RootsChat.Com

General => Ancestral Family Tree DNA Testing => Topic started by: sugarfizzle on Thursday 08 March 18 05:49 GMT (UK)

Title: Don't ignore 5th to 8th cousins at Ancestry
Post by: sugarfizzle on Thursday 08 March 18 05:49 GMT (UK)

My maiden name is Steer, and apart from my first cousin, I have found no other confirmed Steer matches at ancestryDNA. I frequently search matches for Steer in case something turns up. I looked yesterday - same old results, my cousin plus 2 other matches with Steer in their tree, but no obvious connection.

Today I was browsing through my newest matches, 5th to 8th cousins. Skipping through those with no tree and those with private trees, I looked at a few who had largish trees and who had logged in recently.

Found two new Steer matches on first page, one still dubious, probably not a match with this name, might or might not match somewhere else.

The other one, bingo.
A Steer family member, sharing only 6.9 cM, would be ignored at gedmatch. Obviously need to confirm, but if confirmed this would be a match with my 6G grandparents, making us 7th cousins.

To me this is fantastic news, and goes to show the reason not to ignore 5th to 8th matches. You can't go through everyone of the thousands that keep coming, but regular searches of name and place, or a quick look at new matches, does often prove fruitful.

One very happy Steer descendant.

Regards Margaret
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: haliared on Friday 09 March 18 02:00 GMT (UK)
I have found some 5th to 8th cousins with a moderate rating to be mine and others I admin for.
I have found ancestry very accurate and am pleased with the results.
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: GailB on Tuesday 13 March 18 02:09 GMT (UK)
I have a third cousin showing as 5th-8th cousin. I think it may because of cousin marriages on both mine and his side of the tree.
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: melba_schmelba on Saturday 15 September 18 23:28 BST (UK)
I have a third cousin showing as 5th-8th cousin. I think it may because of cousin marriages on both mine and his side of the tree.
Once you get past first cousins, the distances can be very inaccurate because of the random distribution over generations. You may have a very small amount of DNA from one great grandparent, as little as 2% - which ancestry might show a cousin in common with as a 5th cousin, or as much as 20% which ancestry may show as a first or a half sibling. So you should only take these relationship distances as a very vague guide.
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: Palladium on Sunday 16 September 18 00:36 BST (UK)
I think ancestrydna tends to be a little bit optimistic sometimes....
Pall
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: Mart 'n' Al on Sunday 16 September 18 00:49 BST (UK)
Margaret, I know your original comment is from 6 months ago, and I've been following your postings with interest ever since you constructively commented on one of my comments, but saying that "... sharing only 6.9 cM, would be ignored at gedmatch... " needs correction, or at least updating. I will only UPDATE your comment by saying that, as you know, parameters can be set to specify minimum matches of interest.

I know we all want to suddenly discover unknown close relatives, (which seems odd but I admit to it myself), but life is probably too short to look for 8th cousins.  I enjoy all your comments, which I follow closely and find very informative.

Martin
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: sugarfizzle on Sunday 16 September 18 07:03 BST (UK)
Martin, Thanks for your reply. I agree it does need clarifying if anybody reads it in the future.

This is from ISOGG, page modified Oct 2017, I think it's still correct.

https://isogg.org/wiki/Autosomal_DNA_match_thresholds

23andme - For half-identical regions the autosomal threshold is 7 cMs and at least 700 SNPs for the first segment; 5 cMs and 700 SNPs for additional segments and for people you are sharing with. For fully identical regions the threshold is 5 cMs and 500 SNPs.

But limited to 2000 matches

ftDNA - A match is declared if two people share a segment of 9 cM or more, regardless of the number of total shared cM. However, if thereís not a block thatís 9 cM or greater, the minimum of 20 shared cM with a longest block of 7.69 cM applies
(Autosomal DNA uploads from other sites limited to close cousins, more distant cousins only show up for those who test there).

Ancestry - The minimum threshold for a match is set at 6 cMs of IBD sharing across the genome.

Gedmatch - Default set at 7 cM
Again, 'One to many' limited to first 2000 matches.

And from myheritageDNA site, updated last year -
Previously the minimum of shared DNA for a match was 12 cM and now the minimum is 8 cM.

So only at ancestry would this match show up automatically. I have tried to persuade her to upload to Gedmatch, unsuccessfully, so I have to leave it at that.

A group of Steer descendants from RootsWeb Steer mailing list, have uploaded to Gedmatch, and by lowering the threshold to 1 cM the Surrey Steer group all match, but only by about maximum 4 cM. This is all on the same place on one chromosome, so it is a pity that my ancestry match can't upload there as well, to see how much and where she matches with the others.

We have confirmed the connection between me and her though, so I am happy.

Thanks for correcting me, Martin  :)

Regards Margaret

Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: sugarfizzle on Sunday 16 September 18 08:04 BST (UK)
Martin, You say 'I know we all want to suddenly discover unknown close relatives, (which seems odd but I admit to it myself), but life is probably too short to look for 8th cousins.'

I have to disagree with that, especially where we are talking about lines that we are unsure of, or with brick walls.

Another example from my DNA
One name, Laversuch, traced back to marriage in Wiltshire, England in 1745, no obvious parentage, family brought up in Idmiston.

By the time census records are readily available, the family is still in Idmiston, a family of the same name a few miles away. Tracing this second family back there is no obvious connection.

Wiltshire PRs then came online, showing a few 'possibilities' where the name is considerably different, but could be 'manipulated' to be the right name. Tentative details and family entered as an 'island', i.e. not attached to my original man.

