Wow, today is a great day in genetic genealogy-land. After the conference in Houston, which ended just a month ago today, a small group met with the Family Tree DNA team and explained what we, as users, need, and why. We walked through lots of scenarios and everyone did a lot of explaining. The whiteboard was full. We were hopeful.
Bennett made a commitment, publicly, at the conference, to do whatever it took. However, I never expected this feature, the Family Finder Match Matrix, which was very high on the priority list, to make it out the door this soon. Less than one month later. Hats off to the Family Tree DNA team! YOU ROCK!!!
Why is this so important? Because you have two halves to your chromosome, and there is no magic zipper to divide Mom’s half from Dad’s half. So you’re going to match with people who come from Mom’s side, Dad’s side, and some who just happen to match because of random recombination. The best way to figure this out is to see which of your matches match each other as well.
So, in a nutshell, here’s how this works.
- If your matches match you, but not your other matches as revealed in the “In Common With” feature, they are questionable matches. To find who you match in common with one of your matches, use this crossover icon on your matches page:
- If your matches match you and each other, then they are very likely important genealogical matches.
- If your matches match you and each other, and you can identify the lineage based on which of your cousins or other family members they match, you’ve got a hugely valuable piece of information. I discussed this in yesterday’s article, Chromosome Mapping aka Ancestor Mapping.
Here’s the release today from Family Tree DNA. And even better news, they have promised to keep us apprised on new features to come ON A WEEKLY BASIS!!!
From Family Tree DNA:
Today, we are happy to release our new BETA Family Finder – Matrix page. The Matrix tool can tell you if two or more of your matches match each other. This is most useful when you discover matches with wholly or partly overlapping DNA segments on the Family Finder – Chromosome Browser page.
Due to privacy concerns, the suggested relationship of your two matches (if related) is not revealed. However, we can tell you whether they are related according to our Family Finder program. To use it, you select up to 10 names from the Match list on the left side of the page and add them to the Selected Matches list on the right side of the page. A grid will populate below the lists. It will indicate whether there is a match (a blue check mark) or there is not a match (an empty white tile).
You access the BETA Family Finder – Matrix page through the Family Finder menu in your myFTDNA account.
The page starts out with two list areas: Matches and Selected Matches. You add Matches to the Selected Matches list by clicking on a name and then on the Add button.
Here is a screenshot of the BETA Family Finder – Matches page with a few matches added to the Selected Matches list.
You can change the order of names in the matrix by clicking on a name and then either the Move Up or the Move Down button.
To remove someone from the Selected Matches list, click on their name and then the Remove button.
______________________________________________________________
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
But I still can’t tell if they are matching one another on the same segment of chr 10 they match me! Close, but no cigar….I want to work with proven knowns, not ‘maybes’. Just me and my type A personality. I can’t say for certain that we are all sharing a common ancestor unless I can see that we are indeed all sharing on the same segment area of the same chromosome. Once I know that then I know that some common ancestor passed that segment down to us all. I’ve seen instances where two folks who match me in an area, match each other, but somewhere totally different–that could be relevant or it could be totally irrelevant. Step in the right direction though, so hopefully we will see it progress to give us what we truly need-what we already see at 23andMe.
Yes, you’re right about that. Family Tree DNA has some concern about privacy in terms of revealing that information about your matches to each other. They are trying to work through it. This eliminates a lot of extraneous matches. I’ve also seen cases where they match each other, but don’t match me on the same segment, which in itself is still a hint that there is a family connection, just not proof.
I agree. I like this, but it falls short.
I can understand privacy concerns, but I already know where my matches match me. Knowing where they match each other doesn’t blow down some kind of privacy barrier that isn’t already pretty well blown down. I have to think some kind of direct hacker attack on FTDNA is a much more plausible threat than the bits and pieces of information that genealogy nerds could conceivably string together.
What a great tool. Thank you Mr. Greenspan for this and thank you to everyone who goes to the conferences to help them understand what tools are useful to everyone who uses DNA to help with their genealogy. FTDNA is certainly a cut above the rest. Great Job everyone at FTDNA.
