Ancestry Matches 7cM and less

[Pauline]

I have been wondering if I should actually remove the notes and group assignment for those Common Ancestor matches in the 6-8 cM range where I know the suggested connection to be spurious, so as to allow these matches to disappear from my list.

Having read the blog article mentioned above, this still seems like it might be a good idea but I am wondering if anyone else has any thoughts on this.

 

[peter]

I have been wondering if I should actually remove the notes and group assignment for those Common Ancestor matches in the 6-8 cM range where I know the suggested connection to be spurious, so as to allow these matches to disappear from my list.

Having read the blog article mentioned above, this still seems like it might be a good idea but I am wondering if anyone else has any thoughts on this.

The fact that the Common Ancestors connection is spurious doesn't necessarily mean that the DNA match is also spurious (or vice versa, in Blaine Bettinger's opinion). But if the match's tree is sufficiently small that you can look through the list of individuals, you may be able to eliminate the match as being of any practical value.

 

[Bob Spiers]

Having read and hopefully digested these DNA 6-7cM postings, and made a best effort to comprehend the White Paper -fractionally more understandable than Einstein's Theory of Relativity (but aided somewhat by the Bettinger blog article) - I have reached a conclusion.

All the 6-7cM matches I wish to further peruse have been dutifully colour coded or notes added (on 3 Trees). I understand from the moment that Ancestry bring in 8cM minimum default matching, those 7.9 and under will not show. So be it. I shed no tears over this and accept the Ancestry given reasons for the change, without the need for wailing or the wearing of sack cloth.

 

[Pauline]

The fact that the Common Ancestors connection is spurious doesn't necessarily mean that the DNA match is also spurious (or vice versa, in Blaine Bettinger's opinion). But if the match's tree is sufficiently small that you can look through the list of individuals, you may be able to eliminate the match as being of any practical value.

Thank you, and yes I have looked elsewhere in the trees concerned for a possible connection and have found nothing that looks remotely promising.

 

[Pauline]

So be it. I shed no tears over this and accept the Ancestry given reasons for the change, without the need for wailing or the wearing of sack cloth.

I'm apt to agree with you here, and while I have always been cautious about accepting small segments as proof of anything, I now feel more confident in being so.

It has always bothered me that people will point to a distant DNA match as proof of a particular genealogical connection. Even where there is good documentary evidence for that connection, a single DNA match (or a small cluster of distant matches closely related to each other) cannot be used as proof, but all too often distant matches are cited as proof of connections for which there is not only no documentary evidence but which are also highly unlikely or suspect.

 

[peter]

It has always bothered me that people will point to a distant DNA match as proof of a particular genealogical connection. Even where there is good documentary evidence for that connection, a single DNA match (or a small cluster of distant matches closely related to each other) cannot be used as proof, but all too often distant matches are cited as proof of connections for which there is not only no documentary evidence but which are also highly unlikely or suspect.

It depends on the context. Last month I wrote in the newsletter about my discover of 8th cousins once removed who confirmed my research back to my 7G grandparents - and quite rightly a reader wrote in to point out that such a distant match could be spurious or that we could share a completely different ancestral line.

In that particular case the ancestors of my genetic cousins had emigrated to the US around 1700, which greatly reduced the chance of a connection on a different ancestral line; also there were three matches with distant cousins in the US, and whilst two of them shared common ancestors through to 1800, the line of the other cousin diverged around 1740.

Of course, the chance of getting a match with a specific 8th cousin once removed is remarkably small - about 1 in 200. But my US cousins bred like rabbits - it wouldn't surprise me if there are 20,000 descendants of our common ancestors in the US because they had such large families. So being matched with 3 of those cousins isn't that remarkable.

 

[Helen7]

If you want another opinion on the changes see this blog article by Blaine Bettinger, which explains things a lot better than I could.

Thanks for this blog link - an interesting and informative perspective. One part that interested me was the comment on Ancestry's statement that "Our updated matching algorithm may reduce the estimated number of segments you share with some of your DNA matches" when Blaine Bettinger says that "... the number of shared segments is usually inflated. For example, I share 44 and 49 segments with my parents when in fact it should be exactly 23." My son shares 87 segments with my husband and 79 with me, which I'd thought was 'normal', so like Blaine I'll be interested to see whether this goes down to nearer 23 after the update.

 

[canadianbeth]

I have never understood the segments part of the matches, so the fact that I share 77 (1762cM) with my sister's daughter did not raise any red flags. How many should I actually share? Except for her father, we share all the rest of our ancestors. And for my cousin's daughter, I share 14 (372cM). The next closest match is 198cM/9 but he/she is the person who has never replied to my messages so I do not know where that fit is. I look forward to that particular update.

 

[peter]

the fact that I share 77 (1762cM) with my sister's daughter did not raise any red flags. How many should I actually share?

The comments in Blaine Bettinger's blog related specifically to parent/child. There is no 'correct' number of segments that you should share with a niece.

 

[Bryman]

Word seems to be spreading and now there is almost a panic arising because of the imminent loss of a large percentage of matches. At various moments during the NZ day, I have tried checking for possible matches in the 6cM/7cM range and achieved nothing. I keep getting a message . . .

"Our backend services are overtaxed at the moment and we are unable to retrieve all your matches. We apologize for the inconvenience, please try again later."

Checking within other ranges seem to be unaffected. There may well be some truth in the wish to improve match quality but I think that the recent increase in numbers tested, now over 18 million, is leading to insufficient resources available and Ancestry is having to reduce the amount of data to be processed. I think that everyone should take note of some comments suggested by Blaine Bettinger and prepare for further reductions in the years to come.