Then comes DNA - several matches with 3rd, 4th and 5th cousins, to be expected.
'Surname search' at ancestry with variation of Leversuch, a known variation, I found a match with a family I had never heard of.
Looking at myheritage DNA, another match descended from the same lady, but unknown to my ancestry match.

Back to Wiltshire PRs
Stephen Leversidge (my ancestor) and Mary Leversidge (their ancestor) baptised in Chitterne, Wiltshire at appropriate times for their marriages.

The family name that could be manipulated? Leverstretch or Leversidge.

Do my current confirmed 3rd to 5th cousins also match with my ancestry match? Those that have replied to me do, but none have uploaded anywhere else to compare with chromosome browser, unfortunately.

Do the family found 150 odd years later in census also relate to this family from Chitterne? Yes, in hindsight quite easily.

The conclusion? I have almost definitely taken my Laversuch Leversuch Leversidge Leverstretch line back by 2 further generations, by finding 7th cousin matches at myheritage and ancestry. Further matches from different descendants and matching with a chromosome browser could verify it further.

The same result could have happened with 8th cousins, so I am still happy to say 'Dont ignore 5th to 8th cousins at Ancestry'.

Regards Margaret



Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: melba_schmelba on Sunday 16 September 18 09:57 BST (UK)
Margaret, I know your original comment is from 6 months ago, and I've been following your postings with interest ever since you constructively commented on one of my comments, but saying that "... sharing only 6.9 cM, would be ignored at gedmatch... " needs correction, or at least updating. I will only UPDATE your comment by saying that, as you know, parameters can be set to specify minimum matches of interest.

I know we all want to suddenly discover unknown close relatives, (which seems odd but I admit to it myself), but life is probably too short to look for 8th cousins.  I enjoy all your comments, which I follow closely and find very informative.

Martin
Sorry if I am explaining something you already understand Martin, I think many people think the centimorgan match is accurate or very near accurate reflection of the genetic distance, when the reality is it can be very misleading indeed. It is based on the belief that we share 25% for each grandparent, or something VERY close, and then so on further back. The reality is after our parents it CAN start to vary significantly, so we may have 20% of one grandparent's DNA and 30% of another. If that outlier result happened in the previous generations - and the 20% grandparent had a skewed share of their own grandparent's DNA, you could end up with no DNA from a great great grandparent at all.

I highly recommend looking at this video by Andy Lee from Family History Fanatics who shows using a computer model and charts running in real time how the share of our ancestors DNA can vary quite hugely more than people realise:

https://www.youtube.com/watch?v=nlmK0X3I1Lo (https://www.youtube.com/watch?v=nlmK0X3I1Lo)
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: sugarfizzle on Sunday 16 September 18 10:22 BST (UK)
melba_schmelba, Martin and I realise that, and Martin has quoted that video, or something similar, recently.

DNA is not an exact science, as most of us realise.  But it helps. My main tree is private, and I have attached the Leversedge Leverstretch Liversedge etc family to Stephen. In my public tree to which my DNA is attached, the family is still an island, so won't show up in searches by DNA matches and will remain an island until if and when I have further proof.

I have no wish to circulate an erroneous public tree.

Shared cM project gives a good idea of how much or how little some people share with relatively close and more distant matches.

https://thegeneticgenealogist.com/2017/08/26/august-2017-update-to-the-shared-cm-project/

All 3rd cousins, from shared 2G grandparents, have reported a match.

Regards Margaret
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: Mart 'n' Al on Sunday 16 September 18 11:13 BST (UK)
Melba, I think I have previously recommended that video in a much earlier subject. I agree.

Martin
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: melba_schmelba on Sunday 16 September 18 13:42 BST (UK)
Melba, I think I have previously recommended that video in a much earlier subject. I agree.

Martin

Hi Martin! Yes of course I saw your comment on the video after I posted my message ;D ::). I just wanted to try and get across to others, that they certainly shouldn't ignore even what ancestry says are 8th cousins, as we can COMPLETELY lose DNA by the great great grandparent stage as Andy's video shows. Even those great great grandparents showing 2% inheritance would probably be reported by ancestry for a descendant in common as an 8th or even greater cousin when they are in fact 3rd cousins.
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: hurworth on Tuesday 18 September 18 10:52 BST (UK)
Hopefully your matches will upload somewhere with a chromosome browser.  I tend to a bit iffy if it's a small shared segment.  It's possible that you infact match via another Wiltshire ancestor but if and when you get more matches and they triangulate you can be more certain of the link.

Having said that, I look at 5th to 8th cousin matches all the time.  I have access to two kits (3rd cousins to each other) and so the first thing I do when there are new matches is check whether they're a match to both.  Having access to both kits has helped take the trees back further on the Scottish side from the mutual gtgt-grandmother, and we've found that she had a gt-aunt (from a combination of DNA and records).  It was a site with a chromosome browser that clinched it though.

The third cousin appears to have inherited very little DNA from the English gtgt-grandfather though, so not so useful for matches on that side.  It's just how the cookie crumbles.
Title: Re: Don't ignore 5th to 8th cousins at Ancestry
Post by: sugarfizzle on Wednesday 19 September 18 08:03 BST (UK)
hurworth, I agree it isn't confirmed, and I am hoping that further matches will indeed prove it.

However, this is my only Wiltshire family, so it is highly likely to be from this line. I will try again to persuade at least one or two of my 3rd, 4th, 5th cousin confirmed matches to upload elsewhere, if not the more 'iffy' 7th cousin match, who does in fact share a respectable 10.1 cM with me. The Myheritage match shares 12.8 cM.

Not holding out much hope, though.

Regards Margaret