I’ve been happily playing with this new feature. Thumbs up on this! Question though…I chose a match, found the in-common matches with him, and then added them all to the Matrix. I expected all the in-common matches to have a check mark in the column with the original match (the one I generated the in-common match list with) but they don’t. What is that telling me? Wouldn’t they have to match him to be in-common matches?
“Meh” at best.
First of all, I have over 2700 matches in some cases (Ashkenazi). Scrolling through a list to add names is painful. I assume they’ll find a way to make this better, perhaps using an interface like that of the Chromosome Browser.
Second, I’m Ashkenazi, and a nice-size majority of my matches have SOME relationship to each other. Without knowing the level of relationship, or at the very least filtering out “Suggested: – ” matches, it’s nearly useless.
Send them feedback. How would you suggest they address the privacy concerns about sharing information about how 2 people match each other, but not you??
– An opt-out option for people to not share their relationships. This would be ideal for me. Actually, ideal for me would be just to show the relationship, but it seems unlikely to happen. 🙂
– An option for me (the user of the matrix) to not show suggested distant cousins. That does reveal to me that the two people are at least [suggested] 4th cousins, but it may be sufficiently ambiguous to meet the privacy bar.
We talked about both of those. The opt-out might include people inadvertently who don’t want to be included and many people would never notice the opt-in. I don’t know what the final resolution will be.
What exactly are the privacy concerns?
Unless I know exactly where the start points and stop points are, and unless I know which alleles they have, and unless I can assume there is a perfect match at every single nucleotide within the segment, and unless I have some detailed health information on the two people, and unless that health information happens to coincide with the matching segments in question, what could one do to violate privacy with second hand matching segment clues?
The problem is that the release everyone signs says that they will show you matches to you. If they show you locations where other people match each other, but not you, that technically falls out side of what people signed when they bought the kit.
Roberta, I don’t think that’s rightklik’s point, or at least it’s not mine. If I match A and B, I don’t necessarily need to know where A and B’s matching segments are relative to each other. I get where that might violate the FTDNA privacy agreement, and potential legal issues unless a separate amendment is made.
But what I would like to know would be the level of match. Are A and B suggested 1st cousins, or 5th?
I have already sent this suggestion to FTDNA for consideration, but I bring it up here to understand the privacy concerns a little better.
I am an administrator of a Family Finder project and I am able to view each members matches. So I can see how 2 people in my project match each other but not me.
I suppose it would be possible to start the “Fearless FF Project” where everyone who joins becomes an administrator as well.
Anyone want to join?
Yes! I would love to! Did this ever get created?
“If they show you locations where other people match each other, but not you, that technically falls out side of what people signed”
I can understand that point. But judging by that standard, FTDNA is already on thin ice, at best. Anyone who puts their matching segment information from FTDNA into a spreadsheet is going to see a lot of people matching each other. Put that together with ICW and this new matrix and there’s not much left to the imagination.
It seems like the remaining barrier isn’t doing much to protect sensitive information, bit it is blocking genealogy research.
“If they show you locations where other people match each other, but not you, that technically falls out side of what people signed”
I can understand that point. But judging by that standard, FTDNA is already on thin ice, at best. Anyone who puts their matching segment information from FTDNA into a spreadsheet is going to see a lot of people matching each other. Put that together with ICW and this new matrix and there’s not much left to the imagination.
It seems like the remaining barrier isn’t doing much to protect sensitive information, bit it is blocking genealogy research.
Spot on Rightklik, this makes a joke of the whole privacy issue
They aren’t showing you the genetic locations where those people match each other but not you. You can find out where each of them match you by checking the chromosome browser.
I am also of Ashkenazi ancestry – I limited mine to 2-4 cousin matches and found they all matched at least three of my other matches except one – he matches a relative of the same name and me. I only have about 8 2-4 matches. No immediate. So I actually found it very helpful. I don’t even consider the huge list – after a certain distance the likelihood of finding a paper trail seems unlikely. I will look at 3-5 after I give up on the 2-4th responding.
Steve, I’m not sure if someone has already mentioned this, but you should be able to to start typing in the last name of the person to add them, there is no need to scroll thru the long list of names. Just make sure your cursor is in the “Matches” box. (I’m using Firefox, not sure if it works with other browsers.)
