DNA Color Clustering: The Leeds Method for Easily Visualizing Matches
Unsure of how other people were sorting their Shared Matches from AncestryDNA, I developed my own method: the Leeds Method of DNA Color Clustering. This simple and quick method helps you easily visualize how your close cousins are related to you and each other.
I created this method while working with an adoptee, and it works wonderfully for unknown parentage cases where the cousins are strangers and their true relationships are unknown. It also works wonderfully with traditional genealogy. In fact, it is being used to break down some more recent brick walls!
DNA Color Clustering: The Method
Step 1: List 2nd and 3rd cousins.
For this method, I use AncestryDNA and skip any “close family” and “first cousin” matches. Instead, I use cousins Ancestry says are “second” or “third” cousins. However, I do eliminate any of the matches near the top who share over 400 cM. (In other words, use those matches who share between 90 and 400 cM of DNA with you.) The key is to not include anyone who you share two grandparents with. For the spreadsheet, I am using Excel. I add “borders” around each cell – this helps everything stay readable when printed – and I “fill” the cells with color to create the color clusters.
Step 2: Assign a color to your first DNA match.
In this example, I assigned blue to person 01.
Step 3: Open that person’s shared matches and assign each of them the same color in the same column.
In this example, 01 showed shared matches with 02, 06, & 08. Each of these people are now a part of the Blue Cluster.
Step 4: Locate the first person who does not have a color assigned and assign them a color in the next column.
In this example, person 03 is the first person who is not already a part of the Blue Cluster. So, they were assigned the color orange in the next column.
Step 5: Open that person’s shared matches and assign each of them the same color in the same column.
In this example, 03 showed shared matches with 09 and 11. Each of these people are now a part of the Orange Cluster.
Note: Even though it didn’t happen in this example, a person would be assigned as part of the Orange Cluster even if they were already a part of the Blue Cluster! When someone is part of more than one color cluster, I call it cluster overlap.
Step 6: Repeat steps 2 through 5 until all of your 2nd & 3rd cousins have at LEAST one color assigned to them.
Analyzing the Results
4 Columns, No Overlap:
If your results show 4 distinct color clusters, like below, without any overlap, your sort is likely showing matches to your four sets of great grandparents.
Fewer than 4 Columns, No Overlap:
If your results show less than 4 clusters, it is likely these clusters represent 3 of your 4 sets of great grandparents and that you have no matches at the 2nd/3rd cousins levels who have tested for the 4th set of great grandparents.
Some Overlap:
If your results show 4 clusters but some of your matches have been assigned more than one color (for example, 02 and 04 are both blue and orange), your sort is likely showing either your four sets of great grandparents, but also showing you that two of these results (i.e. blue & orange) are on one side of your family. Or, the overlapped clusters (blue & orange) might belong to one set of great grandparents and, in this example, you are missing matches for one set of your four sets of great grandparents.
Lots of Columns & Overlap:
In this example, there is a lot of overlap between all of the clusters except the yellow and brick red clusters. All of the overlapping clusters are on the maternal side of this test taker and visually show a lot of cousins marrying cousins resulting in pedigree collapse. The paternal mother’s side is represented by both the yellow and brick red clusters. The paternal father’s side has no cousins matching at the 2nd/3rd cousin levels. So, even though there are a lot of clusters and matches, this sort represents only 3 of the four sets of great grandparents for this individual.
NOTES:
-
- Your results may vary!
- Color Clustering, as with most DNA results, is not proof of a specific relationship. Instead, it is a clue that can be helpful in determining relationships.
- Although all the results shown in this post are real, the names have been hidden for privacy reasons.
- A special thank you to everyone who allowed me to access their DNA results and gave me feedback! And, a special thank you to John Motzi who provided a lot of feedback!
This method has really helped me to quickly sort dna matches for both myself and for a couple of adoptees as well. Thank you !
I’m so glad, Laurie! Thanks for sharing 🙂
I am confused at the 1st hurdle with this. I maybe able to unravel this with time, but times not on my side. I am only trying to locate my son’s grandfather / wife’s father but have nothing to go on as she doesn’t know who he is and her mum won’t talk about it. I have successfully found my dad’s sisters through my heritage and daughter through other means.
Kind regards Chris
Hi, Chris. The Leeds Method was created to solve an adoption case so it can likely help you with this family mystery! You might try watching this free video to help: https://www.rootstech.org/video/you-can-do-dna?lang=eng
Dana
Thank you for sharing that video. I really appreciate the detailed explanation without it being so complicated. You are amazing to think of this process.
Thanks so much for your kind words, Jacqueline!
Hello Dana,
I clicked on the link but it was no longer there. Would it be available elsewhere i.e. Youtube?
Thank you
Hi, Nancy. I thought it would be up permanently, but I can no longer find it, either. If you have a Legacy Family Tree Webinars membership, you can watch a full-length presentation on the Leeds Method; just use the search feature to search for “Leeds.” This is even better as my part in the other video was only 20 minutes long.
Dana
video no longer available 🙁
That’s disappointing, but understandable. I have a full-length video on Legacy Family Tree Webinars. It’s for subscribers, only, but the site is well worth the cost. Or you can sign up for a trial membership. Just search under speakers for my name.
Dana
This link is not working. 😢
SADLY, after 2 years it appears RootsTech has removed this link. (I get it!!) If you have a membership to Legacy Family Tree Webinars, or would like to do a free trial, I have an updated presentation there! Go to https://familytreewebinars.com/ and search for “Leeds.”
Hope this helps!
Dana
SADLY, after 2 years it appears RootsTech has removed this link. (I get it!!) If you have a membership to Legacy Family Tree Webinars, or would like to do a free trial, I have an updated presentation there! Go to https://familytreewebinars.com/ and search for “Leeds.”
Hope this helps!
Dana
I’m aware this is an old post but for anyone (like me) who is just discovering the Leeds Meathod (thank you Dana) a YouTube search (in 2022😊) has many wonderful little videos to help.
Dolly, Thanks for sharing! Also, there is a full length video of me explaining it in detail and giving examples. It is with Legacy Family Tree Webinars and you would either need a subscription or need to use a free trial to see it. If you’re doing genealogy, I think it is one of the best subscriptions out there!
Yikes. This looks like just what I need, but the link won’t load.
Unfortunately, that video is no longer available. Instead, I have an updated video on the Leeds Method at Legacy Family Tree Webinars. The video is free to members. If you’re not a member, you can try a free trial membership. I am a member and think it’s well worth the price! There are tons of great videos there. I even just recorded a new one a few days ago which is a DNA-based case study that highlights the Leeds Method.
Here is the current link I believe: https://www.familysearch.org/rootstech/session/the-leeds-method-grouping-dna-matches-to-identify-shared-ancestors?lang=eng
Hi, Judith. First of all, my sincere apologies! I haven’t blogged (before today) in quite a while and somehow stopped receiving comment notifications and didn’t notice. I’m catching up tonight!
WOW!!! Thanks for the link of my speaking at RootsTech! I really thought it was gone.
Dana
link no longer works
Thanks! It should be working now. 🙂
I have been using a similar method of color coding the different branches
of my Family Tree but not on Excel.
Unfortunately a lot of my DNA Cousins DO NOT have SHARED matches.
It’s really Weird. Many of them might have the same last name, but
no shared matches show up for them. IT is kind of frustrating.
But thanks for sharing your organization methods.
That would make sense with some small cM matches. It also might make sense with recent immigrant lines where there just aren’t many people who have tested. Or possibly with small families. And you might try another testing site!
I’m so confused. How does this work when there’s ENDOGAMY?
Hi, Onyx. Unfortunately it doesn’t work with endogamy (when it’s across all 4 grandparent lines) or pedigree collapse. My favorite resource for endogamy is on Legacy Family Tree Webinars called “Dealing with Endogamy” by Paul Woodbury. You would need a paid membership or a free trial. Leah Larkin is introducing a new tool for DNA and endogamy at RootsTech in February! It has the name Banyan in it, though I’m not sure of the exact name.
Hope this helps!
Dana
Thank you for sharing!
You’re welcome! I hope you find it useful!
Can this method go out further to 4th cousins without blowing up or getting so complicated that any useful information gets buried in the “noise”
My 2-3rd cousin color clusters are very clean and I was able to identify the greatgrandparents and 2nd greatgrandparent clusters and which one went with which set…Hope to try back one more generation…
Hi, Erik. I’m glad your 2nd/3rd cousin clusters were so clean & you were able to identify the groups by both great grandparent & 2x great grandparents. I haven’t even went to the bottom of my 4th cousin list – or anyone’s – so I’m unsure. I “think” it would get too messy. But, please give it a try and let me know how it goes! 🙂
Hi Dana, I’m looking to find my Mother’s Biological father, and Ancestry was the first DNA company I used. My mother was very close to 100% Ashkenazi and my Dad from a completely different gene pool. However my Ashkenazi ethnicity matches on Ancestry only start at the 4th cousin level. I’ve colour coded using dots on Ancestry and I’m happy to send you the results. Of course it’s an endogamous population, but it’s colourful! I’ve be grateful for some advice here. Thanks for your excellent work. Kind regards
Hi, Alexandra. Unfortunately, working with Ashkenazi heritage is very difficult, as it sounds like you’re aware. And, the Leeds Method has not proven helpful with any type of endogamy. Lastly, when working with DNA matches, people are often encouraged to use only those matches who share 200 cM or more…. and you don’t have any of those.
My best suggestion would be to make sure you test the oldest generation – so in this case your mother if she’s living – and to have that person’s DNA in all 4 of the big pools plus GEDmatch (Ancestry, 23andMe, MyHeritage, and FTDNA). (If she’s not still living, test any siblings you might have, too.)
Here’s an article you might have read, but it also mentions working with those who share 200-250 cM of DNA: https://medium.com/@CleverTitleTK/no-you-dont-really-have-7-900-4th-cousins-some-dna-basics-for-those-with-jewish-heritage-857f873399ff
Hi Dana
Thanks for the link you have shared in your response to Alexandra’s post. I am not Jewish, but I also belong to a very endogamous population. I found some good practical tips for assessing/analysing my DNA matches.
Some of my top matches, I could link to my tree (only?) because I have an extensive family tree/paper trail that spans across almost all of my second cousins and many of my third cousins.
That article now helps me to focus on the better prospects under my top DNA matches to build out my broader family tree.
Hi, Marietjie. Thanks for sharing!
I wonder if we just did 4th cousin just by themselves it would at least group them together so we have a start. Might try when I ever have a spare minute. Thanks for sharing this is great.
Absolutely! I usually work with higher matches first, since those are more easily identifiable, but I definitely “cluster” my lower matches as the results can be really helpful!
Dana, Do you have any hints around identifying 2nd great Grandparent groups using a colour coding method? Is the range of matched cms too varying at this point?
Hi, Sharon. If you mean on Ancestry… I would suggest using 8 colors for your 8 great grandparents. Then you work with a single great grandparent and further break their matches into 2 more colors. I shared a similar method at i4GG last year and I think you can buy just that single lecture. (But I’m not positive.)
Dana
Hey Erik! Just wondering how it turned out using your 4th cousins?
I’d love to know the answer to this too. I only have 13 matches in 2nd/3rd range and 3 of those do not share matches within those 13. Help!
Hi, Jane. I’m guessing those 3 matches are on the lower side of the 400 to 90 cM range. And, you might find matches as more people test. And, you might find matches for them in those UNDER 90 cM. If so, use them! Try to see if you can figure out the connection between those matches. Best wishes!
Thanks so much for this wonderful tool. Want to share this with you, recently was able to quickly cluster matches for an adoptee who is a close cousin match for my niece. The Leeds method of clustering made quick work of sorting her lines. I am certain we could have eventually answered her questions without using the Leeds Method but not so quickly. It was days from start to finish, not weeks or months of following trails and possible dead ends before finding the answers. Whether or not there will be a happy ending for this case won’t be known for a while yet but the adoptee has answers and is very happy.
Roberta, Yay!!! That’s so great to hear. Thank you for letting me know!
Hi Roberta and Dana! I was recently given the link to this incredible tool, as I am the DNA manager for my dear friend who is searching for the biological family of his grandmother. I have an overwhelming number of DNA matches I’m sorting through, but I didn’t have any idea how to identify who, if anyone, among these 4k ppl could be related biologically to his grandmother (b. 1912). I don’t know if those ppl would be 2-3rd or 4-6th cousins, etc. Roberta, was the adoptee someone you discovered outside your tree, meaning that person was given up for adoption by members of your nieces family ?
Dana, into which category would the biological relatives of the adopted maternal grandmother fall? Lots of overlap, some overlap, or 4 columns no overlap? Does this method work independently of the tree information, meaning more reliable than say using the common ancestor info to match ppl against tree info when you aren’t confident that info is 100% accurate?
I am so completely obsessed and throughly invested in genealogy but I really have a hard time understanding the DNA side of it. I am working to “prove up” my friends tree, verifying every single person and relation on it using actual records rather than other tree info provided by other ancestry users. I suppose we should probably go ahead and get the other DNA test companies done as well. I haven’t seen what their results look like so if there is a similar DNA match list, would your method work on that list as well?
Hi, Jennifer. Sorry for the delay! As far as identifying which cluster(s) the family of his biological grandmother falls into, it could be any cluster. Your goal would be to identify those clusters you can identify and then work with the others as this unknown part of the family is likely in the unknowns!
And, yes, this method works independently of the tree information. It is only based on DNA and who shares DNA with whom.
Yes, my method works on all of the sites. You can transfer your DNA for free to MyHeritage and FTDNA. You have to purchase a test for 23andMe.
Hope this helps!
Dana
Hi Dana
Thank you for sharing your method. For me it results in only two colours for a 12 person list. Am I doing something wrong? I have only listed the cousins from my father’s side, as it is he I am seeking to find. Can you help please?
Many thanks in advance.
Hi, Donelle. If you are only working with 2nd/3rd cousins on your father’s side, you “should” most likely only have two columns! Hopefully, these two columns are your dad’s mom’s side and your dad’s dad’s side. I just published another post about identifying the common surnames for these unknown people in the clusters. Hopefully, it’ll help!
Thank you Dana. I’ve done the next step and identified common surnames. However, my father’s surname is not among them. I’ve built a mirror tree in hopes of identifying my father’s mother, using the common surnames in my matches’ trees, but although there are plenty of candidates, I cannot find any record of a birth that could be her. Not sure where to go from here. Any suggestions?
Hi, Donelle. Just trying to clarify, you said you were trying to find your father, but on the second comment you said your father’s surname was not among them. Do you know your father’s name? If you do a “search” of the surnames for your DNA, do you get that surname at all? Feel free to contact me via email.
Hi Dana
Thanks for your reply. I would like to contact you via email but can’t find your email address. Could you send please?
Donelle Staley
Help !! I’m dealing with over 8000 Ashkenazi matches. My paternal line. Ranging 7 to 200cm. I’ve sorted and researched 50 people. Not a single common match going back to gr gr grandparents. I have through MH found a family name “S” which I had a DNA match to but records show most wiped out WW2, 1942. My paternal grandfather had a different last name, identified from a ship’s manifesto when he travelled with his family and my paternal grandmother (the nanny)
Where do I go from here.
Hi, Carol. Sadly and frustratingly, clustering doesn’t really help with endogamy. My number one recommendation is to watch a new video by Paul Woodbury at Legacy Family Tree webinars. If you don’t already have a subscription, it is well worth the yearly subscription fee. The video is here: https://familytreewebinars.com/download.php?webinar_id=1292
Best wishes!
Dana
Sorry! Another issue I need to fix on my new website. And, I’m planning on getting a new email address. Anyway, my email is drleeds@sbcglobal.net (my initials – not a doctor!) 🙂
Hi Dana: I have an adoption paper and the bio father; was proven last year to be friend genealogist’s relative; in 2020; not my bio. I don’t have bio mom’s side info. I am Beverly Allen farmbev@gmail.com; and looking for my bio father. I am hoping you can tell me how this helps me. On Ancestry DNA I have one first cousin; 1189 cM; and 33 segments. Clue for bio father; and I have 5th to 8th South Asian father and son 17 cM cousins; the son emailed me. I know these are clues for my bio dad. I don’t know how this would help me with the colour codes. Would be nice to know how to use these; thank you for any help. If I may be of any help; please let me know by email. Thank you for your time and hope we can keep in touch, I got this sent from my DNA teacher for genealogy; we only have zoom once a month. No time individually.
Hi, Beverly. If you’ve created a chart, the 1189 match can help you. Look and see which color clusters that match matches. Those are the color clusters you should work as since they are on your bio father’s side of the family. If you’re on Facebook, you might post on the group “DNA Detectives.” Someone there might be able to help you more.
Hope this helps,
Dana
Just curious as to why you do not use matches with cms > 400.
Hi, Chip. Good question! When I started doing this method, I was working with adoptees so we didn’t know the relationships of the matches to the test taker. If you include someone who shares grandparents, instead of great grandparents, you essentially will “merge” two of the clusters. So, you lose some of the clustering effects. But, if you DO know the relationships, you can use anyone who does not share grandparents, but instead shares great grandparents with you. You can still use these closer cousins to help you label the clusters, though. Or, you can even include them but put them in two columns for two grandparents. 🙂
Thanks! Really love the visual image this method produces. Just used it with a few folks I am working with and to our surprise of the newer ones appears to have one of her lines unrepresented (immigrants in the early 1900s). Following up now to see if that is, in fact, the case or if I messed up somewhere.
Hi, Chip. From my experience, and what I’ve read from others, branches that include recent immigrants often do not have any/many DNA matches. I think it’s a great idea to double check your work, but I think it is likely you’ll have the same results. Are there any cousins from the home country who could test?
I tried this with my Ancestry.com results, trying to identify my father’s three brick wall grandparents. I only have one third cousin once removed there, but made some headway in organizing everyone and some ideas for searching.
Now I’m trying to do this at Family Tree DNA where I have a half first cousin, once removed and some 3rd cousins. It’s much more tedious to work on there, but i have more people who match, so I will see what happens.
Thank you!
Denise, Please let me know how it goes! I haven’t tried it at FTDNA yet.
Hello Denise,
Did you get anywhere with the FTDNA project? I have a 67 marker, Y-DNA kit there and would like to use Dana’s system with my results but frankly I don’t know where to start.
Hi, Paul. I am guessing Denise meant she was trying it with the FTDNA Family Finder test. That is FTDNA’s autosomal test which is what the Leeds Method was designed for. This method wouldn’t be helpful with a Y-DNA test.
I don’t have any posts about Y-DNA, but you might start with this post from my friend Roberta Estes: https://dna-explained.com/2020/01/02/y-dna-part-1-overview/
Dana
will this be useful when you know the mothers side but trying to figure out the fathers side. Unfortunately there is only one 3rd cousin match at 93 CM and then a whole bunch of 4th cousins from 68 through the 30’s. There are a few decent trees but I am still having difficulty pinpointing any real starting point. thanks for all the great information.
Angie, I actually developed this method while trying to separate paternal from maternal matches. So, yes, it should be helpful in that way. However, it does depend on enough “good” matches. I have not tried it this way, but if you have a handful of 4th cousins who match at 80 cM or more, then see what kind of clusters they create. Also, do you have your DNA at FamilyTreeDNA, MyHeritage, 23andMe, and GEDmatch? You never know where the closest match is going to be!
I’m using your very helpful method in for a friend who has 3 predicted 1st cousin matches – how lucky! Did as you suggested for 2nd and 3rd cousin and then went back to 1st cousins which allowed me to easily differentiate paternal from maternal.
Chris, I’m so glad it’s working for you and your friend!
I am trying to break a brickwall. I found out a few months ago that my paternal grandfathers dad isnt his biological dad. It has been a struggle tto try and piece info together.
I did the Leeds chart and added all the 1st and 2nd cousins.They all seem to match except one , but I knew what line he was from and just put him there.
I dont seem to have any that are not accounted for. Am I doing something wrong or should I be doing something else? Instead of using my matches, should I be using my dads?
This is all new to me and gets a little confusing.
