By Endlisnis – Street limbo 3, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=45772088
A long-requested change is taking place shortly at Family Tree DNA – although we don’t have a firm date as of yet. Final testing before release is still underway.
I received a notification from Family Tree DNA, as follows:
You asked for it and we listened!
Currently, the current matching thresholds – the minimum amount of shared DNA required for two people to show as a match are:
- Minimum longest block of at least 7.69 cM for 99% of testers, 5.5 cM for the other one percent
- Minimum 20 total shared centiMorgans
Some people believed those thresholds to be too restrictive, and through the years requested changes that would loosen those restrictions.
In the near future, as soon as the quality assurance process is complete, the following changes will be implemented to the matching program.
- No minimum shared centiMorgans, but if the cM total is less than 20, at least one segment must be 9 cM or longer.
- If the longest block of shared DNA is greater than 9 cM, the match will show regardless of total shared cM or the number of matching segments.
The entire existing database is being rerun using the new matching criteria, and all new matches calculated with the new thresholds.
Most people will see only minor changes in their matches, mostly in the speculative range. They may lose some matches but gain others.
This is truly exciting news, especially for people with African American heritage whose connections to matches may be several generations back in time and may not have met the previous 20cM total threshold criteria. In fact, these new thresholds may benefit many of us with deep colonial roots, but I’m hoping that people who have had few matches until now will be pleasantly surprised.
As soon as this change is released, I’ll be checking to see if I’ve gained or lost matches, and will report back as either additional information or results are forthcoming!
______________________________________________________________
Disclosure
I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.
Thank you so much.
DNA Purchases and Free Transfers
- Family Tree DNA
- MyHeritage DNA only
- MyHeritage DNA plus Health
- MyHeritage FREE DNA file upload
- AncestryDNA
- 23andMe Ancestry
- 23andMe Ancestry Plus Health
- LivingDNA
Genealogy Services
Genealogy Research
- Legacy Tree Genealogists for genealogy research
Last year while Janine was at Jamboree we talked about the discrepancy of this minimum largest segment and she said that I was the exception. On my mother’s last page she has 4 people with the following LONGEST BLOCK. 6.10cM, 5.79cM, 5.68cM & 5.11cM for totals of 161.88cM, 221.60cM, 141.83cM & 134.14cM. Of course they count every tiny segment as well. Plus there are 16 more matches like these who fall below the 7.69cM minimum longest block.
I’m happy to see this. AncestryDNA should follow suit.
Roberta, So good that you are the one who is passing this great news on to us first (first for me anyway) and I know how vocal you have been about this and you deserve many thanks here.
This is going to make a huge difference for a number of people I can think of personally. Their research should be much more effective and honest results to be soon seen.
Also, very happy to hear that you are getting up and around again. Take care. Hire someone to pull those weeds.
Linda McKee
Looking forward to the change. I hope it will help people better identify their best matches in a way that they can understand.
Like it.
The question now is what are the SNP thresholds?
I had a interesting comparison recently. I normally use 250 SNP’s and 1.0 cM minimum segment size, but I will be criticized for using parameters that are “prone to false matches”. My belief is that if I can find two kits with zero matches using those parameters then the notion that they produce false matches will be debunked. With the amount of common ancestry that people have, it is difficult to find those kits, However I recently did a comparison between my mother and a person with both sets of grandparents from southeast Nigeria. Both kits are 23andMe V4 chipset, so we have no compatibility issues The kits had just one matching segment. It was on #1 from 171.5-173.7 Mbp (266 SNP’s and 1.3 minimum segment). I have reason to believe that the match is real. My mother has several lines which trace through a mulatto lady in Barbados circa mid- 1600s. Also my mother matches a lady (both FF kits) in this same location who has 10 percent African heritage and known slave ancestry. That person matches me on both Family Finder and GEDmatch using default parameters. She was the first connection which indicated a male relative, likely an uncle of a third great grandmother, was involved with slave breeding. When I received my mother’s DNA results, I was surprised to find her related to my father through that or another uncle. My mother is from Canada and the North Carolina connection was not expected.
My point is that parameters of just 250 SNP’s and 1.0 cM minimum segments just do not produce the false segments which many experts say they do.
I do not believe the SNP thresholds have changed from 500.
Thanks, just take it a step at a time, I guess.
How will this this impact endogamous populations, such as Ashkenazi Jews? Will this mean more “noise”? It seems as though it would.
It will be interesting to see how this will change the cousin relationships and id some of my 2nd to 3rd cousins will now show for my brother and sister and vice versa. Many were not even shown as distant for those two vs me!