Thank you, Rhonda, that works in IE too. There’s still some funky and annoying problems (e.g., when you’re in the box and press backspace, it sends the PAGE back one in history), but it’s a start and a help.
There seems to be a matrix bug.
My Mom and her cousin share matches that are not registering on the matrix.
I would love it if FTDNA would add a “remove all” button to clear
everything in the “Selected Matches” box.
This would make things much faster.
Please submit feedback to FTDNA.
Joe, you can use the shift key to select more than one name or to select all names between two names. Then you can click on “Add” or “Remove” to add or remove the whole selected group at once.
what am I missing here – how is this different from the In Common With feature, other than being a graphic representation of whether some matches match others?
If your matches match you, and each other, it’s an indication of a family connection between all of you. If they match you, and NOT each other, it does NOT indicate a family connection between the three of you.
yes, I understand that. But the In Common With feature tells you if your Match A matches your Match B (or not), doesn’t it? Isn’t that all the Matrix is doing?
No, that’s not what the ICW it telling you. You are Barton. Your match is Susie. You ICW with Susie and you both have Martin and Abraham. What ICW tells you is that Martin and Abraham both match both you and Susie, someplace, not necessarily on the same segments. It does NOT mean that Martin and Abraham match each other. This new tools tells you whether Martin and Abraham match each other, in addition. That’s important because if all 4 of you match each other, it implies a family group. If Martin and Abraham don’t match each other, then you are looking at different ancestors.
I don’t see that there’s any new information either, just a slightly easier way to get some of the information we can already get. If I use ICW on a match A and get B and C, I can already tell if B is related to C by using ICW on either B or C. It just saves time.
Furthermore, by looking at chromosomes, I can better tell about a possible common descent for me, A, B, and C than I can with the matrix.
It looks like a plus for convenience.
Privacy, really is the number one issue here, i personally cannot conceive of why anyone would go to the trouble of DNA testing for the purpose of Genealogy and then have privacy constraints, it’s kind of like you show me yours but there is no way you are seeing mine, it’s very annoying as is the limit on the size of the tree you can enter, as for this new feature, my known 2nd cousin and i do not match hmmmmm curious
There are lots of reasons someone might do DNA testing but still have privacy constraints.
For example, they might be interested in their heritage, but NOT interested in reconnecting with the great-uncle who molested them when they were four. So they might want to determine which line you’re from before connecting with you.
Or, they might know that your (still-living) grandmother cheated on your (still-living) grandfather with their (still-living) grandfather. In cases like that, they might want to make actual human contact, rather than letting folks find out by looking at a tree on the internet.
Or, they might be seeking their father, who raped their mother. There again, they might want to make some actual human contact rather than just putting that out there.
Or, they might have a mother or grandmother who worked as a prostitute. That’s another one where contact might come before revelations.
Or, they might have living relatives who violently object to this whole thing (maybe because of some of the circumstances above, or maybe for a million other reasons). However misguided those relatives may be, it’s not reasonable to expect them to tell their close living relatives to suck it up for the convenience of distant DNA cousins. Those people still have to navigate family gatherings, relationships between their children and those relatives, etc.
These are just a few examples from people I personally know. There are lots and lots of other circumstances. This isn’t smooth sailing for all of us, and genealogy is helpful (even healing) for many people with very difficult backgrounds. They shouldn’t be excluded just because they can’t give strangers everything they want up front.
Actually, not that useful for my work.
I’ve taken the detailed chromosome map and have mapped everybody to a segment. While I know that simply overlap does not guarantee that you are related, it’s obviously a requirement if using the chromosome mapping approach.
If we take the ICW list and compare, chromosome by chromosome, segment by segment, it’s relatively easy, but quite time consuming, to figure out who overlaps whom. In the example I’m looking at, I have a person that I ask for an ICW list and then have identified several people who also overlap my ICW match and myself, as expected (and several who do not, which is a different issue with FTDNA).
When I place the people who I overlap, as well as my ICW person, into the matrix, NONE, not ONE of the people are shown to overlap my ICW match or anybody else in the list.
Disappointing.
I was hopeful that the matrix would take it one step further and simplify some of the work for me but in the end it seems to have filtered more out, e.g., the filter length of largest segment, or the total amount of combined DNA in common is below some arbitrary threshold.