Thanks
Susan
Hi, Susan. I would definitely work with your dad’s DNA. Start by making the chart with those AncestryDNA labels a 2nd or 3rd cousins, but keep the matches under 400 shared cM. Then, see if you can label the columns as to which of your dad’s grandparents each column/cluster is related to. You can follow up here or email me at drleeds@sbcglobal.net
Dana
Dana,
Thank you for sharing. I have used the Leeds Method for a few kits I manage including my own. I must have done alright without this technique, but it does add some simplicity, something I’m always looking for to make processes more efficient. I did add one, actually two, more steps that you may like to share. When a cell is filled with a color as you describe, I will add the surnames of the most recent common ancestors (MRCA) if they are known. For example, in my personal chart, I have 2 “second” cousins and 7 “third” cousins in four columns. In one of those columns, the colored cells have “Gordon-Stewart” (my great-grandparents), “Stewart-Dougherty” (2nd great-grandparents in my great-grandmother’s line), and one blank. The blank is an unknown relationship to someone who doesn’t answer emails. I’ll figure that one out eventually. If there are any cousins with an MRCA beyond my great-grandparents, such as the Stewart-Dougherty cousin, I’ll change the color in the cell to a darker shade of the same color if they are in the paternal (Gordon) line or lighter if they are in the maternal (Stewart) line.
Adding MRCA’s allows for an instant determination of unknown relationships. My goal is to fill in the blanks. After applying these two added steps to the Leeds Method, I’ve found working with 4th cousins to be a little easier to visualize, too.
Hi, Brian. I think the “simplicity” and efficiency of this method are two of its top qualities!
Thank you for sharing your two additional steps.
1. Adding shared surnames to the cells – this is a good idea. I’m still trying to keep mine very simple, and think this might take away from that, but maybe it won’t! For now, I’m adding column headings naming the set of great grandparents that each column has in common. But, your method is more specific!
2. Changing color darker/lighter – I love the idea of making the same color darker or lighter for paternal vs maternal lines. Nice! I will see how I can incorporate this in my own Color Clusters.
Thanks!
Dana
Hi!
You say not use this method with people who share 2 grandparents with you but what if you don’t know this information?
I’m trying to find my half brother who was adopted out or a descendant of his. I found someone I share 240cms with and can’t figure out how We are related. I’m wondering if maybe it is somehow through my half brother.
Hi, Leah. First of all, best wishes on finding your half-brother!
I recommend the 90 to 400 cM range as that works well for most people. If you don’t know how someone is related to you, then you can use that range. The only time it’s really an issue is when it’s the top (or key) person in a particular cluster. If they match two of your grandparent lines, they can make these two clusters “merge.”
So do the Leeds Method then try to identify the clusters that formed. Then, to which group or groups does this mystery match belong? Are they on the correct side of your family? If you’re not ready to reach out to this match – or if you’ve tried but they haven’t responded – to they have an unusual name? Or clues to their age or where they live? Do they have a tree? Sometimes we can use clues to figure out who the person is and contact them via Facebook or email or in some other way.
Hope this helps! (And, since I just responded to another question about consulting, I do want to let you know I offer hourly consultations if you find you need additional help.)
Dana, thanks for explaining this so clearly.
I’d like to ask about using this starting with 4th cousins rather than second cousins – any tips ?
That’s essentially my question, though more details about why I’m asking and what I’m thinking are below.
I have one second cousin match – and I know his ancestry and how we are related – but after that the closest matches are 4th cousins. They include a 4th cousin and a 5th cousin I can definitely place on my father’s side, a 4th cousin on my maternal grandmother’s side, and two 4th cousins where I can guess which set of great-great-great-grandparents are involved, based on surnames (and if so these are both related by my maternal grandmother’s family).
I’m in touch with all these people. None yet identified for my maternal grandfather. I have all my great-great-grandparents in my tree and about half of my great-great-great-grandparents.
So here’s what I’m thinking. Just as 2nd cousins should fall into 4 groups, based on the great-grand-parents involved, so 4th cousins should fall into 16 groups – or fewer, if there are some branches where no-one as close as 4th cousin has tested. These would be based on which of 16 sets of great-great-great-grandparents are the common ancestors. (I am looking at this in fact as which of 16 great-great-grandparents had the sibling who is the ancestor of the 4th cousin, because it’s easier for me to visualise that way as I know my Gt-gt-grandparents). I know there could be “half 4th cousins” which might complicate things but in fact I’m not aware of any at that level – could be though)
So here’s my question.Do you think it is worth trying this method starting with 4th cousins mostly and looking for 16 groups? Do you have any tips?
( I haven’t tried this yet. One thought I’ve had, bearing in mind I’m a bit of an Excel geek, is that with maybe 16 groups I might use letters in columns in just the same way you have used colours, but use Excel conditional formatting to turn colouring on. That way I should have the visual effect of the colours, but it will be easier to use Excel to show overlap automatically, i.e. how many columns are filled for each person. This would also mean I could switch colours easily, e.g. if it turns out that letters A and G turn out to be related to a great-great-grandfather and great-great-grandmother couple, then when I have identified this I might use two related colours, e.g. I might set A=red and G=pink)
Any thoughts would be welcome as I try to get more clues about my 148 4th cousin matches!
Thanks
Mike
Hi, Mike. First of all, I think there are some drawbacks or dangers with working primarily with 4th cousins. But, I’d still suggest you give it a try!
I am really not sure how many clusters you would have. Even when using 2nd/3rd cousins, we often end up with 3 or 5 or more columns. And, as the pieces get smaller and the chances of you sharing with a specific “4th cousin” decrease, I just think it will start falling apart.
But, since this is what you have to work with, I would give it a try! Hopefully, even with the first 10 matches or so, you will see whether or not it is working. And, if you only do 10-15 matches and it seems like a mess, you won’t have wasted very much time.
One other question: do you have your DNA on other sites? Is there a different site where you have higher 2nd/3rd cousin matches? That would be worth a try, too.
Best wishes! And, please let me know how it works out.
Dana
Thanks Dana – warning duly noted! But I have been meaning to try to sort out which matches go together, so this may be worth a try when I have time.
I have uploaded to gedmatch – highest cm match is 37.4 which I understand is likely to be 4th cousin approx. I have 60 entries in gedmatch “1-many” above 25 cm. I think I’ll start with the Ancestry matches though – at various times I have added notes against some of the Ancestry DNA matches, so I want to bring that work together first; I’ve noticed some clusters of shared matches, and used some of that information, but this would be a way of recording that in a structured way.
Mike
Hi, Mike. I think AncestryDNA works best for most people because of the sheer number of test takers, number of cousins, and trees. Let me know how it goes! This method is really new and there’s a lot still to learn about it!
I Dana. Your method is fantasatic . It makes everything so easy. I have been able to colour code all my father’s parents very easily and my mother’s father’s also without a problem. My father’s have been coloured code bright pink and light pink and my mother’s father bright blue. However the elusive one is my mother’s mother which is the “mystery” one of which she herself did not know is missing. She was placed in care at an early age and she had Chinese ancestry. Could I assume that all my Ancestry DNA matches without a colour code could be her line. They are all matches to each other and not to the other codes. (If you know what I mean) I chat to my known relatives all the time and they all tally up.
Hi, Helen. I’m glad you are finding it helpful! I wouldn’t say you could assume they’re all her line, but you can work with that as a hypothesis – especially if they match others in that group.
If this is the only line that has Chinese ancestry, you can also do a “search” of your matches for a birth location of China.
Hope this helps and best wishes!
Dana
This is great, thanks! Would you say this method can still work well when endogamy is at play?
Hi, Tania. I’m still working on it with engodamy and pedigree collapse. First of all, for adoptees or people working with a tree they don’t already know a lot about, this method does a great job of showing endogamy. But, so far, I haven’t been able to figure out how to sort those with endogamy and pedigree collapse. If you figure something out, please let me know! 🙂
Wow! What a great tool this is!! It was fun and educational. I have someone I’m trying to help nail down info on her father. She’s 77, so it’s a pretty safe bet he’s dead, but she just wants to know something about his family. (She’s not an adoptee, but might as well be for this purpose.) We can easily identify her maternal matches, so the only ones I plugged into this chart were her paternal matches. Got two distinct groups! I kind of already knew this, but now it’s crystal clear. The 12 matches I used ranged from 192 to 100. I am nearly certain his parents were not married and I’m not very good on “halfs” — does this mean I should be able to find a set of GGGrandparents in each group of the two groupings?
Thanks, Elana! Since you only created the clusters for the paternal side and got 2 distinct groups, these are most likely the paternal mom’s side and the paternal dad’s side. Matches in this range are likely 2C, 2C 1R, or 3C. If 2nd cousins, they share great grandparents. If 3C, they share great great grandparents. So, you can’t tell exactly who they would share, but you can narrow it down to common surnames. Have you tried the 2nd part of this method: https://dnawithdana.com/dna-color-clustering-identifying-in-common-surnames/
Hope this helps!
Dana
First of all, let me tell you how much I enjoy keeping this spreadsheet on my Desktop so that I can consult with it easily and often! I added a column for cMs, so I have every piece of information I need at my fingertips.
I worked on the “2nd part” as you suggested, although I don’t have much luck with identifying surnames since they are much the same all through this group — small community, all of Scottish background. I do have the MRCA of quite a few of my friend’s matches, so I made note of those on the chart as well. Armed with this information, I weeded out all of those ancestors who would not have qualified to be the birthparents — age, geography, etc. I came up with two good possibilities for the parents, plugged them into the tree, and immediately almost every DNA match fell in place with approximately the right cMs for each match. However, I was then told that 8 “segments” were not sufficient for her closest match (192) to be a second cousin. Sigh! Is there a chart somewhere that breaks down segments in a similar manner to all the charts that we have for cMs?
Elana, I’m so glad it is being so useful that you are keeping it on your Desktop! I’m not sure where you saw 8 segments isn’t enough to be a 2nd cousin. I primarily look at cM only. My favorite chart is a free tool at http://www.dnapainter.com where you enter the # of cM and it gives you the probabilities of different relationships. It also has a new tool called “What are the Odds?” which can help you if you think you (or a person) might fit into a tree in one of two or more ways.
Hi Dana, the Leeds method is the first tool that has actually helped me sort matches into usable information. I love it but I don’t know what to make of matches that are likely related to cousin marriages. I’m working on a tree that has parents as first cousins and great grandparents that may also be first cousins or cousins of some degree. The great grandfather is the brick wall with this family. To complicate matters, the great grandmother had a first marriage resulting if half siblings. I have a 2-3 cousin match of 123 cM that is descended from this first marriage. I’ve put this match into a separate column. The largest cM match is 483 and this match is also frequently the first shared match of some 3rd and 4th cousin matches. I have no idea who this person is and I put that match into its own column. The second highest match drops down to 141 cM and I know this match is my 1C2R. The MCRA is likely from the possible cousin marriage of the great grandparents, but not through the cousin marriage of the parents, but there could be shared DNA of the grandparents since they were siblings. I hope this makes sense because it’s got my head spinning. I don’t know what to make of the cMs given the cousin marriages. The vast majority of the cousin matches are below 90, but are shared over several segments. Existing family members aren’t willing to be tested and all but one aunt is still alive. Any advice is greatly appreciated.
Hi, Eileen. Unfortunately, working with “cousins marrying cousins” is something that creates a lot of difficulty with all DNA methods that I am aware of. I have this situation in my one of my four grandparents, too. So far, I haven’t figured out how to work around it. But, I think we will find a way eventually! In the meantime, keep in mind that the projected shared cM represent can be confusing. For example, if you share 200 cM with someone, it might actually be more like 100 from one side and 100 from the other. So, the Shared cM charts won’t work as well.
I will definitely be sharing how to work with these situations when I, or someone else, works this out and lets me know!
Dana
Thanks for responding Dana. I have a follow up question. Ancestry and FTDNA use different categories of cousin relationships. I have far more “close” matches on FTDNA than Ancestry. For your method, you recommend using cMs between 90 and 400. Does that range typically include just 2nd to 3rd cousins? And is anything less than 90 cMs less useful in your cluster method or are they usually 4th cousin and greater matches? I have plenty of those and I can see how they connect with other shared matches. Thanks in advance.
Hi, Eileen. Because the different testing companies list cousin relationships differently, I have suggested using 90 to 400 cM for your original sort. But, if you KNOW the relationships of specific individuals, you can use them even if they aren’t falling within that range. Since you are seeing the connections, and if they are making sense, you are probably doing it correctly!
Absolutely going to try this!! Trying to find anything that I can work with to trace my grandfather. Ancestry matching has been hopeless – I don’t recognize a single name on my aunt’s (last surviving offspring) DNA matches with the exception of my siblings and one cousin that’s recently tested. I’ve not been able to locate him, alone or with his parents on any online census. If I can perhaps get clues from other matches trees, I might actually get somewhere.
Judy, I hope this method helps! It “should,” but it also depends on who has tested. Please let me know if you are able to find your answers!
Thanks Dana! I’ve been color coding matches in a spreadsheet format for a while but your technique has a much better visual representation of the clusters, especially for explaining it to other people. I’ve started using lighter shades of the same color for columns with more distant connections to the same family. I shift them columns to group them when I’m all done. Very helpful. Thank you for sharing your work!
Christine
Love that idea!!
How do we identify the colum names? We have been working for a couple of hours to figure out and understand these instructions, and we are not understanding it, could you give us a bit more detail on how we identify the colum names?
Hi, Robin. Do you mean identifying the surnames that belong with the columns? At first, I just use the color names for the columns. If you know the relationship of a group of people to you – which you really only need to know a few at first – you can name that group as “maternal” or “paternal” or possibly “mom’s mom”, “mom’s dad”, etc.
Robin, Do you mean labeling as maternal or paternal? Or with surnames? If you are using family members you already know (vs working as an adoptee or something similar), hopefully you can identify some of the cousins and label them accordingly. If you are talking about surnames, you need to look at the surnames of the Shared Matches. If looking at the 4th generation, great grandparents, doesn’t give answers, then look at the 5th generation, 2x grandparents. Hope this helps!
Hi Dana,
Have you looked at, or considered using/collaborating/coordinating this method with DNAPainter?
Joe
Hi. I may have missed it but I needed some clarification. After we add people to a cluster, do we not look at each of their shared matches? For example, if persons 2 and 3 go into a cluster with person 1, but person 4 doesn’t, I know to start my next cluster with person 4. But do I not look at shared matches for persons 2 and 3? Or do I look at everyone’s shared matches regardless of whether they have been included in a cluster already? Thanks.
I am not sure my prior question came through so let me try again. When you begin and assign a color to the first cluster, say blue, you say to go to the person without a color assigned. So, if persons 2 and 3 match person 1, do I check persons 2 and 3 shared matches even if they are already in cluster blue. I know a start a new cluster with person 4.
Hi, Lana. With the method has you just go straight to person #4 and start the new color cluster. However, I’m working on a post about “checking” your work & I have found it IS important to go ahead and, in your example, check person #2 & #3. So, you can do this as you work or at the end as a “check.” Hope this helps! And hope to post within the next few days.
All of my 2nd cousins are more than 400cMs except for one, and I already know she’s related to me on both my maternal and paternal sides. I’m afraid she’ll mess me up so I didn’t include her. I only have 6 3rd cousin matches. Can I include 4th cousin? Why or why not?
Carolynn, Unfortunately, yes, you need to skip that 2nd cousin then. I don’t think it will be as effective, but you can try adding some of the higher 4th cousin matches. Or, you might see if you can specifically find and test some of your 2nd/3rd cousins to get more DNA data to work with.
On your diagram you say 2 and 5 blue and brown….it’s actually 2 and 4 that are blue and brown from what I am seeing.
I love this method it has helped me sort threw things easier.
Thank you very much
Thanks! Fixing it now. 🙂 And, glad the method has helped you see things easier!
So I have been having some good results confirming the method. I am looking for an unknown paternal grandparent and have a maternal aunt that I have tested. Using Match-O-Matic (DNAGedcom) I have subtracted out all the maternal matches so I should have a clean stack of paternal matches to sort through. Everything was looking normal until I got a little confused and looked at the matches for someone already assigned a color in the first column, (in your example above the blue color for person 2 in column 1). Now this person is matching different people in the first column that did not match Person 1. I would have not seen this if I hadn’t goofed… I would have followed the procedure and gone to a person without a color and started another color group column. Is this just an odd (one off occurrence) or do I need to check the other people in that first column? Should I start a new color with Person 2 (make them double colored) and indicate their matches separate from the Person 1 group? I didn’t expect this and now worry that I may be missing multiple color block people (I assume pedigree collapse at this point).
Hi, Erik. I don’t think this is isolated case unfortunately. I’ve been meaning to create another post but have been traveling and have a sick child. I will try to post it today. But, I do think we ALL need to create the clusters and then “check them.” What I’m finding is that a few people who end up being related to the test taker through more than one line can throw this off unless you DO check it. I suggest looking at the shared matches for EACH person and making sure they do not actually belong in two color clusters. If someone does belong in 2 color clusters & you used them to start a group, this could definitely throw things off! Sorry I haven’t posted this sooner!!!
Looking forward to this post as I’ve tried both ways and each have their merits and challenges!
And, today I was able to use this method and some follow up research and find the grandparents of a very elderly adopteee. My first attempt at this and it worked out in a few hours!
Thanks for a great method and the ongoing refinements.
Li, I’m so glad you’ve found this helpful! I think it is a terrific system for working with adoptees.
I did my chart for my mom’s DNA. I ended up with 8 columns and a bit of overlap. I was able to determine all of the clusters were for 3 of 4 grandparents of my mom). It’s the 4th we are trying to find.
Sheri, Then I suggest moving on to the 4th cousins and sorting them into the columns you’ve already created. Hopefully, you’ll come across some that do not match the 3 columns/3 grandparents and you can concentrate on them. Another possibility, if your mom’s 4th grandparent is unknown, is maybe that grandparent is actually related to one of your other grandparents.
This is great idea. My problem is, what’s the quickest way to download the matches information from Ancestry? If I just copy paste, I have to format everything for Excel. I have many, many matches.
Any ideas?
Thank you, Patti
I have never done more than 75 or so matches, so I just type them while having two screens open – one with Ancestry & one with Excel. If you want to do hundreds of matches, I suggest one of the automated Leeds-based methods like Genetic Affairs’ AutoCluster or DNAGedcom’s Collins Leeds Method. – Dana
Thank you!
Patti, I know this is an old post, but it is extremely worthwhile spending a few hours watching Excel tutorials, especially to learn tips and shortcuts.
Thanks !
Hi I am having trouble understanding all this but I really want to give it a try.
My question is at the start when you put the first person in and assign the blue colour then go to there shared matches and assign them all with blue, why are there gaps, why are they not consecutive eg person 1, person 2, etc.
Thank you
Elizabeth
Hi, Elizabeth. You make a list of the people first – 2nd & 3rd cousins who share less than 400 cM. On all of the test sites, these people are in order from most to least amount of shared DNA. So, the order of the list doesn’t change. Then, you pick the first person & give them a color – like blue. Then, using that same list – that’s already created & in order from most to least DNA – you give everyone who is a shared match with that person the same color – Blue. Does that help? You might also try a video that was recorded of me teaching at RootsTech. There’s a link on one of my podcast or you can google “You Can Do DNA” & “RootsTech.” There are 3 of us speaking & my part starts at ~40 minutes in, though I recommend watching the entire video. Hope this helps!
I am new. Found my BF easily with DNA test as I had a name. I am trying to help my maternal 1st cousin find her BF. Her DNA shows relationships to names she has never heard of. I have sorted her matches into maternal and paternal and now have used your method in excel to sort the paternal matches into colours. 4 colours with some of them having 2 colours. There seems to be one family name in common with the ones with 2 colours. Most don’t have trees or if they do they are limited or private. Am I on the right track and where/how do I proceed from here?
Hi, Shirley. Yes, it sounds like you’re on the right track! So, within any group – or if you have 2 color groups that seem kind of combined – you do want to focus on a surname (or surnames) you see repeating. Then, you’re trying to build out those trees and have them intersect. That can be a problem with limited or private trees! With private trees, you can ask if those people will share the tree with you (or your cousin) – but take photos of the information you do have before hand in case they delete it. With limited trees, that’s sometimes where some higher skills come in. You can build “quick & dirty trees” and you really only need one person who is identifiable. That could be you have a “unique” name or you have a name with dates and places of birth or death. You just need enough to know you can identify that person. I plan on writing about quick & dirty trees, but in the meantime you can Google a video by Blaine Bettinger if you need help. Hope this helps!