I understand that there may be population groups that get few matches. As Ashkenazi (on father’s side only) I get way too many matches and the cousin relationships are distorted by endogamy. i can only imagine how many more thousands of matches this could add if you have two Jewish parents.
I would love to see someone address this for both the changes at Ancestry and the upcoming changes at FTDNA. Maybe after all is done and you have an opportunity to address how the changes impact the general population you could do a post on endogamous populations?
This is the same question I have. Please give us your thoughts, Roberta. I am 100% Ashkenazi and already have many thousands of matches. My understanding had been that FTDNA had been adjusting matches for Ashkenazim already; if so, how will the change affect someone like me, do you think?
I don’t know yet. I do know that has been part of the evaluation criteria.
http://www.latimes.com/science/sciencenow/la-sci-sn-ashkenazi-jews-dna-diseases-20140909-story.html
just saw this. it might have some info for you.
I also am concerned about the amount of noise for those of us of partial or full Ashkenazi descent. I have one grandparent who was Ashkenazi and as a result I have pages and pages of people who appear more closely related than they truly are. (I also have a Connemara Irish grandparent and a similar, though not necessrily as strong, endogamous effect happens with matches on that side as well.)
But my major concern is whether changing the standard improves the accuracy of the matching process. Can someone explain why this change would improve it or is this just to make customers happy?
I believe the new 9cM threshold was an attempt to deal with too many matches for some, and the elimination of the 20cM total was an attempt to deal with an issue for others. So they are trying to make all of their customers happy with the most accurate data they can provide. It’s hard to hit a happy medium when some people have 10,000 matches and others have less than 50 and want every possibility they can have.
Good news. Thanks!
Many thanks for this Roberta.
Roberta, am I understanding this correctly? FTDNA will still report the total cMs and the largest segment but we will lose those matches with a largest segment less than 9 cMs? My father, uncle, sister and I have a great deal of early colonial New England and have MANY matches where there is only one segment larger than 5 cMs – and it’s less than 9 cMs. I maintain an excel spreadsheet tracking our matches (currently 2284 people) and clustering our segments, and it will be a full-time job to update our data!! But, I think we will lose some of the “noise”… On the other hand, my mother’s background is entirely German (both parents emigrating to the US as adults) and she has few matches with the old system, but will likely gain new matches that were failing the total 20 cM minimum. It will be interesting!
That is my understanding.
It sounds like we all better download our matches (names and emails) and our list of matching segments sooner rather than later. LOL. (I’m still trying to follow up as to why my half-first cousin is no longer showing up as one of my matches, when he matched just 2 months ago.)
This comment is slightly off subject, it’s FYI regarding surnames posted in our FTDNA pages. Roberta, have you experienced FTDNA deleting a list of surnames? FTDNA said they showed no indication there was ever a surname list associated with my brother’s account….. therefore the only solution was to reestablish the list after they disappeared by doing copy and paste from my own surnames entries. The result of the deletion of the group of surnames was that none of his matches showed as common match surnames (highlighted).
What a nightmare it would have been if I had to re-enter all those names individually which seems to be the only way to get them posted according to FTDNA – they don’t reload as a new list if you replace a family tree. There are other things being affected also which are connected to the same issue… like the surname lists which appear in the FTDNA Origins section and within his Family Tree – I have never been able to open the surname list from within the details box of a person even though there is a link. Just saying…. it looks like FTDNA is having issues I had not previously seen – better technicians are needed perhaps?
I don’t know, but I have reported these two items as a bug if they were not already before.
Is it that all the new matches will have at least 9 cM Shared cM, which is the lowest threshold of Longest block? Is it also that all the currently existing matches below 9 cM LB will be lost?
That is my understanding from what they provided. I have submitted a request to confirm this.
Thanks. I hope we get to know before the change happens.
A brief story I would like to share. I have only had my DNA done at 23 and me. I received a message a year ago from a young woman when she showed up as possibly being related. We corresponded periodically after that, trying to find the connection. She was trying to find her birth family. She walked me through uploading my info to GEDmatch. We had determined the match was probably on my maternal side. I sent her what info I had, etc. Then, as an afterthought I sent her info I had from a 1982 family reunion, complete with names, adressess, and phone numbers current at that time (no relationship info was given). That info verified other info she had found on her own. Last I heard she was meeting with a possible sibling! You never know when an oddball document in and old file will be the “KEY”. !!!
Roberta, So glad you are feeling better!
