamiliar, I positively refuse to second guess their meaning, even if I might get close. So what please is an N.P.E cuckoo?

 

[Bob Spiers]

This results in matching that jumps over brick walls and see around name changes...

I have to agree with Helen (who I see has already singled out this part of your posting) I think this a little over the top in what can be expected from the new Beta features introduced by Ancestry. Certainly not in their present incarnation. However I do agree the 'New and Improved DNA' is likely to be the winner in due time.

 

[Andrew Lloyd]

very annoying when your actual ancestor is replaced by an impostor from another tree

My feelings are that it could be counter-productive to always default to the users tree. I am happy to have an impostor from another tree, in the knowledge that I could be an impostor on somebody else's suggestion. After all we all know that there are very many incorrect trees out there. Having an impostor gets you thinking and in the end you can always reject the suggestion.

 

[Bob Spiers]

On the original dashboard there is a link entitled 'Still want to use DNA Circles', which I presume would flag your choice not to use ThruLines.

Sorry cannot find this - can you confirm still there?

 

[jorghes]

Sorry cannot find this - can you confirm still there?

Still appearing on my dashboard - and all of the other DNA dashboards, ironically except for one (unsure as to why) if it is appearing, it should be directly above the green "Explore ThruLines" button.

 

[Bob Spiers]

Still appearing on my dashboard - and all of the other DNA dashboards, ironically except for one (unsure as to why) if it is appearing, it should be directly above the green "Explore ThruLines" button.

That clinches is then. I only have two dashboards and it appears on neither.

 

[Helen7]

My feelings are that it could be counter-productive to always default to the users tree.

I am happy to see suggestions for potential ancestors which go back further than I have in my tree. Clearly they may prove useful in further research (even if they prove to be mostly wrong, there might be something useful there). But if I have carefully researched and found an ancestor that I am sure about (with evidence to back up my conclusions), then for ThruLines to replace them with someone called just 'Esther' (for example), with no sources in the tree to back it up, seems counter-productive to me. What I would like is for ThruLines not to replace existing ancestors in my tree, just to suggest potential ancestors from earlier generations.

 

[Pauline]

in the end you can always reject the suggestion.

Except you can't in Thrulines. Whether an impostor has replaced a genuine ancestor or gone into a gap, there is nothing you can do to reject them except hope the other tree owner will correct their mistake.

Admittedly the impostors in the gaps can just be ignored but you can't get back the genuine ancestors that have been displaced, and so cannot see if ThruLines might have had something useful to say about them. I know I have matches via some of my displaced ancestors.

 

[Pauline]

N.P.E

Non Paternity Event or Not Parent Expected.

 

[Helen7]

That clinches is then. I only have two dashboards and it appears on neither.

Same here. I have no button to go back to DNA Circles on my dashboards on both Firefox and Chrome. I thought it was because I was never in any DNA Circles.

 

[Pauline]

That clinches is then. I only have two dashboards and it appears on neither.

Were you in any DNA Circles before? Looking at the different DNA tests in my account, the two that were previously in circles have the option to switch back. The two that weren't don't.

 

[Bob Spiers]

Were you in any DNA Circles before? Looking at the different DNA tests in my account, the two that were previously in circles have the option to switch back. The two that weren't don't.

Thanks Pauline, that resolves the matter as neither of my Trees had been admitted to a DNA Circle. I dare say that applies to the odd one out for jorghes?