Dana
Thank you for your reply. I am manager for my cousins ancestry account. She barely uses the computer. I watched the video on building Q&D trees and will try that. I am finding surnames that repeat so will divide her paternal side yet again. The 4 colour groups almost condence down to 2 except for a few that in both groups that only have one colour…now to figure what names go with what colours. I know I am working with all of her paternal side.
It sounds like you’re on the right path!
Oh no! I have done my four colors but there are a few with no color. What to do?
Hi, Allison. The goal is not to create 4 colors. The goal is to give every match in the range you’re working on a color. So, keep adding colors if needed!
So what does that mean if we have 5 or more colors? I thought the purpose was to group these matches into clusters based on the testee’s 4 grandparents?
Hi, Meredith. In an “ideal situation,” you would get 4 clusters. But many people get more or less! It all depends on your ancestors and if they were related or not AND it depends on who has tested. With 5 clusters, it is likely that 2 of those clusters represent one grandparent. In this case, you likely have 3rd cousins and not any 2nd cousins for that particular line. Or, it is also possible you need to combine 2 of these clusters. You should do this if you have heavy overlap – in other words, most of the people in one color cluster are also in the other color cluster and vice versa. Hope this helps!
Pingback: Secrets of the Past Revealed – Welcome To DNA Enthusiast
Just want to clarify one question about the “shared matches” I find between me and any particular individual – If I am checking Person 1 in my list, and s/he has a shared match with me on, for instance, Person #6, but not a “triangulated” match (i.e., we don’t all 3 match on the same exact chromosome cMs), do I still assign a common color to that match? Or does each shared match also need to be a triangulated match?
Hi, Susan. You do not need to “triangulate” the matches. If Person #1 shows you Person #6 as a Shared Match, then they are both in the same group. A person only shows up as a Shared Match if they share at least 20 cM with the other person – that might be the issue that you are noticing.
That’s helpful. I’m looking at my matches on MyHeritage, and I get tons of matches, but not nearly as many triangulated matches with each person. I didn’t want to do all that work and find out 80% of them should not have been included, but this works the other way! I find more matches than I expected to:) Thank you.
Thanks for leaving the comment, Susan. I’m glad you have found it useful and found more matches!
Pingback: Coincidence? Perhaps, or Sheer Luck! – Welcome To DNA Enthusiast
Pingback: Prove & Extend Your Tree with DNA Research - Shannon in the Country
I feel like this method is a huge game changer in my research! Over 20+ years ago I was taught “start with yourself (or what you know) and work back”, but that was before DNA came in to the picture. Now with this method and starting with 4th cousin matches, you can actually start in the past and work forward. I was able to help a family member verify some questions in their paternal line by using this method proving who they were really related to instead of who they were told they were related to. This method has also allowed me to figure out one more branch in my own tree that was previously a brick wall. Fortunately for me both my parents had agreed to take DNA tests a few years ago so I will be able to use this method forever. Thank you for making the science of DNA and researching family history so simplified!!!
Lori, Thank you for your kind words! You made my day! And, I’m so happy that you are finding this method so helpful and being able to help other with it, too.
Hi I want to try your method, but I’m a bit lazy. Is there any way to download my 2 and 3rd cousin matches from Ancestry.com rather than retyping them all?
You could use DNAGedcom – which also has an automated Leeds-like tool – but it costs $10 a month. You can try Genetic Affairs’ AutoCluster which is also based on the Leeds Method. You can use it FREE for your first 8 runs, so I would do it but make sure you use 400 to 90 cM to start. You can find out more on my “Leeds Method” page. Hope this helps!
I did a simple copy from my DNA match list from 400cm Dow too 100cM. Then pasted into a word doc deleting everything except the names. The did a copy/paste into excel under the name column. Took a couple of minutes and saved all that typing.
Nice! And, I know it’s a lot of work, but I feel I get to know the names a bit while typing them. 🙂
Great idea, Louise! One reason some of us “enjoy” the typing is because we get familiar with the names. But, this is a great time saver! Thanks for sharing!
Louise, I am doing the same thing! I’ve got my own DNA, my mother, my maternal grandmother, and both paternal grandparents. Those five testees are all uploaded to Ancestry, MyHeritage, FamilyTreeDNA, & 23&Me. That’s a LOT of match lists to go thru!
Lucky you! Great to have those matches.
I used this method for the first time today. My husband has only one 3rd cousin match so I charted his 21 4th cousin matches and I have 18 different colors.
There is no overlap and 4 clusters of 2 matches each.
Are these results common?
Paula, Yikes! Unfortunately, it’s the 2nd and 3rd cousins that usually help tie the 4th cousins together. So, without those higher matches, it is likely that your 4th cousins will not share overlapping DNA segments and so will be forming clusters of one or two. I’m not sure how common this is, but it definitely makes sense – especially in countries or with populations who haven’t done much testing. I hope he gets more matches in the future!
I love this method, in theory. My results, however, have indicated something unexpected that makes any separation and identification difficult at best. My paternal line is coming up as “mother’s side”. I know at least 3 close relatives on my father’s side that show mother’s side. The percentage seems to be small but it indicates endogamy somewhere in my family history. Every match is viewed as maternal. Ugh!!! Any suggestions on how to unravel. Thanks!
Hi, Robin. Is this on Ancestry? And, have you also tested your father? How close are these matches? (Also, I’ll email you and see if I can help.)
Dana
Hi, just found out about this technique after getting the Genetic Affairs newsletter in my email this morning. Is it OK to combine names from FTDNA/23andMe/Ancestry in one single spreadsheet or is there a reason I wouldn’t want to do that?
Hi, Greg. You wouldn’t be able to create the clusters with all 3 testing sites at one time, but after you figure out various clusters you might combine them. Say, for example, you did find clusters for your 4 grandparents on multiple sites. You could combine those on one spreadsheet and include a column that says which testing company they tested with.
Hi Dana – thank you so much for this method. You are getting lots of mentions in the Ancestry DNA Facebook group, so I imagine you are helping lots of people. My question – I have identified all four grandparent lines (one was a complete surprise) and every 2 or 3rd cousin fits in these groupings except a second cousin who doesn’t fit any of these groups and four other matches who share DNA with both of us, but I can’t see any matches in this group that align with my four grandparents. I have marked this group of five with a fifth color. Any idea how that could happen? Thank you in advance for any ideas! Dana (it is unusual to find another Dana 🙂
Hi, Dana. It’s been amazing that I keep running into Dana’s doing genealogy when I’ve only met a few outside of genealogy! 🙂 Here’s my guess… I am guessing that whichever grandparent this cousin and their matches are tied to only has more distant 3rd cousins and the DNA is just not overlapping. So, you actually have 2 clusters for one of your grandparents. This often happens when you don’t have any higher matches – like closer to 400. Hope this makes sense!
Hi Dana,
Very interesting exercise. A couple of questions: 1. Should you also remove half-1st Cousins? 2. On my maternal-paternal split to three clusters – since I only recently learned about this relationship (result of an affair), are there things I should look for in these three clusters? 3. Both sides of my family or from central KY, and going down to 60 matching cms, more overlaping clusters begin to show up – for overlaping clusters how do you peel back that onion?
Again excellent exercise, helped me better understand all this relationships.
Bob
Great questions, Bob. And, I’m glad this exercise helped you!
So, no, you don’t need to remove half-1st cousins. For the initial sort, you only need to remove matches who are related to 2 (or more) of your grandparents. Of course, you can add in the higher matches after you’ve built your initial chart.
As far as the 3 clusters, if there is a lot of overlap you can merge them.
As far as the 3rd question, if your KY ancestors are like my TN ancestors then you have pedigree collapse with cousins marrying cousins. In this case, I would suggest trying an automated tool such as Genetic Affairs’ AutoCluster to create clusters. Or, you can do it manually.
Dana
Hi, Bob. Sorry for the delay!
1. Yes, you should remove half 1st cousins and remove anyone who shares more than one grandparent with you.
2. With the 3 clusters being from one part of your family, I would suggest just working with the matches in each cluster. Try to figure out how they’re related to each other and then how they’re related to you. Also, keep in mind that they are likely 3rd cousins as 2nd cousins – or cousins that share a set of grandparents with us – often create a single cluster.
3. Your Kentucky ancestors might be like my Tennessee ancestors – they had instances of cousins marrying cousins. This gives us DNA matches who are related to us in multiple ways! Depending on how messy it is, you might rely more on traditional genealogy for this part of your family rather than DNA matches. But, these matches can still be helpful and give you hints!
Best wishes!
Dana, on Feb2, 2020 you said, “So, no, you don’t need to remove half-1st cousins”, but then above, Feb 26th, you said “you should remove half 1st cousins and remove anyone who shares more than one grandparent with you.” I am puzzled as half-first cousins only share one grandparent, usually due to remarriage. Could you clarify that please?
Martin
Hi, Martin. Since half 1st cousins only share one grandparent you can definitely use them! If you would please point me towards the Feb 26th comment I made, I’ll be happy to fix it. Thanks!
I’m very confused. I did what you said on ancestry with 2nd and 3rd cousins for my mom, and I have 20 different colors and yes some overlap. My mom is 81 and has lots of matches. Do have any suggestions about what is going on?
Hi, Laurie. Please feel free to send me a screenshot of your mom’s chart and I’m happy to give some brief, but specific, feedback. If you do, please add a column showing the number of cM (centiMorgans) each match shares with your mom.
Hi, Laurie. First of all, how many matches does your mom have at the 2nd & 3rd cousin level? And, I’d be happy to take a look at her chart and give some feedback. You can email me a screenshot of what you’ve done. Please include a column with the number of centimorgans. Thanks!
Dana,
I am working on my matches to place into the Leeds spreadsheet, however, I have discovered that I have some half-1st cousins who are classified as 3rd-4th cousins on Ancestry. All of these are well below the 400 cM level. We do share a set of Grandparents and my family is really a mixed up mess.
What should I base my spreadsheet on?
Great question, Norman! If you know some people share a pair of grandparents with you, you should leave them out of your initial chart. You can add them back in later, but they would likely be in 2 (or more) color clusters.
But, I’m also a bit surprised at them showing up as 3rd-4th cousins. Half first cousins can have a wide range of cM, but I would expect them to show up in the 2nd-3rd cousin range. How many cM are they sharing with you?
Dana,
I have found that I have half-1st cousins who are in the lower 100 cM range and up. I have taken a step to sort all of them out of ones I will use for creating my Leeds Method spreadsheet.
I completed the chart as prescribed and I have 5 colored columns at the end of the day. One I can identify as paternal, two are maternal with one overlapping individual between the two columns, one has two people in it of which one is from the paternal line and one column has one person off on their own. I used my aunts results as mine has heavy endogamy from my fathers side of things. Would I then count the two columns with the single overlapping individual as potentially representing one great grandparent pair for my aunt?
Hi, Tracy. With the two columns, do you know if they represent only one of her great grandparents? If so, yes, you can combine them. But, if you’re not sure, just leave them as 5 groups for now and try to figure out what the groups actually represent.
Ok. Thank you. Leaving it as 5 groups right now as I confirm a few things first.
Hi Dana
I have 1 2nd, 8 3rd and 450 4th cousins.
My mum and dad are 2nd cousins.
I am trying to do a spreadsheet like you mention above.
Do I work with my common matches or all.
Thanks
Nana
Hi, Nana. That’s a tough one! I think you should still start with the steps as laid out and see what your results look like. And, feel free to email me a screenshot of your results if you need some quick feedback.
Hi Dana Thanks for you reply will work on it.
Hello Dana, this is an awesome tool. I am a bit confused though. I ended up with 6 cluster colors. Could this indicate a few half cousins?
Hi, Shadee. If someone were to get 4 groups, half cousins would just sort into 1 of those groups instead of 2. So, that’s not the likely answer. It is more likely that you don’t have enough 2nd cousins and that your matches are lower in this cM range of 400 to 90. Or, it could be that there is significant overlap and you haven’t consolidated them. If you’d like to email me an image of the chart, I’ll give you some quick feedback. If you’d like to do that, please include a column listing the # of cM of shared DNA for each match.
Hi Dana,
I am working with Ancestry DNA results and have used your great method for a friend ( I had actually experimented with something similar before I stumbled on your information).
He and his mother are from single child families, her mother had just one sibling. His father was born in and of Middle East descent so I expected his match list to be on the fairly small side. The results contain just 323 4th cousins or closer. For that reason I charted a bit farther down than just 2nd and 3rd cousins including several listed as 4th cousin down to a 49 cM level for a total of 31 matches. I found 3 distinct non over lapping groups. This does not include the 5 matches he has in a 4th separate column that have the strong Middle East ethic mix with those matches between 71 and 55 cM’s.
Two of the three groups are from known family lines on his mother’s side. Unfortunately at the moment his top 3 matches in the unknown group range from from 527 to 219 cM and have absolutely no Middle Eastern roots . Those high matches would indicate I am looking to find an unknown half uncle or half 1st cousin in his mother’s family.
I guess my question really is, if you get 4 distinct color groups with no over laps does that likely indicate 4 sets of of great grandparents? The two known groups I have charted to seem to be connected to the same set of GGParents without any overlaps, does that sound possible?
Thanks for listening,
Regards,
Vern
Hi, Vern. This is a complicated question! So, it is quite possible to see 3 groups on your mom’s side (or dad’s side). Often, this is because you don’t have 2nd cousins for that line so you’re working with 3rd cousins. So, 2 of those groups could represent the same grandparent line.
Next, you mentioned 3 matches from 527 to 219 cM on your mom’s side and said it indicates you are looking for an unknown half-uncle or half 1st cousin. There are quite a few other possible relationships for that highest match, and many possible relationships for the 219 cM match. You can look at the free tool on http://www.dnapainter.com – the Shared cM Tool – for all possible relationships. If you’ve already done this, maybe you have other information making yo believe it is one of these two relationships.
Your “big” question was about the 4 groups and does this likely indicate 4 sets of great grandparents. The answer is yes, especially if your matches include representatives in each group who are likely 2nd cousins – so people sharing around 200 cM or more with you instead of closer to 90 and the low 100s. If you have lower matches, who are more likely 3rd cousins, then it is more likely that your groups are representing sets of GGparents – probably the case you’re seeing.
Hope this helps!
Dana
I used a variation on your technique when I was locating my birth family. A new feature added to Ancestry since you published this made it possible. Since they added groups, you can do all the work right inside ancestry. Just add your first match under 400 to one color group, then add all their shared matches under 400 to that group. Then I usually mark the shared matches of one of those shared matches, until all the shared matches seem to be marked that color. Then give your next uncolored match a new color and repeat. All the conclusions you draw in your above synopsys work directly on Ancestry without the need for an external spreadsheet.
Hi, Peter. Yes, you can do this right in Ancestry. But, it only works well if all of your matches are only related to you in one way. I still recommend that most people do the Leeds Method first and THEN add the dots into Ancestry. After you get past the initial work of 2nd & 3rd cousins, you can add 4th cousins and higher directly into Ancestry.
When I colour block the second and subsequent columns, should I only colour 2nd and 3rd cousins or all matches, including 1st cousins or siblings?
The family has a brother and sister from one family married to a sister and brother of another family. Is this endogamy too? I imagine it is affecting the results.
Hi, Elizabeth. For the first time through, I would use only 2nd & 3rd cousins. After that, you can add in higher matches like 1st cousins and siblings.
As far as the siblings marrying, it depends if they are direct ancestors and how recent it was. Yes, it can affect the numbers. And, this is called pedigree collapse. Endogamy takes place over many generations.
Hi Dana,
Thank you very much for this method. It makes looking at matches much easier and I recommend it all the time whenever I see a forum post made by an adoptee. However, much though I try, I’m not able to make any sense of it when it comes to sorting through my Dad’s matches.
He has a generous 32 Ancestry DNA matches falling below 400 cM but still in the 2nd/3rd cousin range. These matches can be divided up into not 4, but 6 clusters, with a large amount of overlap between them. Only one group of only 3 matches doesn’t overlap with any of the others.
What’s more, the overlaps themselves are very confusing. My Dad has three known second cousin matches, for example, we’ll call them A, B and C. A and B are first cousins to each other, yet A falls into only one cluster, while B falls into two. B and C are sisters, but C falls into three clusters! I have no idea how to explain any of this.
Do you have any advice? Thanks,
Amelia
Hi, Amelia. You probably need to do the “check your work” step. Since the original method only looks at the first person in each cluster, it is possible that other people in the cluster actually belong to more than one group. And, if you develop heavy overlap – where a lot of people are in 2 different clusters – these can often be combined.
There are 3 ways you can do this:
1. Use Ancestry custom group “dots.” First, enter the colors as you have on your chart. Next, open each match, one at a time, and make sure you have all of the colors they belong to.
2. Just use your chart and open each match, one at a time, and see who they match. Add the individuals to additional clusters as necessary.
3. Use an automated tool such as AutoCluster at Genetic Affairs.
Hope this helps!
I did this with my second and third cousins. I actually had found many of these and when I did this method, I have six color columns, and it’s still missing a line of ancestors. Do I go into 4th cousins and continue? Also, what if some are definite tree connections, DNA connections, but not showing as shared DNA even though they are not distant.
Hi, Debbie. Six columns are fine and it is quite possible to still be missing a line of ancestors. So, yes, you can continue on to 4th cousins. Unfortunately, I don’t understand the last part of your question. Can you give a more specific example? Thanks!
Pingback: The Endogamy Files: What Is Endogamy? – The DNA Geek
Dana, I was SO hopeful when I saw how simple this looked, however, two hours and 39 columns left it seems my ancestors did some serious pedigree collapsing! If anyyyyyone is a pro at making sense of seriously scrambled colour clusters PLEASE let me know – I need all the help I can get!
Hi, Sarah. I am sorry you are evidently finding this frustrating. So, what range of cM (centiMorgans) did you use? I suggest starting with 400 down to 90 cM. At those ranges, I don’t think I’ve seen anyone with 39 columns. And, if you do have pedigree collapse, this method is not very helpful! I have it, but only on 1/4 of my family, so it works fine. If I were to work with my mom’s, though, her mom’s side would just create one big cluster. You are welcome to send me an image of your clusters – please include the # of cM for the matches – and I will provide brief feedback. Hope this helps!
Hello Dana! Feeling a little inept here and so badly want to understand my spreadsheet. I have A LOT of overlap and have 6 columns with my 2nd and 3rd cousins. Would you be able to take a look at it and tell me what I am seeing? Thank you!
Hi, Tonya. Most likely you are experiencing one of two things – either those clusters with a lot of overlap should be merged OR you have more 3rd cousins than 2nd cousins and they are just sorting into more groups. Either way, not everyone gets 4 clusters! And, your goal is to figure out what part of your family each cluster represents. I hope this helps!
Pingback: Ancestry Tells Genetic Affairs to "Cease and Desist" - Dana Leeds
Hello Dana,
I have read and re-read your instructions but my results don’t look right. On Ancestry, I have ten matches between 300-100 cM. I listed them all vertically on a sheet of paper. I went into each person’s shared matches as you said and the top nine names appeared in each others matches – so they were all assigned the same colour, leaving just one person (no.10) with no colour assigned. I went into her matches, and she matches with five people in the first column. So I only have two colums (two colours) and there is overlap on five of them.
In addition, I have a match which is 460 cM which I didn’t include in this, but he matches every one of the 10 people on my list. It’s an unknown father I am looking for.
Have I done my homework wrong? I know that Ancestry has an in-house version, but I wanted to do the paper and pen bit first – as I admit I am old school.
I’ve watched so many YouTube tutorials recently, I think my brain is fried and not working any more.
Any advice very much appreciated. Thanks
Hi, Karen. Yes, you’ve done your homework! And you’ve done it right. So, I would say all 11 people are in one color group. (The 10th person just didn’t happen to match the first person.) In this case, the one cluster/group could represent BOTH of your dad’s parents, or possibly just one of his parents and the other parent doesn’t have anyone who has tested that shared DNA with you above 100 cM.
Do any (I hope!) of these matches have trees? Or can you build trees out? If so, you’re looking for places where they intersect. If needed, you can add in lower matches, too – just look at the Shared Matches of your list and add other matches – especially those with large trees.
Do you know anything about this 460 match?