===============
In another example, I *KNOW* the relationship I have with 3 other individuals. I also know that they share on a specific chromosome with each other, as well as with me. While Person A clearly matches Person B (they are mother/daughter), and although A and B are overlapping C and show “C” in their matches list, the matrix does not show them as being related.
Again, it appears to be a threshold issue, and some bugs will need to be worked out before it’s useful.
My 2 cents.
Regards,
Paul
I’m not at all sure you’re not seeing a bug of some type with none of your matches matching each other. I have forwarded this to ftdna but you may well need to contact them with your kit number.
These look like some very helpful tools.
I wish ancestrydna would add some tools.
Don’t hold your breath. Ancestry is for family trees, DNA is just not their specialty.
I suspect they’re holding their cards close because they want to TELL YOU who your common ancestors are rather than allowing you to see it for yourself. Standing between you and your matching segment data could prove to be more lucritive in the long run: http://www.ashg.org/2013meeting/abstracts/fulltext/f130120903.htm
I sent feedback already, but it wouldn’t hurt if everyone did too.
If you like these suggestion please let FTDNA know:
1. Add “Remove All” button to clear “selected matches” box
2. “Set threshold: (cM) drop down menu to filter
matrix matches by segment size (10, 5, 3 & 1 cM and up)
3. Click on Matrix Match blue box opens chromosome browser with all three matches displayed.
4. Add “Download all Matrix Matches” button (excel or csv file) showing only triangulated segments.
Thanks
Bug Fixed!
Confirmed. Bug(s) fixed. Both of my examples above indicate what they should indicate by doing a specific chromosome/segment analysis from the ICW matches. This will be a significant time saver.
Pingback: Family Tree DNA Weekly Update 10 December 2013 | Ups and Downs of Family History
The Matrix page gives me no information I do not already have.
As many have said, it’s only really useful when you can see what the match is.
I don’t buy the argument there is some insurmountable obstacle to giving the information. 23&me provides the information, through the people who share plus the people who select Public Share.
On 23&me, I send sharing invitations to every match who gives their name. I assume the anonymous ones, the vast majority, aren’t interested. I have 274 who have shared with me, and 147 pending invitations (= those who haven’t shared.)
For comparison with ftDNA, I think this is the meaningful response rate. ftDNA filters out the anonymous people before it sends you the matches.
For comparison, I have 392 matches on ftDNA. I arbitrarily sampled pages 3,19, and 27. Out of 30 samples, 22 provided some genealogical information. 8 provided gedcoms.
I have sent my feedback to ftDNA.
It is unfortunate that size of database of Ancestry, ftDNA, and 23andme is inverse to the usefulness of their tools.
Since ancestry has allowed downloads, I spend most time on Ancestry matches who have submitted their data to Gedmatch. For American Colonial lines, that seems the best of both worlds.
Roberta,
Thank you for posting this. I just tried it out. I have previously built a spreadsheet which has one worksheet of my Autosomal matches (including email addresses, when I contacted them, etc.) and one worksheet of our chromosome overlaps from the 1+cm. Using the Matrix tool, I selected five individuals with whom I match on Chromosome 8. They also match each other. This information I already knew from my spreadsheet. One individual does not show up as a match using the Matrix tool.
Either I don’t understand what a match is, they define matches different than I do in my spreadsheet, or perhaps they have a bug. In this particular situation, I can see what the bug might be.
The following individuals match me at a significant enough level that I think we are all somehow related. I am attempting to find out what line(s) of my family Chromosome 8 represents. I have changed names to initials. From my spreadsheet:
MATCH CH START END
DGC 8 108529309 124725456
DBM 8 118982372 124725456
MKM 8 118982372 124725456
SM 8 118982372 124725456
KAS 8 80116676 127956676
Using the Matrix tool, DGC does not show as a match to the other four. The other four show as matches to each other.
Unless this is a bug, then what this is telling you is that everyone except DGC shares a common family line and DCG is related to you on your other side of the family, not through this line. If you have genealogy or something else that suggests this is error, meaning a bug, then let FTDNA know, for sure. But this is exactly the situation this tool was designed to expose. And singe DCG does not match any of the others, but the others do match each other, it suggests that this is not a bug, but the real McCoy.