 

[Pauline]

Last month I wrote in the newsletter about my discover of 8th cousins once removed who confirmed my research back to my 7G grandparents

I remember reading about this but it was not in my mind when I writing about the need for caution with small segments, and I was thinking mostly of examples in my own ancestry. Since I am unaware of all the evidence you have here I cannot make a judgement call on this, whereas you are better placed to do so.

Nevertheless, except with very close relatives, I hesitate to talk about autosomal DNA matching providing 'confirmation' or 'proof' of a connection. At best it is supporting evidence of what seems from other sources to be likely.

 

[peter]

There may well be some truth in the wish to improve match quality but I think that the recent increase in numbers tested, now over 18 million, is leading to insufficient resources available and Ancestry is having to reduce the amount of data to be processed

The White Paper explains why they're making the change.

 

[peter]

Since I am unaware of all the evidence you have here I cannot make a judgement call on this

All the relevant evidence should be in my previous post; I included it as a guide to others who might be faced with a similar situation.

I will be writing about this in my next newsletter as it's clearly become very topical given Ancestry's forthcoming change and Blaine Bettinger's subsequent post.

I hesitate to talk about autosomal DNA matching providing 'confirmation' or 'proof' of a connection. At best it is supporting evidence of what seems from other sources to be likely

In general DNA is a more reliable source than historical records because DNA cannot be faked or incorrectly recorded.

The uncertainties regarding DNA matches are statistical, so you can set whatever threshold you wish; however, when the DNA results support the information from the surviving records it greatly increases the likelihood that the records are correct.

Of course, it makes a big difference which came first. If the records based evidence is compiled before the DNA match is known it is more significant than if it was compiled afterwards.

 

[Pauline]

All the relevant evidence should be in my previous post

Then if this was my family I would not view these DNA matches as confirmation, only that it seems consistent with other evidence.

But this is precisely the kind of situation where we each have to make our own judgement and much will depend on the strength of the other evidence, and how much weight this DNA evidence is being expected to provide.

 

[Winksetter]

I have been busying myself saving the max number of 7cM into a group. When that is done I will go onto 6cm. Having already loaded my DNA to several other sites as security and better analysis of the chromosome level matches. I hope others are doing the same. It is a matter of time until ancestry stop that as well. It is perfectly clear that Ancestry DO have resource problems and that are certainly using this change to justify this change and for holding back on chromosome analysis tools.
We normally take 2 full memberships each year, that is under review now. Not sure whether we can commit much longer, they are restricting our own information for which we have paid richly is pushing us to the brink. I no longer understand their motivation in the DNA environment, its not so much serving their customers as it once was. I would also add that DNA affords me with vital pointers to family lines where names and records can be so unreliable, increasingly so as you go back in time. Without it I would have been totally unable to do key research into relatives from distant times and frankly into less distant times. I would point out that 7cM matches can be g g grandparents! Not very distant, not irrelevant. The mathematical DNA gurus say in effect that they are no use. I beg to differ.

 

[Bryman]

The White Paper explains why they're making the change.

The White Paper explains ONE reason why they are making the change and I understand the reasoning behind that. However, I think that Ancestry is also suffering from its own success and needs to reduce the growing load on its resources. Removing matches for 6cM and 7cM would also go a long way to achieving such a reduction in storage and processing requirements. I am just trying to warn other members that this may not be the only 'justified' reduction in availability of test results in the future. I would not be surprised to find that 9cM becomes the minimum in a few years time if the number of tests increases at the present rate.

I am sure that you must appreciate the need to minimize the cost of ever increasing hardware provision, such as when 504 timeouts occurred at LC when increasing amounts of data had to be processed. It is not something to shout about too loudly and scare the users but no business can afford to provide unlimited resources.

 

[Winksetter]

I think that people equating segment with chromosome should go back to basics. The only time they come to exactly 23 is when you compare you with yourself.

 

[Helen7]

I think that people equating segment with chromosome should go back to basics. The only time they come to exactly 23 is when you compare you with yourself.

I agree. I can’t see why the number of segments inherited by a child from a parent should be 23. Surely it should be more, when taking account of the mixing of chromosomes inherited from grandparents.

 

[Bob Spiers]

Having already loaded my DNA to several other sites as security and better analysis of the chromosome level matches.

Interesting that - about the word 'security' when considering 'other sites' offering chromosome analysis. Yesterday an Irish genealogical site of which I am a member gave the following advice..."WARNING Do not use gedmatch until further notice - it has been compromised" Follow on advice indicated fake matches with suspect emails. It seems Gedmatch were aware and were investigating. Whether or not Gedmatch resolve the issue it is not the first of such.

In November 2019 I read an article about security issues with Chromosome matching sites and (referring to Gedmatch itself) ..." over the massively crowd-sourced DNA database GED match, MIT Technology Review reports that computer science researchers designed targeted attacks that breached the GEDmatch database by making complex search strings that let them guess much of users’ DNA..."

I have not used Gedmatch for some time now and not just because of scaremongering. I was however swayed by the closing paragraphs of the article I mentioned where it highlighted the reason that DNA matching sites like Ancestry, 23andMe (et al) do not suffer from such vulnerability is because they don’t let users share their own data. Food for thought and would welcome Peter's views on the matter...or of course point me in the direction of an article written on the subject in a long forgotten Newsletter.

 

[Pauline]

The White Paper explains ONE reason why they are making the change and I understand the reasoning behind that. However, I think that Ancestry is also suffering from its own success and needs to reduce the growing load on its resources.

I think the White Paper actually acknowledges that one reason for the change is the need to reduce demand on data storage. For example, on page 14 they say:

"The cutoff of 8 cM was chosen after considering several factors. The first factor is data storage. Since the number of matching segments grows exponentially with decreasing length, we dramatically reduce the storage requirements of our matching database by increasing the cutoff."