I also suggest that any person looking for a biological parent have their DNA at all of the big sites. You have to test at Ancestry and 23andMe, but you can transfer your results for free to MyHeritage and FTDNA, though you’ll need to pay to unlock some features. (About $20-$30). Hope this helps!
I’m planning on launching a new coaching service soon. If you get stuck, that might be an option!
Hope this helps,
Dana
Thanks Dana for the superquick reply! only 4 of the matches have family trees, with only about 5 people in each tree, no dates, and some first names only. One has 51 people in it, and I will ask if they will permit my access. The 460cM guy is really interesting, he is a Greek Cypriot and is kindly sending two DNA kits to his parents, to see if we can shed more light, he also has a grandmother still alive. Its going to take a couple of months to get the results. The other two 400+cMs people haven’t replied, nor from another site where I have a 450cM match, and that is disappointing, I’ll try again in a few weeks, not everyone is obsessed with this as me! I’ve uploaded my raw data from Ancestry to MyHeritage, FamilyTreeDNA, and GEDmatch. And I’ve used the Chromosome browser on FTDNA. My new 23andme test has just been posted off, as this is the only way to access their database, and I’ve chosen the health element too, as I thought I knew both parents until my DNA results showed differently! I have started to build a fake tree with bits of paper and sticky tape so that I can move things around. All my unknown father’s side matches are Greek Cypriot names (despite my Ancestry DNA saying 43% Italian) and I have joined a couple of Facebook groups which focus on Cypriot genealogy. I also joined Ancestry DNA matching and Genetic DNA for Dummies. Now I see that I have used your method correctly, I will try it again with the other sites’ matches, so thank you so much for being such a brilliant mind, and helping us all out. Kind regards
Karen
Hello again Dana, I forgot to say, that my unknown father is from a specific country overseas which has very identifiable surnames, so in my list I only included those matches with that country’s surnames.
Kind regards
Pingback: Clustering Without Downloading Data Using Shared Clustering - Dana Leeds
I have barely started color clustering and I already have 9 columns due to French Canadian ancestry. I have 2 matches between 400 and 90 cMs (161 & 94). How far down my list should I go?
With matches below 90 cM, you get less and less matching, so you would get more and more clusters. So, in that case I would probably just look at what questions you’re wanting answered. Are you trying to solve a particular family puzzle? Or, are you trying to identify matches?
I’m working with my FTDNA matches. As soon as my Ancestry DNA results are ready, I will probably have better matches. I am just trying to connect everyone to my tree. The color matching will help with my maternal side as I have a lot less matches with which to work and have a known match to her father’s side. My father’s side will be more difficult due to French Canadian ancestry.
Yes, you should have a lot better matches at Ancestry. But, it is great to work with your matches at all of the sites. And, yes, your dad’s French Canadian ancestry will be harder to work with, but it’s good you already know that! 🙂
Thankfully, I was a research analyst before I retired so I think this will be fun. I made color coded worksheets for years!
Enjoy then! 🙂
I’m trying to figure out a tangled mess. I found out that my Dad’s father is not who we thought, based on a YDNA done years ago. Only one match at 37 markers and that was 35 out of 7 but i ran with it. Certainly haven’t tied the 2 families together yet but based on that infomraiton I was able to track my Dad’s family (somwhat) thru Ancestry DNA. I did color colding and figured out which ones were from my Father’s paternal side. I’ve narrowed it down to 3 brothers but at a stoppoing point. Nobody around to do DNA and after researching the family I can undersatnd why. 🙂 But as I was doing Thrulines and looking I realized that I have no matches at all based the the family I have configured from DNA and Thrulines I realized I have NO matches with my 2nd great grandmother. Then I found a birth certificate that showed another woman as my great grandfather’s (John) mother. BUT…nothing comes up for her either. Is there a way to use the color coding method to distinguish between maternal and paternal matches on his great grandfather (John)in Ancestry or through your method?? I can trace his paternal side back to Germany based on DNA matches, and the maternal side back to my 2nd great grandfather. But none of his wives (there were 3) show any matches to him in Thrulines. Not even the one listed as his. So I don’t know whether there is a way to look at segments (considering no one responds, that would be tough) or breaking down the color coding even more. I realize it will be a lot of work and realyy don’t want to try it unless there is a chance. I might add, I have 1 DNA at 144cm but she is pretty non-responsive. Hope I haven’t been too confusing.
Hi, Sandy. First of all, I’m sorry that you found out your dad’s father is not biologically his father. That can be very tough to hear!
If you’ve narrowed down his father to 3 brothers, usually the only way to confirm is to have the actual brother or one of his descendants (or a sister’s descendants) test.
I wouldn’t be as concerned about ThruLines. If other people have wrong information on their trees, the ThruLines won’t necessarily show up. And, go ahead and put the names away and just work with your matches. If your highest match, though, is only 144 cM, it will be hard. Where else have you tested? If you aren’t in the 4 biggest databases, I’d do that. You have to pay to test at Ancestry and 23andMe, but can transfer to MyHeritage and FTDNA.
I hope this helps!
I have 13 columns..is that strange? Thanks!
Hi, K. Is this for matches from 400 to 90 cM? If so, are most of the matches under 200? And, yes, this is perfectly fine. If you have “heavy overlap” – meaning most of the people in one color are also in another color – then you can consolidate those into one cluster.
Pingback: Clustering at Genetic Affairs | Jennealogie
Your clustering method works very well and is simple to develop, a great tool. I am writing to discuss the inclusion of the 2nd and 3rd cousins. It is the perfect group to chart but wanted to speak to how those specific cousins are identified. There are other relatives that may be indicated as second cousins by the various testing companies but are not. Half first cousins are commonly indicated as a 2nd to 3rd cousin. This places the generational match in a different location, grandparent vs great grandparent. Assuming the half 1st cousin matches 2nd and 3rds then it does reinforce the relationship of one grandparent and perhaps two great grandparents. I am working with a person who has this situation, her fathers father is unknown and the grandmother was born out of wedlock with an unknown father also. Most of her 2nd cousin matches are really half first cousins with the grandmother as the connection to all.
Hi, David. I am actually in the midst of writing a blog post about this same situation – though I’m not sure when I’ll post it.
I set up the default as 400 to 90 cM as it works for most people. But, if you know how your matches are related to you, my suggestion is to avoid any match who shares more than one grandparent with you in that initial chart. Afterward, you can add in those 1st cousin matches who share more than one grandparent line.
Thanks for sharing! It is something that needs to be pointed out.
Dana, thank you so much for this! I’ve started working on it for my dad’s side of the family on MyHeritage, after getting what little I can from my matches on Ancestry. My mother is Ashkenazi, but her ancestors immigrated to the US in the late 1800’s so there’s a lot to work with (and a LOT of matches). My dad was born in northern Italy, and while I’ve managed to go back pretty far on his mother’s side, his dad’s mother was adopted on the “foundling wheel” in Verona and I’ve hit a complete dead end. (Also, records in his dad’s hometown are VERY limited online, so I can only go back one generation further anyway.) It’s a mystery!
I’m trying desperately to figure out which of my matches are part of his family, so hopefully MyHeritage gives me more Italian matches than Ancestry! I did finally convince my dad to take the test, so hopefully in a month or two I’ll be able to narrow things down using his test. I’m grateful I’ve found a method that also uses my obsessive need to organize things on Excel! :-p
Hi, Jill. Thanks for your kind words and I wish you the best. If you tested your dad at MyHeritage, I hope you’ll consider testing him at Ancestry, too. Although you might not have that many Italian matches at this time, things might be very different in 3-10 years!
Hi Dana –
Thank you for offering your expertise with this easy to use method of research. It’s awesome!
Okay, so what do you make of 12 color columns for 2nd-3rd cousins? Yikes! I used my mom’s DNA to cluster her paternal line only. I was easily able to decipher between my grandparents and remove my mom’s maternal cousin matches since my grandmother was full Finn/Saami. My grandfather’s clan was tied to three Native American tribes. Endogamy, for sure – but 12 color columns! How do I decipher this – plus patterns of surnames are slim-pickings. What does this type of result mean?
Please advise.
Thank you!
Thanks for your kind words! One possibility is, if you don’t have many 2nd cousins and are relying more on 3rd cousins, you can get a lot of clusters. Another possibility is that you have overlap and need to consolidate or combine some of your clusters.
Clustering doesn’t really work with endogamy. But, since you mentioned 3 tribes, those should separate from each other.
Hope this helps!
Dana
My mother’s parents were 7th cousins in two different families. Will this affect the outcome?
Hi, Allison. No, with the relationship being that far back it should not affect your Leeds Method chart.
Hi Dana,
Through dna matches, I met a 76 year old 3-4th cousin from 23&me who is adopted. I came across your method last week and have done that with his matches. Now what? At this point do I determine who the common ancestor is in each color group? If so, there should be 2-3 levels of separation between common ancestor and dna match? And the common ancestor is the adoptees great grandparent? What does it mean for a person to share two colors?
We are unsure of which side is maternal and paternal. Is there a way to just know that by looking at his matches??
His matches start at 15.7% female cousin on 23&me and 916cM on ancestry. He has 3 matches on ancestry showing 916-900 cM shared. When interpreting the dna, does it matter that he is half related to each side?
Both trees have been built and most people have been placed in the trees but I’m stuck. Do you have any recommendations on next steps?
I am so excited I came across your method! Wish I had seen it 6 months ago. 😊 Thank you in advance for help!
Hi, Melissa. Your new-found cousin has some great matches! Let’s see if I can answer some of your questions.
Yes, at this point you should try to determine a common ancestor (CA) for each color group. And, if you are using 2nd & 3rd cousins, the CA should be about a great grandparent or great, great grandparent level for most of the matches.
As far as a person sharing 2 colors – it really depends on the person’s clusters, but the 2 clusters are likely related meaning either on dad’s side or mom’s side. To answer the next question, you can’t tell mom’s or dad’s from looking, but from these people who match more than one cluster, you can tell the dad’s side from the mom’s side – just not which is which.
It does not matter that he is only half related to each side, except to say that if you are looking at possible half-siblings or half-first cousins you should keep that in mind.
As far as being stuck, I do offer a 1-hour consult where I can get online and look at the matches with you and help you analyze both the chart and your matches. You can contact me if you’d like more information.
Dana
Hi Dana,
Thank you and yes, I am interested in the hour and would be thrilled to have your assistance. It took me a bit to see your response here, will we set up details via email to be sure I don’t miss anything?
Melissa
I tried the cluster and I have 4 colors corresponding to both sets of grandparents, but I have some matches that do not fall into any of the columns. I cannot find the link between 8 dna matches (who match each other) and myself. The cM’s run from 105 to 286. I used the color dot system on Ancestry and put all of the shared dna matches that I have with these 8 people and none of them fit into either of my sets of grandparents. I am an NPE and suspect that all my matches with these people (419 so far) were created by an NPE situation but I am not sure how to go about searching to find out. I do hope that this isn’t too confusing. It is driving me crazy because I do not know how we connect. I do know that it is on my mother’s father’s line though because most of them match my half brother and there is one match that has a common ancestor with me and that ancestor is my grandfather’s line. Any suggestions on where to start would be greatly appreciated. Thank you so much.
Hi, Naomi. I hope I can help! First of all, can you clarify…do you have DNA matches that actually link back to each of your 4 grandparents? If you do, then why do you suspect you have an NPE situation? It is possible that these matches all connect through an NPE. But, I need more information.
Thank you for sharing your simple yet brilliant technique. It has added fun to my research, even though I’m staring at 6 columns and wondering what on earth fits where…lol
You’re welcome! And, with 6 columns your next step is probably one of two things. 1. You need to consolidate. In other words, if you have some a lot of people in one color cluster that are also in another then you should combine those. 2. If your matches are more 3rd cousins than 2nd cousins, these color clusters won’t form nice groups around your 4 grandparent lines. It is possible a couple of your groups are based on a great grandparent line.
Now, you start working with each cluster to see who the common ancestor is for that group so you can identify what the group means.
Enjoy!
I’ve been trying to cover more generations and it seems quite similar. I’ve used a color to represent where my cousins diverge from the paternal lines of my great grandfathers. Then 2nd, 3rd and so on.
As I’m extending the line, the matches get sorted, in what seems the same method.
Once sorted, a simple note can be used to free up the group markers for the next grouping.
Pingback: Fast Ways to Cluster Your DNA Matches at the Beginning of a Research Project – Family Locket
Using my Ancestry results I got a confusing mess. Part of that cleaned up when I eliminated my known 1C1R and 1C2R who were below 400 cM. But one cluster had total overlap with a second cluster, except for the cluster leaders, because the two cluster leaders match each other at less than 20cM despite being 2C1R to each other. Is there a better way to handle that?
Mark
Hi, Mark. You’ve brought up a great point. If you know one of your matches is a first cousin of some type who shares more than one of your grandparent lines with you, then you should leave them out of your initial chart. Otherwise, if they happen to be your highest match, they will “group” both of those lines into one cluster. So, you will get only one cluster when it would be nicer to have two.
I think you handled it find…just don’t include those 1C1R/1C2R if you already know how they’re related to you! You can add them back later.
Thanks for writing,
Dana
I was sent this link by a person doing genealogy. I’d like to try and determine parentage for myself and my half-sister. I know nothing at all of my father; she know only a name.
Can you explain how you determine which color block to put a person in? Are you using surnames?
Thanks for your time.
Terri
Hi, Terri. I developed this method to help someone identify their unknown biological parents. You are not sorting based on surnames. You are sorting based on shared matches – people who share DNA with both you and one of your matches. So, you don’t have to know how any two people are related!
You might try watching a video that was recorded of me at RootsTech: https://www.rootstech.org/video/you-can-do-dna?lang=eng
Also, I am planning on starting a workshop group – hopefully in October! If you’d like to join in, watch either my website or my Facebook group: “Leeds Genealogy”
Thank you for the quick reply. I will look at the video and see if that helps.
Hi Dana, I received information about your site from someone else. As I have been searching for my grandfather for 11 years, now very exhausted, I decided to give it another try. All I have is what was believed to be his name, but besides the fact that he was married to my grandmother and had several children, one of whom is my father, I come up with nothing but brick walls. At this point I’m not sure if that’s his real name. I do match a lot of people with his last name, but I’m not sure what to do after that. Can you please help me to get a new start on this? I would be so grateful. One more thing, does 6 cM and 5th and 6th Cousins mean anything? Thank you. I hope to hear from you soon.
Hi, Doris. If you’ve created a Leeds Method chart, feel free to email me and I’ll take a quick look at it. (drleeds@sbcglobal.net) I would go ahead and set aside the name for now and work with your matches and clusters. How many clusters were created when you used 2nd & 3rd cousins? Can you identify which part of your family each group belongs to? If not, make that your goal for now.
As far as 6 cM, those are very small matches and most have been deleted from Ancestry at this time. 5th and 6th cousins are also quite far back for the mystery you’re trying to solve. Hopefully you have closer matches!
Pingback: An Adoptee's Search Part Two: DNA Results Are In! *
I tried this method, using my three 3rd cousins and about half of my 4th cousins. I have no 2nd cousins tested. Virtually all the results can be traced back to my father’s mother, with only two people matching my father’s father and one person connecting to my maternal side. There was a handful of folk who remained unidentified. This was particularly frustrating because my great great grandfather on my father’s father’s side is unknown and my great grandmother a NPE. Has anyone else had a similar experience?
Hi, Diana. This could be the result of several things. First of all, it could be that some of your grandparent lines are related to other grandparent lines – whether fairly recently or more distantly. It could also be that some lines just haven’t tested. This sometimes happens with recent immigrant families, but it also happens with small families or just families who have not tested.
First of all, I suggest trying to determine what part of your family each cluster represents. Then, since the mystery is on your father’s father’s side, work with those matches and expand down into 4th cousins by looking at the shared matches.
Best wishes!
Hello there. I’ve fairly successfully done this chart for myself and it has been very helpful. Thank you! I was going to do this for both of my uncles’s matches, but as I look through them I don’t see true second or 3rd cousins. Only 1st cousins once or twice removed or great neice/nephews. Will this method still work doing it based on cM’s and not caring that I know this about these people?
Charlene
Hi, Charlene. If a match shares more than one grandparent with you, you are only dividing your matches into mom’s side and dad’s side instead of the 4 grandparent lines. You would have the same problem with nieces and nephews. So, you might skip all of those. If your uncles don’t have any true 2nd or 3rd cousins, they can still create clusters – they just won’t be close to representing their grandparent lines. But, they will still “cluster” around shared common ancestors and potentially everyone in a cluster is descended from that couple or from that part of the tree. Hope this helps!
Dana
I am trying to work out who my biological grandfather is as I have DNA links to another family at that point. I have colour coded all but am still confused?? Please help
Hi, Sandra. Feel free to email a screenshot of your Leeds Method chart to me at drleeds@sbcglobal.net and I’m happy to give some quick feedback. Please add a column showing the # of cM you share with your matches (if you haven’t already done that).
What do you think about combining your matches from all of the sites you have. I have Ancestry, 23andMe, FT DNA, MyHeritage, Living DNA, and DNAged.com. I would tage each by source, eliminate obvious dups, and then do the clustering on them as one group, and not separate. Is there something wrong with this concept?
Hi, Karen. Since clustering is based on who shares DNA with whom, it would be hard to do that across testing companies. It would be possible, however, to create 4 “clusters” for each of your 4 grandparents and mark those across the different testing companies.
Another issue – and the reason I primarily use Ancestry for all types of clustering and the Leeds Method – is a numbers game. If you are doing the original Leeds Method using the range of 400 down to 90 cM, this is how many matches I have at each of the companies:
Ancestry – 28
MyHeritage – 10
FTDNA – 6
23andMe – 2
LivindDNA – 0
So, besides MyHeritage and possibly FTDNA, clustering matches is just not very useful at the other sites.
But your results might be different!
Dana
Hi Dana
I have accounts in Ancestry, ftdna and myheritage. My parents are of mostly west asian endogamous ethnicity, but there were slips in conversations regarding me being mixed in the past, which is why I tested (without bringing them into it). I look less west asian than my parents and siblings (more European features) so wanted to see who those people might be to unveil a possible family secret.
My 4th cousins at ancestry (nothing closer) are of mostly the same west asian ethnicity and some are mixed (only 514 matches in total including 5th-8th cousins) and only 1 in 4 is of west asian ethnicity starting in the 5th-8th category. Now, ftdna shows only 11% of total matches to be of west asians of broad regional ethnicities (328 matches in total there), while myheritage shows only 3% west asian broad regional matches out of the total of more than 700 matches there.
So I figured most of my blood relatives must live in Europe because most of my matches appear European. Since ethnicity estimates gave me different results in the three companies, I decided to piece out my percentages using myheritage autoclustering.
I do not have anyone closer than 3rd cousins on ftdna and myheritage. My autoclustering gives me 13 different clusters. My top match sharing 78 cM (from Europe) is not on the list of clusters at all (was a stranger to me). The match revealed that they are 1/4 of the same endogamous ethnicity as my parents, though 3/4 Scandinavian. The second largest match is on one cluster sharing 69 cM (same west asian ethnicity as my parents). Another cluster has west asians but they are of a different ethnicity.
The rest of the clusters (11 of them) appear to be of European descent (5 Finnish clusters, 3 Slavic, 1 Lithuanian, the rest are Germanic/Austrian/Scandinavian).
The only thing, when I lowered the cM threshold to 20 cM minimum most clusters had at most two people in them.
Now, it appears these clusters are not overlapping. There is no one over 30 cM in the list except the endogamous top 69 cM match.
Are these clusters different gg’s, ggg’s, gggg’s or even beyond? I know you said 3rd-4th cousins and beyond is frequently inaccurate for clustering, but these European ethnicities are not really endogamous as far as I can tell.
Any thoughts would be greatly appreciated!
Hi, Robert. I’m glad you wrote! So, first of all, ethnicity estimates are just that – estimates. They are great to the continental level, but not as good at the country level. But, you are most likely seeing more European than Asian matches because more Europeans or people with European ancestry have tested. So, it’s really a question of which matches have tested.
As far as your AutoClustering – especially on MyHeritage. Most of us just don’t have as many matches on MyHeritage and the clusters end up being small and more distant.
No matter how many people are in a cluster or how closely they are related to you, the procedure remains the same. The goal is to figure out how the people in that cluster are related to each other and then how you are related to that cluster. However, it gets more difficult as the matches get more distant.