REALLY? Wow! In that case, I’ll read your post again, in more detail. I have not yet received reply from DGC, and of the five, only KAS and I are the closest to knowing how we match. That is on my mother’s side. So you are saying that the others would also be related to me on that same line on my mother’s side, but not DGC? Even though they share Chromosome matches? I have much to learn.
Yes, the ones who all match each other are from the same line. The one that does not is matching you from the other side of the family. Remember you have two halves to every chromosome = Moms and Dads halves – so you will ultimately have different people match you on every segment who are not related to each other because one group comes from Mom’s side and one from Dad’s. There may also be some random matches too that you can’t track to either but they also won’t cluster.
WOW! Okay, now I’m super excited. 🙂
I’m not getting an option to hit “Reply” again to my earlier post, so I’m just seconding what Dave said. The Matrix feature is just a more economical way to do what ICW does; instead of individually confirming that each match matches other of your matches (or not), it tells you that in fewer keystrokes and in a nice graphic. I’m not poo-pooing it, but I think it’s important to be clear that it is not giving us new or different information.
I was quite excited – I entered all my 2-4 cousin matches (which are my closest) and they all matched at least 3 of my other matches – except one who only matches another person to whom he is obviously related.
With this information we may be able to find who each group has in common. If they respond to our email.
Wow! I’ve had a lot of fun with the matrix tool. When combined with a spreadsheet of matches sorted by chromosome number and location, it makes a powerful tool.
It is a good visual. In my opinion it would benefit from some further information beside it together with a disclaimer to check segments at smaller segments in the chromosome browser and to encourage further testing (if you believe it to be in doubt – otherwise it could create havoc for newbies). I have 8 FTDNA family profiles and I have 1 cousin whose matching segments to 2nd cousins is lower than normal but then on the flip side he has had wonderful matches to 3rd through to 6th cousins with common ancestors to all of us. If I were to rely on the matrix I would think he wasn’t related to certain family members when I know from other cross-referencing checks he is.
Roberta, I just did In Common With for one of my matches. He came up with 8 ICW matches. When I put him and those eight in the Matrix, not one of those 8 matches each other. What does this mean?
I t might be a bug. Enter person’s name you match (let’s say John) and compared against and then the rest and see if they match John. If not, it’s perhaps a bug.
So what I think you said to do was an ICW on the 8 individuals that John matched to see if John shows up as a match for each of them. I did and he does. So if John matches me and those 8 other people, but those 8 other people don’t match each other, and if it’s not a bug, then would it mean that the matches are so small that they don’t register in the Matrix tool? Or does it mean that John is one one side of my family, and that the other 8 are actually on another side of my family? I will do some comparing in my chromosome spreadsheet.
What it means is that they aren’t clustering as a family, so they could just be very far apart on whatever line. They could also be IBS or circumstantial matches. Or yes, very small segments. Using the chromosome browser is the next step to take a look to see what is going on. It could be a bug but it doesn’t sound like it. I’m working with raw data here and I’m finding a lot of that kind of thing, more that don’t match than do match.
I took a look at my chromosome matches. Even just on Chromosome 6, three of the matches for “John” match each other, meaning that their start and end positions overlap (right?). Wouldn’t it be the case then that these three should show up as matching each other in the ICW (they do not) and in they would overlap in the Chromosome Browser (they do)? The number of centimorgans matched with me for the three on the common section in Chromosome 6 is quite low: 1.30, 1.21, 1.56. “John” does not at all match me on Chromosome 6.
I randomly picked “John.” He just happened to be the first one on a list I had. Our total shared cM is 36.21, with no individual segment larger than 8.02. If very low cM matches are ignored in the ICW and the Matrix tool, then neither would be of value for detailed analysis (I wouldn’t think, anyway).
As a test, I chose another match, “Jane,” who I know is my 3rd cousin, 2x removed. She and I know our common ancestors and how we descend from them. I did an ICW on “Jane,” and three people showed up. None of those three match each other. I looked at the Chromosome Browser and my spreadsheet. None of them overlap “Jane.” I think I’m losing my mind.