That is some exciting news. To take it off me personally, I think it will be some relief to Gedmatch. Now all those thousands of people will not have to automatically upload their info to Gedmatch to see matches below that threshold. As of now I have less than 80 matches on Family Tree DNA and a quick glance shows only 1 match has a longest block cM less than 9. That longest block is 7 something cM’s, but the two pencil points touch on another chromosome indicating mostly definitely an ancestor in common, but that I can only see if I use the drop down menu box. I think that match will probably be removed in the new process, so I have written down all the contact info and will email her tonight. As an African American who has about 26% European ethnicity according to FTDNA, I would not expect a lot of matches with the current default minimum. That minimum would leave you in the about 4th cousin range, and I pretty much knew all of my first, second, third and even fourth cousins without DNA testing. I like the new changes, but I still think the best solution would have been to just have a drop down tab for cM’s, with maybe a pop up window that you have to click to go forward that states matches below this point are highly speculative. The 9 cM minimum is good, but I like many people I have seen triangulations and confirmed matches well below that threshold.
I wish they’d instead spend their time (and my money) on more worthwhile projects. What use are more matches on a site with next-to-nothing for tools, minimal trees, what what seems like a higher percentage of members who do not reply to inquires. Oh well, at least i have Ancestry. I’ve learned more real world information and made more connections in 6 months than I have in 2 years of FTDNA. FTDNA is good for the data side of your DNA and that’s it.
I still believe that the thresholds are too restrictive, particularly for us African Americans. The bulk of my African American matches at Ancestry fall in the 6-7cM range- the good news there is, I can actually triangulate those segments on Gedmatch. If we have 16th-17thc slave ancestry, then it is doubtful that we will carry a 9cM segment. I honestly do not see how this helps.
There are so many errors on Ancestry. Do not trust any thing they have transcribed or any other trees. There is too much false information.
I guess this is where I’d say ‘Prove it’. In less than 6 months on Ancestry DNA I found my true paternal gr-grandfather. Something 2 years of FTDNA was unable to do. And with the known relatives who are on there, I’ve yet to see any errors in the DNA levels or connections/shared matches. In fact, they fall exactly where they should considering some of them are only ‘half’ relatives due to the situation with my gr-grandfather.
Only a fool ‘trusts’ anybody else’s tree. Use them as reference? Sure. ‘Trust’ them? Never, unless you know the researcher and their methods.
Haven’t FTDNA customers also requested incorporating X in matching algorithm?
Why hasn’t FTDNA included X in the matching algorithm?
When will they?
The X is included if you also have an autosomal match.
Roberta,
I love you, but derp.
You know what I was asking.
Just say “Yes, customets have asked for that change. I don’t know why FTDNA has been short-sighted and does not use the X to find matches. I do not know if FTDNA will ever make that change, but i will inquire.”
derp??
I am a little confused also. I have 51 pages (so between 500-510 matches) of 7.69 to 8.99. This is over 1/3 of my total database.
What happens to them? If they are deleted, I would not consider that “minor changes”
There is a significant amount of confusion, but it does not appear that you will lose those matches. I hope that Family Tree DNA will issue additional information when they make the change.
Pingback: Family Finder Update Update | DNAeXplained – Genetic Genealogy
Pingback: Family Finder Update Comparative Results | DNAeXplained – Genetic Genealogy
Pingback: Family Tree DNA Introduces Phased Family Finder Matches | DNAeXplained – Genetic Genealogy
Hello Roberta, Any update on this? I still seem to be on the old system, with the 20cM shared cut. The new criterion sounds almost exactly what I would choose.
Hi Paul, This was implemented when I wrote about it. Everyone is utilizing the new thresholds.
Hi Roberta 😊 I’m not sure if the Family Finder thresholds have changed since 2016. I have a match on Gedmatch at 20.5cM total, largest segment 13.6cM (this is on the one-to-many tool). We both tested at FTDNA but don’t show as matches – which we should, under either of the two threshold criteria. He’s as puzzled about this as I am. Can you think of any reason why we wouldn’t appear on each other’s FTDNA matches list? Many thanks, Fern
PS Gedmatch’s one-to-one tool verifies we match on the 13.6 segment (the only segment we share >7cM)
Not a clue. I know the different companies use different “healing” routines to over come bad reads or missing segments. If you’re using genesis, it’s known to be buggy.
No, not Genesis. Thanks … I’ll try the FTDNA forums 🙂
Just received clarification from the FTDNA helpdesk. So I thought I’d share it:
“GEDMatch examines different segments from us, and has different thresholds for matching. By way of example to illustrate this, if they look at a chromosome and look at the segment comprised of markers 100-200, and we look at the same chromosome and look at the segment comprised of markers 1-155, and you only match on markers 154-200, then you would look like a non-match to us, and a pretty solid match at GEDMatch …… sometimes it just so happens that the markers looked at by other companies tell a different story than the ones that we look at.” 🙂
Pingback: DNA: In Search Of…Your Grandparents | DNAeXplained – Genetic Genealogy