Hope that helps!
Dana
Thank you for your thoughts Dana.
As far as west asians not being tested enough on myheritage, I ruled it out for the endogamous ethnicity of my parents. The 69 cM endogamous match (100% this same ethnicity which my parents have) revealed to me that about 98% of his matches on myheritage are not only west asians in general but of this same endogamous ethnicity. And he is from the same region as my parents are. So my theory is that my west asian consists of this endogamous ethnicity together with others – those for which there are not enough people that have yet tested at myheritage). I also know that since my parents’ ethnicity is as endogamous as Ashkenazi Jews are, and at least a quarter of them live in the United States, I should have had way more of these matches at ancestry (I only have 21 fourth cousins, and nothing closer). And half of them appear mixed. One of them is from a neighboring non-endogamous population there, so that raised my suspicions and eyebrows too. This does not help figure out percentages (no autoclustering unfortunately either for ancestry!), but definitely supports the idea that my other west asian ethnicity is not really represented in the u.s. and definitely not tested at Ancestry as a result. Only 514 matches at Ancestry and 701 matches at myheritage is interesting to me. My parents are not eager to test (nor my siblings), so this is my personal journey. However, my dad encouraged me to test myself when I described how useful it would be.
Totally agree that the clustering is just a hint for distant matches. I guess I will have to wait for closer matches everywhere! Thanks for your input.
Pingback: Ancestry's Changes Affect Those with Enslaved Ancestors: A Guest Post - Dana Leeds
I hoped that your method might work for me, since I did have six 2nd-3rd cousins. (I eliminated my one 1st cousin, and my other 507cM 2nd-3rd cousin).
However, 5 of the 6 are all from the same set of g-grandparents (on my paternal side). The sixth cousin is almost definitely from my maternal grandfather’s side.
Is it ever possible to use 4th-6th cousins to try to pull out the other sets of grandparents?
Loralee
Hi, Loralee. First of all, you can use the 507 cM cousin if they are a known 2nd or 3rd cousin. And, yes, you can go ahead and use those lower matches! Just be aware that you will not be seeing clusters based on your 4 grandparent lines. They could be based on great, great grandparent lines or even more distant. But they can still be quite valuable! After creating the clusters, your next step is to determine what each cluster means. Hope this helps!
Pingback: Attempting The Leeds Method For Sorting Unknown DNA Matches – Lee Genealogy
Hi Dana, My Dad aged 89 youngest of 5 very close siblings who never knew any of their Dad’s family. My GF never talked about his family. Dad’s last surviving. Have his DNA across all platforms including Big Y. All that was known and believed was from Marriage cert. Dec 1917 Wagga Wagga NSW he was then aged 28, father with same name & Mother Kate Moloney farmers and both deceased. He died 23/8/1965 in hospital metro Melb Vic, AUS believed to be day before 76th BD. In 70’s my mum (died aged 79, Alzheimers 25/1/20) did a lot of work trying to fond his ancestry. Was believed he had a brother, and some tragedy befell the family related to loss of parents young. I’ve found a small very old envelope I think likely given to mum in 2010 at a family reunion in country Vic relating tp Dad’s mum’s paternal line, Dad hadn’t seen this. It had on it, from his much loved sister it was for him, “when I’m gone.” And on photo inside, “This is Dad’s brother William Gray, so he always said.” Man in photo looks nothing like my Da. And more recent find were small bunch of photos including a few very good quality, I thought Da and this supposed brother in Vic country town Wangaratta @ recognisable studio, I’ve had this dated and bingo, Dec 1917 re wedding photo seems correct. This photo special. The other man could be person in first pic. Amongst those photos 3 of a woman & on her younger and a man. We can see likeness between the man and supposed bro. No likeness at all between them & Da. The women has a lovely smile but is overweight and is hard to see resemblance to either Da or bro, except for Da had a relatively low hairline and thick curly and lol in pre wedding pic looks blow waved! hair and if this woman was his mum, he got her hair. Both bro and apparent Dad are losing theirs. Well, from DNA match’s have given up on names of parents as have so few match’s with those surnames 7 in trees, absence is conspicuous. Helpfully for separating Dad’s Paternal & maternal matches, is his mum’s mum’s family came from Prussia. His mum’s Dad’s Dad’s paternal line to Wales, and maternal English. And his Dad’s mum Irish. And Scottish locations identified @ FTDNA were a surprise, but even bigger after recent Ancestry ethnicity update separated Irish & Scottish and shows Dad with 39% Scottish! Totally unexpected. But SNAP in relation to the common ancestors I see amongst his match’s. Also 16% Wales more than the English and German. And lastly first evidence I’ve been able to find of Da’s existence apart from Dad’s knowledge of Galloping Race Horse trainer Da first worked for (and have very fragile reference from him where date is missing); is on 1914-17 Census list’s show him living in assume a room in building still standing Dad has identified as very close to old trotting track, aged 24-28 occupation (for rest of his life) Horse Trainer – trained & rode in Harness Racing. Right, so apart from revelations above, still up against solid brick wall. And have nearly finished first leads chart using Ancestry match’s. Dad has more match;’s than any of us! But highest c> is a match on MH 160cM estimated 1st 1 x removed – 2 1 x removed unresponsive. And on Ancestry, different line 131 cM 3-4. On Ancestry six 3-4 cousins 131 cM to 92 cM. And six 4-6 cousins 69 cM to 53 cM. Well that top Ancestry match has no tree and manager unresponsive, first column perfectly clean, but distinct absence of trees :/. 2nd column much longer as lovely lady at top shares lots of apparently valuable mutual matches & was so excited to get her response! Almost everyone with a tree falls from a couple from Scottish Isles Argyllshire who landed in Warrnambool coastal Vic town with their first of eight children. And a few fall from another couple with same surname who had landed in USA mainly North Carolina, or another appears landed in same suburb Dad was born in in city of Melb. I have other surnames around the pair who landed in Warrnambool, that also relate to the others with same surnames of each. And am very grateful to a match who has some of these in his tree who is great at communication but we are still stuck working out our relationship too. So got a bit off track there. My question relates to having added people to the chart from mutual match list with the person at top of second column. Which resulted in two clean columns. So 10 in 1st. 17 in 2nd. And 5 in 3rd … with 2 from 3rd also matching 2nd. And then 4th starts with person 54 cM added – and felt need to colour a block upwards matching him to her person below very top match in first column. and i’ve already added another two 4-6 cousins to 4th column, that don’t match anyone else in 1, 2 or 3. well, Dad has a lot of 4-6 cousins. Surname coming up the most is that of female in the common couple & comes up on strongest matches. Stands out. Oh yeah, two biggest at MH Also fall from them. And not so much in stronger matches the male surname. But @ FTDNA most of Dads 111 match’s have nominated that as their furtherest back surname, and he’s been put in a group of that name. I’ve been able to find names in trees of more distant match’s given to me by a lovely lady, daughter of one of Dads 111 matches now deceased who was adopted, where she discovered his parentage after his death & has his ancestral line back a long way as was well known family. And I’m finding smaller match’s via sister of male in the common couple. And very recently stronger has come up on MH and I see two more middlish strength related to that match, one her daughter, on Ancestry. I must wrap this up. I’ll continue, but would appreciate your feedback re the cross column connections I’ve noticed so far. I’ve seen going into 4-6 not recommended. But while I see Dad’s as valuable as I can find the bunch of connected names all over the place. Feel like I’m swimming in them but not getting anywhere.
Hi, Kerryn. I’m happy to give some quick feedback, but there’s a lot of information here and it would take a long time. If you’d like to email me an image of your Leeds Method chart – and include the # of cM for the matches – I’m happy to give some quick feedback. But, for this initial chart please just use cousins who share 90 to 400 cM. Thanks!
oops, correction in above (massive -sorry :/) message. The 2 cross column match’s are between 2 ppl from 4th to 1st columns, not to 2nd. 2nd & 3rd very clean. Yet to identify any common surnames in columns 1 or 4, and less trees there must say. Going to review all tree lists after I get to end of putting names on Leeds chart and hoping to make some progress on columns 1 & 4 too.
What does it mean if you only get three columns when using this method? I am helping someone find their birth father. Two columns clearly relate to her mother’s side and then their is only one column/color representing her father’s side. I used 2nd and 3rd cousins.
Hi, Carla. It can mean several different things. You might have a 1st cousin in that group and so it is really two clusters that are combined. Or, you might not have anyone who has tested on that 4th grandparent line – so you might need to go into those 4th cousins. Hope this helps!
Wonderful goods from you, man. I have be mindful your stuff previous to and you’re simply too excellent.
I actually like what you’ve obtained here, really like what you’re stating and the best
way by which you assert it. You’re making it entertaining and you continue to care for to keep it smart.
I can not wait to read far more from you. That is really a terrific
website.
Thank you!
Hi Dana
I’m just getting my head around this! I leave off 1st Cousins when listing the initial matches with me. When I then look at the matches to those on that list, if there is a match to a 1st Cousin do I include them? I think not, but I’m not sure.
Hi, Bob. When you create your initial chart, you do not list your 1st cousins. But, you do use everyone in that list of 2nd and 3rd cousins – even if they match one of your first cousins. When you’re done with that initial report, you can add back in those 1st cousins. They will likely match 2 or more of your clusters.
Hope that helps!
Dana
Hi Dana,
I have used your method with many relatives and clients on Ancestry with no problems at all, it is fantastic! I have just now used it for the first time with a cousin who is only on 23&me and came up with 18, yes EIGHTEEN, different color groups!! I have only used her 2 and 3 cousin predictions. 23&Me gives you the shared DNA percentage not shared cms, so I don’t know if that is the problem.
I’m adopted into the family but I’m the one who is obsessed with genealogy lol.
Hi, Kim. Thank you for your kind words! And, with this cousin, you might find the post I just wrote to be helpful: https://www.danaleeds.com/leeds-method-time-to-consolidate/
Another possibility is that the cousin doesn’t have many (or any) 2nd cousins and the matches. We need 2nd cousins to group our matches together to get something close to 4 clusters.
Hope this helps!
Dana
Dana
Just read your new post. And this is what I figured I needed to do. She has a lot of clusters where there is only 1 person who only has only this new color, the other shared matches this one color person has matches with others that would now have more than one color to them. (hope this makes sense, I retyped this many times lol)
She does have low 2nd cousin numbers. She has 9 predicted 2nd cousins and 60 predicted 3rd cousins.
Matches with # of colors:
1 color = 33
2 colors = 16
3 colors = 14
4 colors = 4
5 colors = 1
6 colors = 1
My other question has to do with Ashkenazi Jewish. She and 2 other first cousins have Ashkenazi Jewish, 1.6%, 1.3%, and 2.8%. This is all news to us as we thought our Grandparents were Polish Catholic. How far back would a Jewish ancestor be with those percentages? And could that be messing up anything.
Thank you so much for responding!
Kim
I’m also having the difficult time of determining which colors to combine, there are odd multiple overlaps. Ugh, lol
Kim
Pingback: Leeds Method: Why Do I Have More Than 4 Clusters? Or, Time to Consolidate! - Dana Leeds
I’m not sure if you are still monitoring this or not but I have a question. I was wondering if you could help me. I have six columns in total. I have three columns of colors with almost all are matching except a few random people. The three random columns are as follows:
Column 1: person 5, person 6, person 9, person 11, person 13, person 15, and person 16.
Column 2: person 9, person 12, person 13, person 15, and person 16.
Column 3: person 6, person 9, person 11, person 14, person 15, and person 16.
Each person individually seems to lead back to the same surname how come they all don’t match each other then. Sorry very new to this.
Hi, Dawn. Great question! It sounds like these three columns have what I call heavy overlap and I would combine them. I actually just wrote a post about this yesterday! (https://www.danaleeds.com/leeds-method-time-to-consolidate/) I hope you find it helpful! Let me know if that doesn’t answer your question.
Hi Dana –
When you say “use 2nd-3rd cousins”, I assume you are not referring to people whose relationship to us is know, but rather people who fall within a certain cM range, and whose relationship is inferred but unknown. Is that correct?
Will this method work with matches from other sources, such as GEDmatch, 23andMe, FTDNA, etc?
Hi, Pierre. Yes, I am referring to the categories as they are listed on Ancestry. BUT, if you already know the relationships, you would want to choose 2nd & 3rd cousins. (I’m coming up with a new blog post soon.) Most importantly, you want to avoid anyone who shares more than one grandparent with you.
And, yes, it will work on those sites. However, many people who have plenty of matches on Ancestry do not have enough to use this method on some of the other sites. Also, many people just don’t have enough information or trees on those other sites to be able to work very well with identifying the people in their clusters and the meaning of their clusters.
Hope this helps!
Dana, I would like to sincerely thank you for developing this method and sharing it with us all. It was the breakthrough I needed to figure out who my dad’s birth father was. I will forever be grateful to you and can’t even begin to express my appreciation in words! <3
Amy, Thank you so much! You really made my day and I really appreciate it. I guess I shared your comment with my daughter and forgot to reply here, but please know how much I appreciated your response. And, congrats on finding your dad’s birth father’s identity!
I am working on figuring out who my biological daughter’s birth father is. I know I should know this but hey I can’t remember I was such an adventurous teen. I have used the method but still at a loss. I get close but no boys. gonna try again. I think I am ready to retry. How far down the line do you go to the end of matches or just to 90cm
For the initial chart, just go down to 90 cM. But, if you have very few matches, you can go a bit farther. Or if you have plenty of matches, you can add lower matches after you create your initial chart.
Hi Dana,
I am very new to all of this and I wonder what on earth I could be doing wrong since I get stuck right away trying the Leeds method. On Ancestry I have 10 matches between 90 and 400cM (4 of which are listed as 1st-2nd cousin). I have listed the 10 persons in Excel and color coded person 1, and then looked at his shared matches and given those the same color in the column.
Now comes the strange thing, person 1 matches all of the other 9 listed persons which means I never get to the step where I am now supposed to choose the next of my matches who didn´t get a color in step 1, and give that person another color code.
What am I doing wrong? Should I not have included the 4 persons listed as 1st-2nd cousins, even though they are in the range of 90-400cM?
Hi, Monica. That’s a great question! First of all, do you know these top 4 matches and how they’re related to you? If so, are they related to you through one grandparent (some type of 2nd cousin) or through two grandparents (some type of 1st cousin). If any of them are 1st cousins, don’t use them.
If you don’t know their identities then see what happens if you look at some of the other matches. If you look at the 5th person, or the 8th person, are they still related to everyone? If so, then it sounds like you just have one cluster. That could mean it represents either your mom’s side or your dad’s side. Or it could represent one grandparent line. The other part (or parts) of your family most likely have simply not tested at this 2nd or 3rd cousin level. This can happen with small families, families from countries that haven’t widely tested, or just families who happen to not have tested yet!
You can also try working with some of your 4th cousins and seeing if you get other clusters.
Hope this helps!
Dana
Hi Dana,
I know 3 of the top matches (below 400cM) and they are only related to me through one grandparent. The 4th of the top matches I don´t know/recognize, and he doesn´t have a tree either.
Yes, when I look at person 5 and 8 they are also related to everyone. So it looks like you say that I only have one cluster. In this case I know it´s on my mom´s side. This is unfortunate since I am trying to find my (unknown) birth father (the main reason I have tested).
I am Scandinavian and I think less people have tested here compared to the US and some other countries.
Apart from Ancestry I have also tested with FTDNA and I do have some unknown matches there on 2nd-3rd cousin levels (and below that). Can I use the Leeds method with FTDNA as well?
You suggest that I could try working with some of my 4th cousins, how many do you suggest?
Thanks for your help so far!
Hi Monica,
I just saw your post from a long while back while reading Dana’s comments. I have Scandinavian roots… mostly Swedish. I found my links to family there via MyHeritage. It seems to be more commonly used there than Ancestry.
Good luck!
Jill
Hi Dana – I have had the luxury of having my DNA tested as well as my parents, my sister and my daughter. I am interested in helping a DNA match on my Dad’s side. He shares 80cm over 5 segments with my Dad, and we do know that the match had a grandfather who was adopted in 1888, his grandmother’s name is known, and we do not match through her, so looking for the biological father. We are able to narrow down to a couple of family lines that merge, but not the generation. This man is my Dad’s top “4th cousin match” – though I actually believe he is a third cousin. How does doing the Leeds method assist me here?
Hi, Amanda. Great question! Since you are working with a specific match and trying to figure out how your dad is related to him, I would just look at the shared matches between your dad and this match. If possible, get in contact with some of these matches and see how many cM of DNA they share with your match. That will give you a better idea of how the match fits into the family.
Well, I can’t make head nor tail of the method. I’m getting on a bit, maybe that’s the reason!
I’m sorry to hear that! If you’re more of a visual learner, you might try this video – my part starts at about 40 minutes into the video: https://www.rootstech.org/video/you-can-do-dna?lang=eng
Dana
Hi Dana, I am new to this and a bit confused. Someone told me that if you have 2 colours where they overlap, that you should combine the columns. How do you do that if not everybody matches the other? I have 6 columns and I don’t know how to consolidate them. Any help?
Hi, Julie. Try this post – https://www.danaleeds.com/leeds-method-time-to-consolidate/ – and this post – https://www.danaleeds.com/consolidating-clusters-example/
Hope that helps!
Dana
So Dana, I’ve charted 90, 2nd C-4th C. Unfortunately my largest cM is 109 and my smallest is 48 cM. I have color coded the information of the top 14 individuals. The cM’s ranged from 109-68. With those 14 individuals I’ve gone out to 12 color groups the largest of which were 10, 9, 8, 8 members in the group.
Should I assume those four groups are the grandparents or is the information so skewed that no conclusion can be drawn because of the limited size of the cM’s?
Hi, Steve. That’s great that you have four good groups. But, I wouldn’t conclude that they represent your 4 grandparents. Instead, now you can work with each group and figure out how they’re related to you. Of course this is more challenging if you are seeking biological parents or grandparents, but I developed this method to solve an adoption case and it works!
Best wishes,
Dana
Hello, I first want to thank you for this method! Now my question–I am trying to figure out some mystery DNA matches. It started with a 2nd cousin level match who has no known connection to my tree. I have multiple other matches who also match with this assumed 2nd cousin, but still have not discovered any connection. I HAVE been able to connect many of these matches to each other in a dummy tree, but my family does not intersect with theirs at all.
I have done a chart with my 2nd and 3rd cousins. I have 5 columns/colors and am able to definitively trace 4 of the colors to my known ancestors (two for mom’s side, two for dad’s side). One column contains all the mystery matches that I have no connection with–and these colors do not overlap with any others (in my chart) HOWEVER it is clear from looking at their shared matches that they are from my dad’s side.
Because the top level mystery match is at the 2nd cousin level, I started working with the theory that my paternal grandmother’s parents are not actually who is listed as my great-grandparents–but actually her birth mother is who she was told was her older sister, and her birth father is unknown. She apparently mentioned to family at one point that she believed her sister was really her mother, but it was never formally acknowledged.
Is there anything I can do to try to narrow down or confirm this? All the people involved are long gone, the only family member I have above my generation is my mom, all other aunts/uncles and above are dead.
Incidentally, when I first did the chart, I included my top match, who is my first cousin once removed on my dad’s side. When I included her, her group contained all but one of the mystery matches, with many overlaps. When I removed her from the chart and re-did it, the colors are much clearly defined, with only one overlap (on mom’s side)
Hi, Jennifer. Do you have someone in that mystery match’s family that would likely be in the right time and place to be your grandmother’s father? Also, you should work on trying to see if you have any matches to who you were told was her father. If you aren’t seeing any connections, this could add to your theory.
That’s an interesting question if the birth mother is actually the older sister. I’m thinking it through and think I’ve figured it out. She would likely still have all of the same close matches, but she would have less than the amounts of DNA expected for some relatives. For example, her great grandfather would actually be her great, great grandfather. So those she expected to be her 2nd cousins would actually be 3rd cousins. So these “2nd cousins” would show up with lower amounts of DNA.
Hope this helps!
Dana
Thank you for the reply. Many of the mystery matches are clustered relatively close to the area where my grandmother’s family previously lived–the mystery matches are all on the US side of the Great Lakes and my grandmother’s family on the Canadian side in Ontario. My grandmother was born in 1904 in Manitoba (allegedly) although her birth was never registered (unlike all her siblings, who’s births were registered immediately). My grandma registered her own birth at the age of 60. My grandmother’s family does appear in Manitoba in the 1901 Canadian census however, so it does appear that they were living in Manitoba at the time.