Should I trust the spreadsheet more than the ICW and the Matrix?
Remember that matches are based on the 7.7 cM threshold but your spreadsheet downloads as low as 1cM. That may be the difference in what you’re seeing. Since I can’t see your data, I’m forwarding this to the QA person at FTDNA for them to look at.
Thanks. It would seem that if that were the case (that ICW and Matrix tool both use 7.7cM), with as low as these matches are (as shown in my spreadsheet which came from the 1+cM Chromosome Browser level), then these three people wouldn’t show up as matches for “Jane” at all in the ICW or Matrix tool.
I agree. We’ll see what they say. I forwarded your comment to the QA person along with your e-mail address.
OK, I am a newbie that wants to learn…regarding the matrix: I am the navy/black section right? I’ve got 3 people on chromosome 8 in exactly the same place, that means we are all related, right? Now what?
I’m lost here. The Navy black boxes are where the people who match you also match each other. You can then go to the chromosome browser to see if you all share some common segments. If that is what you did, then the next step is to contact these folks because you share a common ancestor someplace.
Hi Roberta,
This question is like Liz’ earlier one, but I didn’t see that she had used the ICW tool, and I’m trying to understand how the two relate.
I did an ICW for one of my matches, “N”, and came up with two names, M, and Q. When I compare all three in the chromosome browser, they all match me on chromosome 20 with lengths ranging from 9.6 to 29.6 cM. So my first guess would be that we are all descended from the same group. However, when I use the Matrix (or for that matter use ICW), M and Q are not related to each other.
So, M, N and Q all match me on C 20, but M and Q are not relations. Given the size the segment on C 20, which should be enough to trigger a match, my conclusion is that M and Q are coming from different sides of the family. N is matching one of them on C 20, but we can’t tell which one; and N must be matching the other person somewhere else.
Does this make sense? If so, I can see where the Matrix can really make it a lot easier to double-check your work. When I saw the initial chromosome browser view with the large segment that M, N and Q were sharing, it looked apparent that C 20 was the source, and I had a family group, but that doesn’t look to be the case.
Yes, this is exactly the purpose of this tool, and you are exactly right. These two people are coming from different sides of your tree.
Clare on December 12, 2013 at 7:06 pm said: “N is matching one of them on C 20, but we can’t tell which one; and N must be matching the other person somewhere else.”
If all three N, M, Q all match on C-20, why wouldn’t N match either M or Q on the matrix but not the other one?
Karen
Because there are two halves of the same chromosome, so you can match on the same segment but actually be matching a different half. Think of it this way. Half of your chromosome was all the letter A, and half was the letter G, you would match anyone with all As, all Gs or some As and some Gs. The As would be a valid match from one side of your family, the Gs from the other side of your family, all matching you on the same location, but not matching each other. The ones who are As and Gs combined would still read as a match to you, but they would not be a match to each other.
Pingback: One Chromosome, Two Sides, No Zipper – ICW and the Matrix | DNAeXplained – Genetic Genealogy
Pingback: 2013’s Dynamic Dozen – Top Genetic Genealogy Happenings | DNAeXplained – Genetic Genealogy
Hi Roberta
I have 2 ( call them B and T ) people that are a Matrix match, when I select them in my Chromosome browser at 1cm range, they do not overlap with each other in any chromosome bar. Does this indicate they are related closer to each other but less so with me?
They are 27cm( B ) and 24 cm ( T)with me. ( biggest single blocks are 11cm ( B ) and 15 cm ( T ))
kindest regards
If they are a matrix match to you and to each other, then should have matching or overlapping segments with each other. Report this to Family Tree DNA. Sounds like a bug.
Pingback: Autosomal DNA Matching within Projects | DNAeXplained – Genetic Genealogy
Pingback: Surname Projects | DNAeXplained – Genetic Genealogy
Pingback: Family Tree DNA Announces *Free Autosomal Transfer from 23andMe and Ancestry | DNAeXplained – Genetic Genealogy
Pingback: DNAeXplain Archives – Intermediate DNA Articles | DNAeXplained – Genetic Genealogy
Worth hunting around I did the matching thing before the holidays forgot how to use it or where it was. So now know they haven’t removed it and I am not losing my mind thank you.