All the DNA matches I have on my dad’s side trace directly up the paternal side so far (nothing that might branch out to include my paternal grandmother’s line).
It’s interesting that you say that about the amount of DNA being less than expected. My top DNA is my first cousin once removed at 391cM (didn’t include her in the chart). My next highest matches are two known second cousins (mom’s side) at 330cM and 309cM. My next match and highest on my dad’s side is a known second cousin at 287cM, then comes the first mystery match at 272cM. However, I also have a known second cousin at 110cM (dad’s side) which is only an 8% probability according to DNA painter.
I’m now wondering if it’s possible to “triangulate” my known matches on my dad’s side to see if there might be clues there?
Hi I am trying to figure out why I can’t find 4 lines only 3. I assume I know what that means but I am not sure I am willing to say that yet. ugh this is so confusing.
Hi, Alex. It is most likely a simple, not disturbing, answer. In most cases it means one of two things:1) you only have descendants of 3 grandparent lines that have tested at this 2nd and 3rd cousin level or 2) two of your clusters are combined so one of them represents your maternal or paternal ancestry instead of a grandparent line.
Hope this helps!
Dana
I am really struggling with identifying which line of my family some of my matches are on; are they paternal or maternal? My dad is the only child of his parents, but he has 1/2 brothers and sisters from each parent.
I have attempted the Leeds Method, 25 people between 340 to 92 cMs. I have 8 colors and 1 overlap. How do I figure out which color belongs to which grandparent/great grandparent? I know the person with 340 – we share the same grandmother, different grandfathers. I have another match, 231, her grandfather is my Uncle on my Mom’s side, but she doesn’t overlap with anyone…and then I have a match at 137 that overlaps with 1 other person, and I don’t know any of these people.
I am trying to figure out my paternal grandfather’s line … and I have no idea what I am doing …
Any guidance is most appreciated…
Hi, Cindy. If you haven’t already, read this post: https://www.danaleeds.com/consolidating-clusters-example/
It is likely you can combine clusters. If not, then it is probably because your matches have lower cM (so closer to 90 than to 400) so they just aren’t sharing overlapping DNA segments. That makes it a bit harder, but it’s still the same process. The goal is to find out how the people in a cluster are related to each other and then how you’re related to them.
Since your main goal is to identify your paternal grandfather’s line, you can most likely eliminate anyone sharing DNA with anyone on your mom’s side. Also eliminate those sharing DNA with your dad’s mom’s line. Then, you might have several groups left – try to work on those. If you don’t have enough matches to really make a group to work with, try adding in those lower matches. (Also the higher ones if you haven’t already!)
Hope this helps,
Dana
Hi Dana!!
I have five colors. I am not sure how that happened. I tried to search previous comments before posting but I didn’t see any with the same issue. Also, I am not sure how to tie this information all together. I am trying to help a cousin locate his bio family. He was adopted at birth. I felt like I was on the right track and now I’m just not sure anymore.
Thanks!
Nikki
Hi, Nikki. First, try this post: https://www.danaleeds.com/consolidating-clusters-example/
Second, when you are working with a single cluster, the goal is to see how the people in that cluster are related to each other. By doing that, you then hope to figure out how the test taker fits into that group. For example, let’s say you have a cluster with 4 matches all going back to a particular couple. In that case, that couple is also likely an ancestral couple of your test taker. (Or at least they are likely closely related.)
Hope that helps!
Dana
Dana, thanks for your tutorial and this great tool! I am trying to determine my stepdad’s birth family – he was adopted at birth. He has about 80 matches in the 90-400 cM range. I ended up with not four but FIVE clear columns/colors/groups. Any thoughts on what this could mean? There is very little overlap in the chart – only two people as best I can tell. The fifth group doesn’t start until a match with 171 cM shared. I can give you more info if it would help you, but at this point I’m not sure what more you would need to know. Any insight would be appreciated!
Hi, Justin. Sorry for the delay. It sounds like you have great matches! I wouldn’t worry about the fact that you have 5 instead of 4 clusters. I would just work with them “as is.” You might find this article helpful: https://www.danaleeds.com/leeds-method-a-lot-of-clusters/
Dana
Hi,
I’ve only used people who match between 90 and 400cm.
I have no over lapping but 7 groupings rather then 4 does that mean it’s gone back a further generation? :/ thanks
Hi Lyndal. It is more likely that you don’t have enough 2nd cousins to pull those matches together. You might try this post: https://www.danaleeds.com/leeds-method-a-lot-of-clusters/
Dana
Hello Dana,
Apologies if this has been asked already, but will this method work for 2nd cousins that are once removed and match with one of the 16-Great-Great Grands instead?
When I separate my ancestry.com matches to show my 2nd-3rd cousins (90cm – 400cm) only, I get 11 matches. Half of the matches, the connected relationship is currently unknown but when I review those matches trees and look at ages, theirs a high probability there’s a generational difference.
The other half of those matches, are confirmed relatives (my shared CM with them ranges between 90-209). The connection from my end ties to one of the 16 Great-Great-Grands and tracing back from their end it’s one of their 8 Great-Grands.
A lot of my grandparent’s siblings either didn’t have children or the ones that did, I know them and they have no interest in DNA testing. So with my scenario, I’m wondering if I could still get the benefits of this method using only 2 cousins 1x removed instead.
Thank you.
Hi, Brittany. Yes, this method would work fine with a 2C1R. Basically you just want to AVOID any matches who share more than one grandparent with you for the INITIAL chart. As far as the generational difference, we often see this with either “older” or “younger” test-takers. But, yes it should still be beneficial! Hopefully you can identify some of those unknown matches and then use your knowledge to start working with those lower matches.
Best wishes!
Dana
Thank you Dana!
I did have a follow-up question, if you don’t mind. I completed clustering my 11 matches, and ended up with 5 separate groups with no overlaps.
Am I getting 5 groups because of the generational differences?
Thank you.
Brittany
Pingback: How to Create Genetic Clusters Manually – Family Locket
This is a great resource, thank You. I have done BigY test with FTDNA and mtDNA test and Family finder test with them. My question is should I use this method for each of the tests separately? Right now I am primarily focused on the BigY test and would really like to understand where my surname “George” came from. I don’t match any of the main George lineages or men on BigY, other than my Grandfather, most of my matches are coming from various Scottish lines.
Jason, It’s been my pleasure creating this resource and seeing how helpful it has been to so many!
So, the Leeds Method was designed to work with autosomal DNA matches – not Y or mitochondrial.
With your Y-DNA test, if you go down to 37 markers do you see any men with the George surname?
Autosomal DNA is more my specialty than Y. I would suggest testing at AncestryDNA. They have a MUCH larger database and more trees. You could then search your matches for the George surname and see how many generations back you can prove. You might also test at 23andMe, the 2nd largest database. And, even with the results from FTDNA you can upload your data for FREE to MyHeritage.
Best wishes!
Dana
Pingback: RLP DNA e-course: Assess Your DNA Matches and Analyze Your Pedigree – Family Locket
Pingback: RLP DNA e-course Part 8: Following Your Plan, Research Logging, and Writing As You Go – Family Locket
I’m a bit confused. I worked my way through but ended up with basically a 5th column. After doing my first 4 columns I had 3 names left (who all match each other as well as match me) but don’t fit into the other 4 columns.
What does this mean?
Hi, Rachel. 3 of the last 4 posts I wrote deal with this topic. Hopefully you’ll find your answer there! If not, please ask again.
Hi Dana,
I got six columns no overlap…
My tree is already well researched so I know that two of those columns relate to half second cousins, so to two different great grandparents. Additionally, two columns relate to 2G-grandparents and one to a single 3G-grandparent, (another half cousin relationship). One great grandparent is totally unrepresented.
My point being to reinforce your caution that drawing any conclusions from a Leeds chart alone should only be done with further evidence.
regards
Gavan
Excellent point, Gavan!
I think I am confused by this. I took my 2nd-3rd cousins and went down from 313 to sharing 90cm. In this chart I only had two columns and only 2 of those people overlapped. Not sure what we do next or if I am supposed to have more than that.
Yenny, The number of columns (or clusters) anyone gets are dependent on 2 factors: the composition of your family AND who has tested. You very likely could have only two clusters! It might be that only 2 of your 4 grandparent lines have tested at that 2nd/3rd cousin level. Or it could be that you have included some 1st cousins. Or it could be that your grandparents are related to each other – even a marriage several generations back could cause this or something much more complicated like endogamy.
The next step is to work with each cluster and see how the people in a cluster match each other and then how they match you. I do this by diagramming: https://www.danaleeds.com/visualizing_clusters_2nd_3rd_cousins/
Hope this helps!
Dana
I am adopted and did the cousin chart using 2nd and 3rd cousins 400 to 90 cM. I came up with 9 colors over 25 people. Very confused at this. Also did 23andme came up with 4 lines with 2 stand alone 16 people. MyHeritage had only 3 matches that met criteria in 2 lines and family tree dna had no matches meeting criteria. Anyone have any clues where to go from here?
Hi! You might find these posts helpful: https://www.danaleeds.com/leeds-method-a-lot-of-clusters/ and https://www.danaleeds.com/leeds-method-time-to-consolidate/ and https://www.danaleeds.com/consolidating-clusters-example/
Dana
Hi. I’ve done this using the color dots on Ancestry, then expanded further up the line to find common ancestors where I’m stuck.
I have a great grand whose parents were 2nd cousins, and for whom 2 grandparents were 1st cousins. Additional intermarriage on that line really confuses the issue, plus remarriages. I may be descended from a couple, but related to 2 different people each by only one side of that couple, for instance.
Still, it’s a valid concept and generally very helpful.
Hi Dana and thank you for sharing how to sort them into family groups. I do have a couple of questions please which I hope you can help with.
I have just received Ancestry DNA results for my dad and had ‘narrowed’ my search to between 90 and 400 as you recommended, however when I went back to look at them, I found that 3 (2 from 1 group and 1 from another) of the 33 matches were 1-2 cousins with a cM of between 341 and 396. Do I leave these out? The other cM ranges were 103 to 336.
I also have 10 color columns, which I have only been able to narrow down to 5, 1 of which only has 1 link. What am I missing?
Any guidance you can provide would be very much appreciated.
Hi, Kristina. If you know how some of these matches are related then you USE the ones who share only one grandparent with the test taker, and you do NOT use the ones who share two grandparents with the test taker. First cousins who are once or twice removed can share one or two grandparents with you depending on his is of the older generation.
And, while it’s nice to get 4 clusters, the number of clusters is dependent on the test taker’s genealogy and who has tested in that cM range. So, the key is to start working with the clusters and figuring out how the people in a cluster are related to each other and to the test taker.
Also, there are often “clusters” of one because they just don’t happen to share any DNA in common with anyone else in the cM range you are working with – in this case from 90 to 400 cM.
Hope this helps!
Dana
Hi Dana,
Just started this process and it’s making sense! Is there anyway to incorporate names from 23 and Me into the spreadsheet. Is there anyway to add 23 and Me data into the spreadsheet? I have a much larger amount of 2-3rd cousins on that platform.
Thank you.
Hi, Terri. I keep my data from the various testing companies separate. But, you can also create a master spreadsheet showing all of the DNA matches across the various testing sites like Ancestry and 23andMe. In this chart, you could either label them as being from a particular grandparent line, great grandparent line, or some other category. I add two columns for the MRCA male and female and a third column showing which child of that couple is the DNA match’s direct ancestor. You could also color code this chart and have it match up with your Leeds Method chart and use the same color for the same part(s) of the family.
Hope this helps!
I’m not particular savvy on DNA. I thought I’d give your method a try. I have 52 people listed from 90-400 cm. I ended up with nine columns. Ugh. One thing I do know is that my mother has 15 siblings and two of her sisters married brothers. Is that why my chart is such a mess?! I don’t think I’m going to be able to get a clear picture of these 2nd-4th maternal cousins because of this.
Hi, Diane. If you see any descendants of your siblings, then don’t use them and you’ll be fine! Actually, if you know any of the matches between 400 and 90 cM then don’t use any who share more than one grandparent with you. You are trying to only use those matches who share only one grandparent line with you.
Search my site for “consolidating” or “consolidate” and you’ll see my advice for dealing with a lot of clusters. Most likely you just need to combine some of them.
Hope this helps!
Hello Dana,
I am working on my mom’s line and I have input 15 2nd to 3rd cousins and all of them are from one line which means they are all in the same clusters. I am sure that they are all from my 1st Great-Grandparents. In this line, they had adopted kids into and also adopted kids out from the family, which is making it hard to find who is their parents or their bio father is. Any suggestions on how to handle this?
Hi! I’m not quite sure if you are trying to find the bio parents of some of these 2nd and 3rd cousins or of your great grandparents. And it’s often helpful to collaborate with these cousins, if that’s who you’re trying to find the bio parents of, after making sure they’d also like to know the answers.
Best wishes!
Pingback: DNA Color Clustering: The Leeds Method for Easily Visualizing Matches - Dana Leeds | DNA kit Advisor
Hi Dana,
Thank you for many thoughts that help.
Wondering, would you be able to interpret MyHeritage screenshot of Autoclusters for feedback?
If yes, where can I send the screenshot? I lack info on my background, and did some analysis on the clusters and also drew lines extending the sides for each color group squares to make it easier to see possible distinct grandparent lines after combining multiple clusters or endogamous groups. Also separated many groups by distinct ethnicities. Was wondering if what I did could be valid.
Thank you in advance.
Hi, Martin. At MyHeritage, most people don’t have enough 2nd and 3rd cousin matches and enough matches with trees to make the tool very useful, in my opinion. You also can’t change any of the parameters to make it fit your data better. So I don’t really interpret those. If you have higher, known matches, it could be very helpful to look and see which clusters they match.
If you are able to combine clusters to represent the various grandparent groups, that’s great!
And I would be happy to take a look at it and tell you whether or not I see endogamy. You can email me at drleeds@sbcglobal.net
Dana
Unless I missed it, this may be a dumb question but I did not see it specified. For the people listed in Column A of the spreadsheet, are these manually entered? With potentially hundreds of relatives, that seems like a very tedious task.
Hi, Chuck. Yes, I do type them. I like it as it gets me familiar with the names. And, if I was going to do hundreds of matches I would use an automated tool. I don’t think I’ve ever done more than 70-80 by hand. Hope this helps!
For my wifes Ancestry tree, I have hit a major roadblock at her paternal grandfather. There seems to be no information about his existence prior to 1923 when he was married for the 1st time. The marriage license lists his mother and father including their “from” location (state only and not necessarily where they were “born”), yet each of them are a complete mystery and do not come up in any other respective records – as if they were not real people. Almost as if they died, or the ones noted were not actually his parents, perhaps other relatives or even his grandparents – I suppose that’s possible.
I wanted to try your color-coding method using my wifes DNA matches from ancestry, but frankly, it does not appear there are any matches to that paternal GGF line..seems like a dead end?
Any thoughts?
Hi, Chuck. It could be that his mother and father were immigrants from a country that hasn’t tested much. Or there could be another scenario! I’ll email you.
Dana
Pingback: How Low-cM DNA Matches Can Help You Solve Genealogical Brick Walls – Josiah Schmidt, Professional Genealogist
Hi Dana,
I only have 10 matches between 400 and 90 cM, and I had to drop the top one (340 cM) because she is my 1C1R and shares 2 of my grandparents.
Of the 9 remaining matches, I had 6 in column 1 (my mother’s maternal line; I know the top 3).
In column 2, I had 2 matches: a mother and daughter who match me on my dad’s maternal line (I was able to tell them about their ancestors, even the ones not related to me).
The remaining match had, of course, no matches. However, this person is a match to another 1C1R and is therefore a known maternal relative, but I can’t identify if they are related to my mom’s mother or her father. Even though the person has over 8000 people on their tree, none of the surnames match any of those on my tree, even out to the 7th generation and sometimes beyond. So I can’t assume the person represents my mom’s paternal line. The tree *is* linked to the person’s DNA. I am at a loss.
The most frustrating part, however, is that no one shows up as a relative to my dad’s paternal line. My paternal grandfather’s parents are my two major brick walls. I know for certain my direct paternal great-grandfather was Ashkenazi Jewish, although he pretended to be French. DNA testing has proven this beyond a shadow of a doubt. Unfortunately, my highest strongly Jewish match is only 41 cM. This also holds true on other DNA websites. And because I don’t know my ancestor’s real name, I can’t look for him in anyone’s trees. I’ve reached out to people without success., and a search angel has come up empty so far.
My ggf’s “wife” claimed to be English, but we can find no record of her before they supposedly married. I have yet to identify any matches related to her. (I have over 45,000 matches on Ancestry.) Since my father’s maternal ancestors were also English, her ethnicity does not distinguish her. We know from the amount of Jewish DNA various family members have that she could not have been Jewish. My maternal background is Scots-Irish (with a trace of Norwegian).
I have been trying to use triangulation to find my Jewish ancestors. I realize endogamy makes this extremely difficult, but a professional genetic genealogist examined my data and came to the conclusion that this is my only hope.
I did try taking my Leeds chart down to 41 cM, but I just kept getting more and more people who matched no one else in the list.
Is there anything else you can recommend?
Thanks.
Since it sounds like your Ashkenazi Jewish heritage is only maybe 1/8th of your admixture, I suggest creating clusters using the Collins Leeds Method. (https://www.danaleeds.com/the-leeds-method/ and scroll down – there is also a great Facebook group for users of DNAGedcom.) Then I would concentrate on any unknown clusters and try to find common ancestors.
Another method would be to just go down your list and work with your highest unknown matches. But, with each one, look at their shared matches and work with the group as a group. Hopefully, you’ll find some connections between these people!
Best wishes!
Thank you for your speedy response, Dana!
I have actually been doing this (clustering my highest matches) for over a year. I am currently attempting to triangulate 8 people who match one another and me. No common ancestors have appeared yet. Given that my highest matches are in the 30-40 cM range, and they likely have endogamy in their backgrounds, these people might not match me very closely at all.
My brother has identified all the descendants of our ggf up to around 12 years ago. It does not appear that any of them have tested, not that this would help me. Ideally, I need someone descended from his parents, i.e., a 3rd cousin. That hasn’t happened yet.
I shall keep plugging away. I do not give up easily.
I should clarify my statement about none of my ggf’s descendants testing. Four of us who descend from my grandfather have tested: myself, my brother, a paternal uncle, and a 1C1R descended from another paternal uncle. My ggf was “married” to a different woman before he met my ggm, and my brother has traced her descendants. None of *them* appear to have tested.
Sorry for any confusion.
Hi and thanks for the excellent explanation! I’m looking for the biological parents of my great-great grandfather, and I’m able to work with my grandfather’s DNA (so, I’m able to identify my grandfathers 2nd-3rd cousins). This being said, I’m left with only 6 individuals with 90-150 cM) and they all overlap… As all his grandparents were supposedly from the same village (with 700-1000 people) I suspect a pedigree collapse (and that a big one!). Any advice appreciated! 🙂
Hi, Camilla. Yes, that does sound like pedigree collapse! (Or possibly even endogamy.) The Leeds Method would not be helpful in that situation. It’s a subscription site, but I recommend watching “Dealing with Endogamy” by Paul Woodbury at Legacy Family Tree Webinars if you can.
Best wishes!
Dana
Hi Dana, thanks for this great method for sorting DNA matches. I have created my spreadsheet using matches =90 – only 23 people. I have 8 columns! 4 of the columns I can trace to my paternal side, but I’m not sure should it have broken into 4? The other 4 columns should be my maternal side. The issue I have with my maternal side is that my grandfather is not confirmed for one, and my great grandparents appear to have just materialised out of nowhere. Would it be worth coding the 4th cousins, of whom there are many? I will check your site also for more useful tips 🙂
Thanks
Michelle
Hi, Michelle. With the 4 columns for your maternal side, do you have overlap? (Search my site for “overlap.”) If so, you can likely consolidate them. If not, it might be that your matches are 3rd cousins instead of 2nd cousins. Which is fine! You just need to create the clusters and then figure out what the clusters mean.
As far as the grandfather who is not confirmed… on the paternal side do you have clusters you do not recognize? If so, concentrate on these. Try to look at the trees and find connections. Try to build out the trees to also look for connections. And go to the Shared Matches of the matches in this group and do the same.
Yes, extending to 4th cousins can help.
Best wishes!
Dana
I have created my spreadsheet using matches between 90 and 400 cm, but I only get 8 matches. They result in only 2 columns. Not sure what to do? Any help would be much appreciated.
Hi, Camille. It’s still very helpful to see those 2 clusters. If you have higher matches, go ahead and add them. And try going a bit lower. But now the main thing to do is to figure out why those clusters formed – so how are the people in the cluster related to each other AND how are they related to you.
Best wishes!
Dana
Trying to us the Leeds method to sort out my ex FIL’s DNA matches. Unfortunately he only has 12 matches in the 400cM – 90cM range which only gives me two of his grandparents and only on one side. If I lower the cM levels down to 20 I pick up another grandparent but I have lots of clusters that might match one or more persons in the original two clusters. I’m not quite sure if I should combine clusters when only a few match and a lot don’t or just leave it for now. I tried using the dots on Ancestry based on “common ancestors” shown with ThruLines (and yes I know ThruLines should be seen as just a suggestion) and that seemed to work for the most part, but that became a fat mess too.
Any suggestions as to what I might try now.
Thanks for your help.
Hi! It sounds likely your other grandparent is either from a small family or a region of the world that hasn’t tested much. Or perhaps it’s a group that just happens to have not tested!
As far as combining clusters, my rule of thumb is to combine when most of the people of one cluster match most of the people in another cluster. So if it’s a minority of people who are matching, I wouldn’t combine them.
Although it’s wonderful when we see 4 clusters based on 4 grandparent lines, that is not the goal. The goal is to form clusters of people who are all closely related to each other. So, no matter what clusters form, our next step is usually to determine how the people in the cluster are related to each other and then to determine how we are related to that cluster. Usually, that’s because we all share a common ancestor or are otherwise closely related!
Best wishes,
Dana
Thanks that does give me some direction to go in. Had lots of clusters that only had one match to another cluster.
Pingback: Should I sort my 23andMe matches by percent or strength of relationship? – Family Locket
Hi Dana
I am trying to use your Leeds method but I am not getting very good results.
My Grandfather was a foundling and I am trying to establish who his birth Father was. The results I am using are my Uncle’s so it is his Grandfather I am looking for.
For My Heritage matches he has 19 matches between 158cms and 40 cms (I went lower due to few matches). When I put them into a spreadsheet and colour code them I end up with 13 separate groups! I can merge two of them but that still leaves me with 12 separate groups. How can this be?
Ancestry is completely different. Using only the 37 matches he has between 168cms and 40cms I have come up with 3 groups (there were more but I was able to merge them). One group has 27 people in, one has 9 and the third has just one person.
I have been trying to work on this for years but feel I may have to give up. Looking at other people’s Leeds methods they seem to have much better outcomes.
I’d appreciate your opinion. Do you think I should just give up until I get better matches in years to come?
Hi, Julia. I’m guessing your grandfather was either an only child – or from a small family – or that his family were recent immigrants so you don’t have a lot of “cousins” who have tested. It is possible you will be able to solve this case, though!
First of all, I would also suggest transferring your uncle’s DNA to FTDNA (FamilyTreeDNA) for free to see if you get any good matches. And, if possible, have him test at 23andMe.
Then work with your highest match at Ancestry – 168 cM. Just look at the shared matches. See if you can find a connection between that group of matches. Then, like the Leeds Method, go to the highest match who isn’t part of that group and do the same. You might have to build out trees for these matches! Maybe you will see a country that is jumping out at you.
Best wishes!
Dana
I used the Leeds Method with various degrees of success. After having my DNA tested, I discovered that my mother’s father wasn’t her biological father. It seems my grandmother had an affair. While mom always suspected that this was true, grandma assured her it wasn’t. I used your method to see if I could find mom’s lost family.
I came up with five branches. Three were what I expected, my Paternal Father’s family, Paternal Mother’s, and Maternal Mother’s. I even found some cross over between my Paternal Father’s family and my Maternal mother’s family that was explainable.
Researching archives in Ancestry as well as newspaper I believe one of the remaining branches is my Maternal Paternal Branch, and that through shared DNA, I believe I found the name of mom’s biolgical father. I share 267 cMs of DNA with his great granddaughter, making her my half 1C1R. Other relations I share with this branch are the grandsons of my newly found grandfather’s brothers with enough DNA to be 2C.
The 5th branch is the kicker. My newly found 1C1R and I share a strong DNA connection to the branch-5 as well as to the members of branch-4, however, none of the branch-4 2C, share a connection to branch-5.
In my mind the only possible way that my new 1C1R and I could be related to branch-4 and branch-5, while my new 2C are not DNA related to Branch-5, is if my newly found grandfather had a different mother than his brothers. However, there is strong documentation that refutes this, it is the only possible answer based on the high DNA cMs we share to both branch-4 and branch-5.
Any suggestions or tools that might help me solve this puzzle?
Hi Dana, after reading your tutorial and watching your presentation at RootsTech2019 I noticed two things: 1) AncestryDNA relationship predictions are not always accurate, and 2) some matches fall well below or above the 90-400 cM thresholds. An example for the first observation, my confirmed 2C1R and I have well defined and sourced trees including family photos…AncestryDNA indicates 51 cM shared DNA with me and 56 cM shared with my father and predicted relationship 4th-6th cousins. Odd. An example of the second observation, again my 2C1R has a confirmed 1C2R relationship sharing 167 cM. Again, odd. When building my cousin’s Leeds table would you include me and my father, and would you include the 1C2R? Thanks in advance.
Hi, Greg. You’re absolutely right! The relationship predictions at Ancestry – or any of the testing sites – are often inaccurate. We need to remember that they are estimates! And “we – whether as people or as a computer program – cannot determine an actual relationship based solely on the amount of shared DNA. We have to also do traditional genealogy.
The 90 to 400 cM range was given after a lot of trial and error. This range was for those with unknown biological parents who did not know how their matches were related to them. If you DO know how some of your matches are related, you want to use those who share only 1 of your 4 grandparent lines with you. And it is often helpful to work with those matches who share less than 90 cM. But I recommend you start with that range if you do not know your matches or your own family history.
Dana
This is very helpful!
On 23 and Me, they have an auto-cluster feature. Is this essentially the same thing? Have you used the feature and compared the accuracy vs. doing it manually?
We found my husband’s unknown father using a similar method, and making random phone calls. I highly suggest to anyone searching who isn’t getting messages replied to via DNA sites to not be afraid to look up phone numbers and make calls the old fashioned way, especially if you want to contact an older generation relative. A 3rd cousin gave us the key clue to finding the closer relatives that weren’t in our DNA matches and that’s what ultimately led to the end.
Now I’m trying to assist a cousin with an unknown father and this will method will be very helpful!
Thank you again!
Hi, Ashlee. Do you mean on MyHeritage? They do have AutoCluster which is based on the Leeds Method. However, you cannot choose any parameters and it often misses your closest matches. It is accurate, but most people just don’t have enough good matches to make a lot of progress there. But it can still be helpful!
That’s great that you were able to find your husband’s unknown father by reaching out to these cousins! And I hope my method works with your cousin.
Sincerely,
Dana
Pingback: Research planning with DNA – Generations Genealogy
Hi,
My Great Grandparents were 1st cousins. So, I am thinking I will have to have extend to my Great Grandparents instead of just my Grandparents. Is my thinking correct?? If not correct any advise?
Thanks
Hi, Janie. Since your great grandparents would be the parents to just one of your four grandparents (if that’s what you’re describing), it actually doesn’t affect the method. In a best-case scenario, you get 4 clusters – one for each grandparent line (or pair of great grandparents). This doesn’t change in cases like yours.
Best wishes,
Dana
Thank you so much for developing this method! I am hoping to be able to use it to break down a brick wall surrounding my 3x great grandmother, as conventional research has not been sufficient. I guess the only thing I seem to have missed is: let’s say that I’m on Ancestry, looking at my DNA matches, between 90-400cM. At what point would I go to look at shared matches? Your instructions seem rather vague (to me) about this bit of information…. Thank you for your assistance!
Hello Dana,
I carefully followed your method step by step as I so wanted to work out who my father’s Grandfather was. I was hoping for ‘4 Columns, No Overlap’ but of course I got ‘7 Columns, 2 Overlaps’. My 1st & 4th columns are definitely my mother’s side, the rest my fathers. Now I’m just confused & still no idea who my Dad’s Grandfather was.
Hi, Debby. Now you need to concentrate on the groups that are on your father’s side. Can you figure out which group or groups are on your dad’s grandmother’s side? Do you have one or more “unknown” clusters left over? If so, concentrate on these! Look at the trees. Look back to 8 great-grandparents or 16 great, great-grandparents. Hopefully you are seeing people who are on more than one of these trees. If so, start studying this family.
I just spoke at i4GG where I identified the unknown parents of my mother-in-law’s grandfather. It turned out he lived under an assumed name! But by following the DNA matches, I was able to figure out who his parents were and then use traditional genealogy to figure out which brother had “become” her grandfather.
Best wishes!
My dad’s father is unknown, and I’ve been working to try and identify him. I followed your method and came up with 6 clusters. 4 of the clusters have pretty heavy overlap. The other 2 don’t overlap at all with any other clusters. I’m not sure where to go from here. Any suggestions?
Hi! I have several posts about consolidating, but you can also work with the 6 clusters “as is.” I’m not sure if this is your test or your father’s. But your next step is to figure out which part of the family tree the various groups represent. For example, if this is your DNA, then you might be able to identify some as your mother’s. The others should be your father’s. So then start working with each of these cluster(s) that are likely your father’s father. Try to figure out how the people in the cluster are related to each other. And then diagram (search my blog) the cluster and see if you can see a likely candidate for your dad’s father.
Hope this helps!
Dana
Thanks for your reply. To clarify, this is my dad’s test. I have also tested, if that helps in any way.
Thanks Dana. I have a 2nd – 3rd cousin who was and adoptee and trying for find her father. I have done the leeds chart and know which line she belongs to. What would you recommend doing next to identify which of my many cousins was her Father? Thanks for you help.
Hi, Karen. I just recorded a webinar for Legacy Family Tree Webinars yesterday. It’s free until May 11th (and then available to members). It is broken into sections. I think you find the section titled “identifying surnames” and “case studies” to be helpful.
Hope this helps!
Good Afternoon,
I am working on my mothers tree (her and her brother both took DNA tests), they share 1700cM and it says they are only half siblings). Could this be because their grandparents (on their fathers side) my be related. Both of the grandparents had the same last name before they were married. They are both from Croatia and were born in villages about 10 miles apart. When I make a Leeds chart for my mom or uncle, they each only have 3 colors (neither have many matches in the 90-400 range). Thank you
Hi, Mike. Ancestry cannot determine exact relationships based solely on the amount of shared DNA. But, unfortunately, 1700 is extremely unlikely to be a full sibling. (It might be impossible though a few people have self-reported these amounts. It is likely they were not aware these were half-sibling relationships.) As far as their grandparents being related, this would actually increase the amount of DNA they shared.
If you can possibly identify some matches as being only from their dad’s side and others from only their mom’s side, you can see if they BOTH are matching these two sides. Or is one of them missing one side? (And it’s usually the father’s side that one of them would be missing.) You might also need to test cousins from both sides of the family to figure this out for sure.
I know this can often be difficult both to figure out and to come to grips with. Best wishes!!
Pingback: I got my DNA tested, now what? | Jennealogie
My dad and two of his sisters all took an Ancestry DNA test.
All three of them are full siblings.
Is it helpful to do the Leeds methods, treating their matches as if they are one set of matches vs. 3 individual ones? If so, are there any special hints or suggestions you would have?
Or is it better to do them individually?
We started a spreadsheet listing all the matches that each of them had in the 90-400 Cm range. Some of them had unique matches so everyone’s matches were put on the sheet.
When a person matched all 3 siblings, due to varying amounts of Cms, the starred match would not always be the same person. So, would we always use the highest Cm match, no matter which sibling it matched? Is it better to use one sibling for identifying the starred match, or look at the one with the highest amount of Cms no matter who they match?
Hi, Connie. When siblings test, I recommend doing a separate Leeds Method chart for each of them. Then you can combine them into one master DNA chart. (I should blog about this, so thanks for the reminder!)
Dana
I am unsure which persons to use for the second color. I have used 23Ansley, if that makes a difference. Do I list the people who are listed for me and person one both as a 3rd cousin?
Hi, Patricia. I think you meant 23andMe, right? Most people have fewer matches at that site so the Leeds Method doesn’t work quite as well. I would work from about 5.38% to 1.21%. Just to highlight the difference, that gives me only 9 matches on 23andMe while I have 36 matches in this range on Ancestry.
Best wishes!
Hi Dana,
I’m attempting to identify the identity of birth father. I have separated known maternal matches (mom tested so this was easy) I have 35 matches from 392-81 that are suspected paternal. I listed them in excel and started with my first color and ended up adding that first color to ALL the others, as they all matched the 392cM match. What should I do to determine where the next color should be applied? Should I use smaller cM matches? Thanks!
Hi, Lorraine. Yes, go ahead and start with a lower match. It is likely that this person is related to both of your paternal grandparents. If it happens again, start even lower.
My father was adopted. My spotlight question is to identify my father’s birth parents. In AncestryDNA, I have 53 2nd-3rd cousins (between 90 and 400 cM). Of those, nine are identified as having a common ancestor with me. Seven of those nine are the same common ancestor. The other two are one ancestor each. All nine has a common surname. Six of the remaining cousins are related (via trees) to my mother’s side of the family.
My question is how do I handle the remaining cousins, especially since AncestryDNA doesn’t identify any common ancestors for those cousins? I may be able to assume common some ancestors based on family trees, but not sure whether this is acceptable practice.
Hi, Lynn. First of all, I hope you used the Leeds Method to create clusters. If so, the next step is to figure out how the people in each cluster are related to each other and then how they’re related to you. When available, I start with the trees that these matches have and try to identify people or surnames that are repeated in more than one tree. Sometimes, though, I have to build trees for these matches. This is, in my opinion, an acceptable way to start the research. But when I think I have the answer, I verify all of the parent/child relationships myself.
Hope this helps and best wishes,
Dana
Dana, thank you so much for this method. So helpful! I have a question about the “lots of overlap” paragraph/graphic. In your example, you stated the sort represents 3 of the 4 sets of great grandparents. and that the paternal father’s side has no cousins matching. This is exactly what my LEED chart looks like – all of my overlapping clusters are on the maternal side (my great grandparents were cousins, and think this was common practice w/in the family). I have a few matches on my paternal side and I am able to clearly connect with my fathers maternal line. I have found no matches on my fathers paternal line. My LEED chart looked between 90cM – 400cM. Should I expand it to below 90cM to see if fathers paternal line will show up or will the endogamy on my mother’s line blur the findings? By comparing my chart with a 1C, I know that father’s paternal line has tested and I don’t seem to be finding any matches. Could endogamy cause the chart to show NO matches at all on my father’s paternal side (I know they also married cousins on that side as well). Thanks so much.
Hi, Elizabeth. Yes, I would go ahead and work further down and see if you eventually see that 4th grandparent line. Your mother’s side won’t effect this. BUT, if there is endogamy on your father’s side, you might only see one cluster.
However, you mentioned that some on your father’s paternal line but you aren’t finding those matches. Are you on the same testing site as they are? Are they 2nd cousins or closer to yourself? If there are matches that are out there that you should be seeing but aren’t, that could be another issue. (You can email me at drleeds@sbcglobal.net
Dana
thank you so much! this is very helpful. i’ll send an email.
Thank you so much for this!
I started doing my tree during Covid Lockdown primarily for something to do, but also to try and solve the mystery of my adopted paternal grandma’s family. We had her mother’s details on her birth cert before she was adopted, but father was ‘unknown’. Thanks to this method, and my DNA, I was able to not only identify her paternal familial line, but also the identity of her father! As a result, I now have contact with many cousins that have been able to provide me with stories and photos of her dad and her many siblings!
That is wonderful!!! Thanks, Christine, for sharing this success with me. 🙂 I love to hear those stories!
Dana
Hi Dana, I have only two matches who share 90 cM or more. Is the Leeds Method still useful to me?
Hi, Sebastian. Unfortunately, it really won’t be very helpful. Instead, I just suggest working with Shared Matches. So look at that top match and see their Shared Matches… and look at the 2nd match and see their Shared Matches.
Best wishes! You might also upload your DNA results to MyHeritage and FamilyTreeDNA, too. Or try testing at 23andMe if you haven’t already.
Hello,
I am very confused! According to my DNA matches on Ancestry, I have 182 2nd-3rd cousins! This is my second attempt doing the Leeds method, (using 245-90), because it just doesn’t seem right. There are lot of people who have at least 4 colors in the same row. I am on column O, and still have at least 10 people that still have blank spaces. What am I doing wrong? Any help will be greatly appreciated, because I’m quickly losing patience.
Hi, Nancy. Do you know much about your family history? When someone has that many 2nd & 3rd cousins, it suggests possibly pedigree collapse, endogamy, or just large families. If it’s pedigree collapse or endogamy, unfortunately the Leeds Method does not work that well in those cases. (Pedigree collapse basically means you have the same ancestors on your tree more than once – usually because cousins married often in the 1800s. Endogamy happens in specific populations where they intermarried generation after generation like Ashkenazi Jewish or island populations as examples.)
If you look on my site for “DNAGedcom’s Collins Leeds Method (CLM)” under the Leeds Method tab, you might try that. You can also email me if you are still unsure what is going on.
Dana
I have a problem with all the advice I have seen to date, which is that it invariably starts with the assumption that both grandparents are shared. In my case my mother was born illegitimate, my unknown grandfather went on to have more children by 2 other women and they in turn had offspring. I have identified second cousins, ie we share the same 2 great grandparents, and a half first cousin (grandfather in common but not grandmother) but it is difficult to figure out which of several sons of those great grandparents is my grandfather. And the reason for that is the absence of advice dealing with such a situation, although one would imagine it is common enough. Would you care to comment?
Hi, Alex. As I’m guessing you’ve read, the Leeds Method often sorts matches into 4 groups based on your 4 grandparent lines. So what if one grandparent had multiple wives? It will really depend on who is in that cluster and who is the top person in that cluster. But you might see just ONE cluster for that grandparent. Or you might see several clusters. If you’d like additional feedback, create a chart and then email me an image of the chart and I’ll try to give you more specific feedback.
Best wishes!
Dana
Hi Dana,
Is it possible to take one of the clusters and separate those the same way? I have my maternal grandmother’s relatives in a cluster, but I’m trying to get a group of her father’s relatives, which we have no information on. If so, what range would I use?
Hi, Jolanda. I still recommend starting with the same Leeds Method using the range of 90 to 400 cM and then adding any higher matches. Then determine which cluster or clusters represent this maternal grandmother. Everyone in that cluster should either be a descendant of her parents OR a relative of her mom OR a relative of her dad. There are a variety of steps you can take from here, but one is to use 3 different dots to mark that group (if you’re in ancestry). If her father is unknown, you could mark those you know are descendants of her parents and those you know who are relatives of her mom’s. You’ll have left overs. Some of those will be unknown relatives of her mom. You can see by who they match. Hopefully you’ll also create an unknown group and this should be her dad’s family!
Hope this helps!
Dana
Hi Dana and thanks so much for sharing your methods so generously. My tree has pedigree collapse, my parents are related as their mothers were first cousins ( their grandfathers were brothers) . What should my spreadsheet look like if I am doing this correctly?
Hi, Elaine. So we *hope* to see a cluster for each grandparent. You will see pedigree collapse with those two grandmothers so they will be mixed in one (or more) clusters. But your two grandfathers should likely have their own clusters. So, in a best case scenario, you’d see 3 clusters.
Hope this helps!
Pingback: Ancestry.com Sideview ™ Review | learnalittleeveryday
Pingback: Still here…. – Fitting into my Genes
Pingback: DNA and Google Maps: Breaking Through Brick Walls to Reveal a Love Story – MyFamilyPattern.com
Hi Dana
My testee’s 400-90cM matches are 14 people who fall into 6 distinct colour with no overlap. One with 134cM is the only one in the group with no connection. What could this mean? I am going to go down to 50cM and see what happens.
Hi, Clare. That likely means you are seeing 3rd cousins instead of 2nd cousins. You need those higher matches to form clusters based on a single grandparent line. But this can still be quite helpful! The next step is to figure out how the people in a cluster are related to each other. And, yes, you can go both higher (if there are any) and lower now.
Dana
I tried to follow this system, and I cannot tell what it means. My charts has many overlaps with about 30 people almost evenly divided between my maternal and paternal sides. Can anyone tell me how to make this chart useful?
Hi, Jim. I have 3 posts about consolidation that might help you. You can use the search feature on my website to find “consolidate” or try this link: https://www.danaleeds.com/?s=consolidate
If you still have questions, please feel free to send me an image of what you’ve done and a brief description and I’ll give you some feedback.
Sincerely,
Dana
I just did this with my matches. I have 8 columns of colors. Lots of overlap. I don’t think mine is pedigree collapse or endogamy though. I have 2 brothers who married 2 sisters on my maternal side and I have a brother and sister who married a brother and sister on my paternal side. I do not have overlap of maternal and paternal. I have lots of overlap of paternal with paternal and lots of overlap of maternal with maternal. Hopefully that makes sense. Is there something else I can do to help sort these down more to get a clearer picture of what’s going on? Thanks!
Hi, Erika. It depends on how far back these sibling marriages are as to how it would affect your DNA. You might read this article: https://familylocket.com/endogamy-pedigree-collapse-and-multiple-relationships-whats-the-difference-and-why-does-it-matter/
Basically, if you have two grandparents who have recent common ancestors then they will have a lot of shared DNA. But if it’s further back, you should see separate clusters. If you haven’t looked at my “consolidating” posts, search for that term and see if it helps.
Best wishes!
Dana
Hello Dana,
I have just found about using the Leeds method and began using it. I am using Ancestry. I have a DNA match (2% Shared DNA: 138 cM across 7 segments) and when I look our Shared matches, they do not match any of the people that I found between 90 to 400 cM list. I am wondering what that means.
Thanks
Jim
Hi, Jim. Great question! Perhaps this match is from a part of your family that are recent immigrants or a part of your family that has had few children. It might be helpful to see if you have all 8 of your great grandparents represented.
Dana
My wife is 3/4 Sicilian and 1/4 Heinz 57. I did the test on her matches and I could very easily determine the 4 clusters. I am of Acadian descent and there is so much pedigree collapse in the Acadian descendants that I cannot determine anything from my test. I did the test on a relative who has 3 Acadian grandparents and 1 Jewish grandparent. As I thought would happen beforehand, her test had one big cluster (the Acadians) and one very small one. Some in the small cluster had obvious Jewish surnames. So, I found the test very useful in some cases, but for populations like the Acadians, the results will be inconclusive.
Faron, Thanks for sharing! Yes, in specific populations test takers are genetically related to “everyone” so the matches form one large cluster. The main criteria is how much of your family is in one of these populations. If it’s the entire family, you’ll probably have one large cluster. Just one parent’s side? You will hopefully see a cluster for them but two (or more) clusters on the other side. And if it’s just one grandparent line you should see all 4 grandparent lines.
Sincerely,
Dana
Hi Dana,
I’ve just been doing your method for the ten matches I have between 90-400, and the thing is I have five distinct groups. I know that my paternal grandmother was the only child of an affair, so would the fifth column mean that one set of my matches are half-relatives through her unknown father?
Hi! That’s quite possible! After creating clusters, the next step is always to figure out what each cluster represents. You can now add in higher (if any) and lower matches to help.
Best wishes!
Dana
Pingback: DNA and Google Maps: Breaking Through Brick Walls to Reveal a Love Story - MyFamilyPattern.com
I already know my great-grandparents and those of my wife. (And most of their descendants.) What cM numbers would you use for more distant relatives? Or are those too far out for the method to work?
Hi, Wesley. I recommend you do something similar but not on a spreadsheet. If using Ancestry, I would suggest assigning 8 colors to your pairs of 8 great, great grandparents (or you can think of it as your 8 great grandparents). When you find 3rd cousins (or those sharing a pair of these 8 2x great grandparents), assign that match the correct color. Then assign that color to all of the shared matches. If that’s a bit confusing, I’m actually working on a presentation for this right now!
Dana
I don’t have access to the ancestry info, so I modified the method on GEDMatch as follows:
1. Get the matches for my kit.
2. In the first fifty matches copy all with “Gen” 2.0 or more and 5 or less into a spreadsheet.
3. Get the matches for the first kit on that list
4. In the first fifty matches for that kit copy all with “Gen” 2.0 or more and 5 or less into a file.
5. I didn’t see the point of colors, since fifty colors would be difficult to get, and since they are in different columns, putting a “1” in the column seems just as good as a color.
6. I didn’t do all fifty kit columns, just the first eighteen. (Tedious process—I might try to automate it.)
7. I inserted a column that totaled how many ‘1’ were in each kit row and sorted by that, moving the ones with no shared matches (or not checked) to the bottom. The first eighteen are the closer cousins anyway.
Now, here’s the weird thing: My kit is actually Y-chromosome from my mother’s male cousin. But the three kits with the most shared matches (5, 3, & 3) are from females!
Im Very Confused is this copy and Pasted information from my DNA or is it manually added? is my DNA stuff Stuck in Ancestrydna.com?
Hi, Chris. I manually type it into a spreadsheet.
Hope this helps!
Dana
I’m confused, I have 5 columns. I have 13 2nd-3rd cousins lower than 400cM broken down as follows:
1st column = 7
2nd and 3rd column = 2 in each
4th and 5th column = 1 in each
Hi, Paige. It’s likely that you just don’t have enough 2nd cousins to “gather” the matches together. Here’s a post which might help: https://www.danaleeds.com/leeds-method-time-to-consolidate/
Dana
Hey Dana – I have had a ton of success using the leeds method in helping others with their research and I really appreciate all the work you’ve done. I now have my late grandfathers dna test and have never seen anything like it. I have TONS of matches but only 3 very clear groups. Only 1 person overlapping. I’ve even added in 4th cousins (after high cm matches/sorting clusters of course).
The majority of the matches go in the first column (blue cluster). That column has trees with surnames matching both of his maternal grandparents. Example: Person A has maternal grandfather last names, Person B & C has maternal grandmother last name, etc.
Does this just mean that that family line is related to one another?
Hi, Sam. I’ve been meaning to write a blog post about how age can affect the Leeds Method. Sometimes the oldest testers show only 2 groups!
That’s likely what’s happening in your case. Do you know how the first person in the blue group is realted to your grandfather? If they match BOTH of his grandparents, you need to skip them. If you’re not sure who they are, you should skip them. Start a group with someon you know is only related to ONE of his grandparents. (Unless you know the two grandparents are related to each other, which it doesn’t sound like is the case.)
Hope this helps!
Dana
The noosphere is working – I was going to write to you about a difficult case I just solved where age was the culprit.
A person born in 1955 was looking for his father.
Over 60k matches, over 200 90cm+ matches.
2 maternal clusters, and 1 massive paternal cluster. Overlap pretty much everywhere. However, no signs of cousin marriage.
Turns out that most of the matches were not 2nd cousins, but actually 1C3R : the father was born in 1889.
Hopefully, descendants of 2C had tested too. Once the descendants of the 1st cousins were removed, the clusters lined up perfectly with very little overlap.
Hi, Kim. In my talk at i4GG earlier this year, I explained how age can affect our clusters. You have the *perfect* example here! So, first of all, I need to blog about what I shared. And, second, I’m going to reach out to you and see if I can share some of your results.
Thanks for reaching out!
Dana
You mention you will have 4 clusters of colors, and the what if if you have 3, or the overlap of clusters,but what if you have 5 clusters? One of my clusters overlaps with another, but not all matches are shared within the 2 clusters, but that caused the additional (5th) cluster.
Hi, Rachel. First of all, my sincere apologies for the delay! I haven’t blogged in quite awhile and somehow I stopped receiving comment notifications and hadn’t noticed. I’m quite behind but catching up tonight!
First of all, having 4 clusters is a best case scenario. At the end of the post, I explain what it could mean if you have fewer or more color clusters. Not everyone gets 4 clusters.
5 clusters is perfectly fine. You can even work with thme as is! The next step is to figure out how the people in the cluster are related to each other and then how they’re related to you (or the test taker).
Sincerely,
Dana
I tried this for the first time and ended up with 6 columns – I can’t see that this has been covered here. Am I missing something?
Hi, Jill. First of all my apologies. Somehow I stopped receiving notifications of comments so I’m quite behind.
Try this post: https://www.danaleeds.com/leeds-method-time-to-consolidate/
Also, just search my blog for consolidate or consolidating.
Hope that helps!
Dana
I keep looking at these posts about the charts and how easy they are… but I really MUST be stupid cause I really DO NOT get it!
i have MyH and Ancestry and I have people I now know on there (maternal and paternal) but that is it… I just cannot get my head around this…. I feel rather frustrated TBH
Hi, Tamaa. First of all my apologies. Somehow I stopped receiving notifications of comments so I’m quite behind.
I’m sorry to hear this. A video might be helpful? I recommend the one I recorded on Legacy Family Tree Webinars which you can access with a membership or a free 7-day trial.
Hope this helps!
Dana
I don’t have such high matches above 90cM for my fathers DNA.
My best three matches are around 70cm, and the next three best around 50cm.
Can I seperate his parents (one unknown) and grandparents either?
Hi, Peter. First of all my apologies. Somehow I stopped receiving notifications of comments so I’m quite behind.
By looking at your email address, it looks like you’re from Germany. I know countries outside the U.S. often have a lot fewer matches.
It is unlikely you’ll be able to separate your grandparents using these lower matches. If you’re on AncestryDNA, it should already sort paternal verses maternal. It is soimetimes possible to identify unknown biological parents from these lower matches, though. The Investigative Genetic Genealogy field often works with these low matches. Often, you have to find additional people to test. For example, if the 70 cM match is on the unknown parent side, then you see if you can test one or both of their parents to determine which side of their family you are related to. Or find an aunt, uncle, cousin, or someone.
It’s not easy working with low matches! And I’m not that familiar with testers from Germany, but in the States we usually recommend that those seeking biological parents be in all 4 bigger databases: AncestryDNA and 23andMe – both of which you have to test at – as well as FamilyTreeDNA and MyHeritage – both of which you can transfer results to.
Best wishes!
Dana
Hi Dana,
I have put together a spreadsheet with 146 DNA relatives from Ancestry, ranging from 90 to 392 cM. I’m up to 14 different colors with a couple of folks not explored yet. At this point, I don’t thing 3 more matter. 🙂 Either I am way off base with my following your very clear instructions, or something else is going on. I really don’t believe I have cousins marrying cousins, I have found a few 2 or 3rd cousins marrying, but this seems extreme. Any idea of what may be happening? Thanks! Dave
Hi Dana,
I have done this spreadsheet for two people (who are related) but we are trying to see if they are only half cousins (Only sharing 1 grandparent) as we think one of the fathers was illegitimate. Does the chart work for half cousins or do you have to widen the cMs to compensate?
Hi, Karen. You would be creating independent Leeds Method charts, so it wouldn’t matter that the person is a half 1C. The main way to identify if someone is a full vs half 1C is to see if they have matches to both or only one of your grandparents on that side. (Endogamy and pedigree collapse change this!)
Dana
I don’t understand how I determine who is person number one. Do I just take a match at the top of my list (90-400cms) and count them as number one? Then go to the next in their shared matches and count as number two? and so on.
Hi, Sharon. I’m glad you reached out! So, you won’t be using “Person 01,” “Person 02,” etc on your spreadsheet (or piece of paper). Instead, yours will be a list of your cousins. You will list them in order from the highest match who shares 400 or less down to the lowest match at 90 cM or above. That highest person would be your “Person 01,” for example.
If you’d like to see a video, I have one at Legacy Family Tree Webinar. You’ll need either a membership – which is well worth it – or a free trial. Here’s a link: https://familytreewebinars.com/webinar/quickly-organize-your-dna-matches-with-the-leeds-method/ If that doesn’t work, find the site and then type my name in the box and look for my “Leds Method” talk.
Best wishes!
Dana
I have been looking for a way to “document”
matches across websites.I think this method is interesting but wonder if /how it can be useful for me? I am from a 100% Ashkenazi background with LOTS of endogamy. I cannot as yet identify second cousins (I just recently found the names of my great grandparents. – only first cousins with many “removed.” Is there any way you can suggest I can make progress with this method?
Sincere thanks.
Hi, Anne. Unfortunately the Leeds Method doesn’t help for endogamy. I wish it did!
AutoCluster is a tool based on the Leeds Method. And about a month ago they released a tool for working with endogamy at GEDmatch! My friends, Everet-Jan Blom & Jarrett Ross, were involved in its creation. I haven’t checked it out, but maybe you can give it a try! Here’s a link to a post: https://hawaiiandna.wordpress.com/2024/02/27/autocluster-endogamy-tool-at-gedmatch-com-part-1/
Dana
Hello Dana, I am so happy to have found you! <3 Thank you for this method and I watched your video as well.
Recently I had my DNA processed on the Ancestry Canada website because I have never known for sure who my father is. My mother told me she was "100% certain" it was one person, who I did get to meet and know my half-sisters through him when I was younger. However, none of them actually believed I am related. To this day, they don't believe. So, I did my DNA on the site I know many of them are on (but I don't think my father is..). A few of his siblings and their kids as well as his daughters are there. I have now learned that either they are not sharing their DNA results with matches OR, they are in fact NOT my family!
Now I am trying to find out who actually is. With no information about them at all.
My question to you is – Could I use this method using only my paternal DNA matches at Ancestry, or do I need to include both sides for best results? The site shows either Maternal or Paternal matches ("Parent 1"/"Parent 2") and I was able to single out my Paternal side via that option. If I did it that way, would I then only need 2 color columns rather than 4?
One other thing I'm finding is that my Grandfather's family on my Mother's side seem to be non-existent there! I find that hard to believe, so now I am thinking my Mother also isn't aware of her bio-father as my Grandmother would never have admitted such a thing. I no longer am on speaking terms with either of them and my Grandfather passed in 2019. That is another thing I need to figure out and I think this method would maybe help with that, too.
Thanks for your time.
Hi, Tia. Thanks for the kind words and I’m glad you found me, too!
When I created the method, you could not see paternal vs maternal. Yes, you could just use paternal. I don’t usually recommend it, though. But you can. 🙂 And in a best-case scenario you would get 2 clusters then. But you could get one or qutie a few! Search my blog for condolidatoin/consolidation.
As far as your mom’s father, I would do the same thing. I would actually suggest you do the method as usual and then try to identify each cluster.
I wish you the best! I know this can be difficult. Feel free to either reach out to me or ask on Facebook at the Genetic Genealogy Tips & Techniques group if you need more help. (Or DNA Detectives!)
Dana
Pingback: Using Excel with DNA test results – Generations Genealogy
Pingback: Using Excel with DNA test results - Generations Genealogy
tldr – Parents only have 4 90-200cM matches and this bottleneck of 3rd-5th cousins. Can I trust 40-90cM matches if my dad seems to have matches on both sides of his family, and has very few paternal matches? (my grandfather’s paternal line and grandmother’s maternal line seem to join together somewhere)
I’m one of those with very, very few 90-200cM matches on Ancestry for my parent’s kits (most of my great-grandparents immigrated between 1905 and 1916). My dad (Polish side) only has 4 within the range, all from the same grandparent line (her siblings). This line is actually Lithuanian-Ukrainian on one side, Polonized Germans on the other (grandmother’s maternal line). My dad also has this 3rd to 5th cousin bottleneck with only 1 4th cousin match (who doesn’t know about his ancestry, either). My grandmother’s paternal line has a large tree, and many matches here, too. Ancestry showed only 150 paternal matches, and 6 “both sides” (all 5th-8th cousin level).
It’s my Polish grandfather’s line that I’m trying to figure out. His mother died when he was 3, and I have almost zero info about her ancestry. My dad’s paternal line is the one whose surname (when I searched with Ancestry’s ProTools – was free this weekend) matches a surname on my grandmother’s Polonized German side. Other matches seem to be Polish-Ukrainian matches, so the Austrian partition (these are also in the 35-66cM range on MyHeritage, the next highest range. There’s a few 70s, but those are 2nd cousin’s grandchildren).
My biggest problem is that it seems like my grandmother’s maternal line makes up over 80% of all matches, and her paternal line makes up another 15%, with my grandfather’s paternal line seemingly making up the last 4.9%, and even then, those matches wind up matching my grandmother’s paternal line (her father’s great-grandparents were from Southeast Poland most likely), the Polonized Germans (I suspect through one of the brick walls), and the Lithuanian-Ukrainian side. There’s an impressive 47.2cM Ukrainian match (Kharkiv region) with 6 segments. They don’t match my grandmother’s side directly, but a “5th cousin of a 5th cousin of a 5th cousin” does. Even my parents match if I look through enough Polonized German cousins.
A lot of Polish-Ukrainian matches moved into areas where my grandmother’s side was from. This is the problem I ran into when trying to do the Leeds Method the first few times – a cousin of a cousin would always lead into the other side of the family.
My mom also has 4, although 2 on each parent’s side, and none of them know who their German or Italian ancestor is. Even with the cousin matches that are on the Italian side, they’re all from the same 3 small villages, making the Leeds Method difficult.
Hi, Jennifer. Unfortunately, people who live in certain countries or who have recent immigrants from certain countries often have very few matches. And, as you are experiencing, this makes it quite difficult to work with!
My best advice would be to work with sets of shared matches. Find the highest match you can for whichever line you want to tackle. Then, look at their shared matches. Study their trees. Build out their trees if needed and if you can. Then try to find connections between their trees. When you can start seeing those connections, that should help you identify how that cluster is related to you. Then you can work on another cluster.
Also, if you aren’t in all 4 big databases, you might do that. You have to pay to test at Ancestry and 23andMe, but you can transfer for free at MyHeritage and FTDNA (FamilyTreeDNA).
Best wishes!
Hi, I’ve just done this with my 90 to 400 cm matches and have only 3 distinct groups. I know who my great grandparents are for each of those groups. We are still trying to figure out who my Dads grandfather was. Is it worth going lower than 90cm or does it just get too messy?
Hi, Shirley. For an initial chart, I recommend stopping at 90 cM or it gets “too messy.” But if you’ve done that and aren’t seeing the 4th group, I do recommend going lower. Another possibility is that these matches are “hidden” with your dad’s grandmother’s group. So you might try “diagramming” those matches – or try skipping the highest match in that cluster and see if it breaks into two groups.
I have a new YouTube channel and this video might help: https://youtu.be/pJSbPdFWevc?si=Lt8KfIaW5L83fq0R
Dana
Pingback: Drop in Meeting - Genealogical Society of Ireland
Hi Dana,
How would I go about this when I don’t really have any matches between the 400-90CM range?
I only have one match that is 94CM, then next one is 74, and then 53, and then the rest 30 or below.
My highest match with 94CM is on my maternal side, but I am trying to find out information on my paternal side as I don’t know who they are.
Would love to know if this method would actually help in my case when I don’t have many close matches.
Hi, Julia. I would suggest this instead: Using hte 94 cM match, look at their shared matches. Then work with the whole group or enough matches to be helpful. If this is on Ancestry, Enhanced Shared Matches might help. I have a couple of videos on that. Here’s one: https://youtu.be/IP6X1Fg7GcE?si=2kksxlQASQMA4NPH
Hi,
I only have 17 people in my 400-90 group. I end up with 5 groups (Ancestry said 3 maternal, 2 paternal) – 2 of the groups only have 1 person in them; no overlaps.
I do have a match at 519cm that I don’t know and she isn’t responding. If I look at her, she matches with 3 people I had grouped that I know are linked to me by my great grandmother. I’ve no idea how I’m linked to the 3rd person and again, no response when I reach out.
What would you recommend as the best next step?
Hi, Alison. I would work with her shared matches and use Enhanced shared matches – a Pro Tool. I have a video that should help: https://youtu.be/IP6X1Fg7GcE?si=2kksxlQASQMA4NPH
And here’s another video: https://youtu.be/DOlKVlQaG-M?si=A83fDE3eDLhhaYw5