3 Improvements to the Leeds Method
I have made some improvements to the Leeds Method that make it faster and easier to use. If you are new to this method, you can read the original steps on my Color Clustering post. These steps are still correct, I have just added a few more!
Improvement #1: List all DNA matches through 3rd cousins.
Instead of just listing the 2nd and 3rd cousins (below 400 cM), we can list all of the DNA matches through 3rd cousins. In other words, list the immediate family, close family, 1st cousins, and higher 2nd cousins. This makes it easier to work with and get the most out of those higher matches. And, if you have closer cousins (i.e. 1st cousins), the chart will likely show you which color clusters or genetic networks are on one side of the family vs the other side of the family. (UPDATE: Do NOT use these 1st cousins and higher 2nd cousins to create your sorts! You are still CREATING the sorts based on the first person who shares less than 400 cM with you or the test taker.)
UPDATED NOTE: This step confuses some people. If you find it useful, please use it. If you do not find it useful, please start by listing only 2nd and 3rd (who share less than 400 cM) and THEN add in those closer matches later.
Improvement #2: List cM for each person.
We should have been listing the centimorgans (cM) for each person from the beginning, but I was trying to keep it as simple as possible. You can add the cM in a separate column, like in the examples, or add it in parentheses after the name.
Improvement #3: Add a column to indicate tree vs no tree.
This can be quickly down right after entering names and cM and it saves time later when you need to know who has a tree and who does not have a tree.
Example with these 3 improvements
Above is an example of a chart I made from a client’s actual data. For privacy, I have not shown the names (column A). For simplicity, I am only showing matches down to 133 cM, but the original chart goes down to 90 cM which is the last of the 3rd cousin group on Ancestry.
Using this chart, we can easily see the asterisks (column C) to identify which matches have trees in each color cluster or genetic network. Then we can work on identifying surnames that repeat in each color which you can read more about at “Identifying ‘In Common’ Surnames.”
How does each improvement help?
#1 – If you have higher, already known, matches, you can immediately see where these color clusters belong. In this case, you can immediately tell the maternal from paternal matches.
#2 – Listing the cM helps you see how close the relationship is you. And, if you have someone else look at the chart (like me!), it helps them interpret the chart.
#3 – As explained earlier, this saves time when you start trying to identify surnames that repeat within a specific color cluster group.
Helped an adopted friend- without the Leeds Method I wasn’t sure where to begin. It made all the difference, helping concentrate on one line at a time. A convergence finally became apparent. Thank you! I will also admit I did use 1st cousin matches from the beginning and noted those with trees.
That’s great! I love hearing about when and how the Leeds method has helped others. 🙂
I love these updates when you post them. For the first time, I am able to organize my DNA matches and now, I see a pattern of relationships to the ancestors because of your color coding. Keep those great ideas coming (-:
Yay!! That’s great!
I also enter the cM in the colored cell so if I decide to sort the spreadsheet by color (i.e. by ancestor), the cM are right there to compare.
Elizabeth, That’s a neat idea. I have done it a few times, but not for any specific reason. But, I like your reasoning!
I also indicate the amount of people in a tree. The more individuals, the better. Some people have only a couple of people in a tree.
I never did understand why the high cM relatives were supposed to be ignored
I wanted to also ask: If your previous method is still correct, then do you START with those under 400 cM (and just list those above 400) or do you start with the highest who would then “wipe out” a whole host of lesser cM matches? I’ve tried both methods and still can’t get very far.
Jim, you can email me if you’d like. But, you create the chart based on those who share less than 400 cM by looking at their shared matches.
I would usually have ignored them because they will likely have more than one color. Let’s say they are going to end up being red, green and yellow as in the example. So if you use them to create the groupings, you will end up mixing red, green, and yellow matches into one group. That’s useless… By starting with the matches that have under 400 cM’s (the ones that are likely second cousins) you will pick people who are only related as second cousins – sharing ONE set of great grandparents. So it will be clear what the colors of the others should be.
THEN you can go back and put the three colors into that top match.
Hi I helping a friend who adopted, she tested on Ancestry and upload to gedmatch and family tree dna. Her matches are 1x 2nd cousin, 1x 3rd cousin and 185 4th cousins and then distant cousins. Is your method still effective with 4th cousin matches?
When you indicate add high “4th cousins” – what is the cM range you are referring to? The highest 4th cousin matches showing for my adopted friend are are 57cM, 53, and 50. The next 47, 46, 42 and 41. Then a bunch in the 30s.
Thanks!
I would keep adding them until things get too messy – probably too many columns.
It is!
I am working to find my husband’s grandparents. His closest match is my children- one 3rd cousin who doesn’t cluster with anyone and a myriad of 4th cousins. Do you think I can use the Leeds cluster method if he ONLY has the 4th cousins to go by? Should I go through your steps that you suggested with 2nd and 3rd cousins- but just use the top say 15 or 20 4th cousins. They do show shared matches- so in a sense I could cluster them. Wondering if you think this would work?
I have tried this method pen to paper for my 2nd/3rd cousins under 400cm but I’m unable to get 4 neat columns ,so the finished chart looks messy. I know my maternal lines a little so that helps somewhat but unable to determine paternal lineage with what I have. What am I doing wrong?
Zowie, Although it is great when we get 4 nice clusters, many people do not! It depends on how many matches you have, how well each of your 4 grandparent lines are represented, and if your 4 grandparents have any common ancestors or relatives. Sometimes, it’s a matter of seeing overlap between 2 or more columns and merging them. I’m always happy to give some brief feedback when people send me a screenshot of their chart with the # of cM included. 🙂
I already had my maternal matches seperate from my paternal-so I should just end up with two color clusters correct? One for my grandmothers line and one for my grandfathers line? I ended up with 4 clusters and 2 people who match all 4?! I do know that some of these matches are my half cousins and there children-how are they related to all 4 colors! I am so co fused-not by your system-but my results?
Hi, Sandra. If you are only using maternal or only using paternal, you would *potentially* get 2 clusters for each. But you could get 1 or many. It depends on many factors.
You mentioned your grandmother had several husbands with children from each. Those could form seperate clusters, too.
I’m happy to provide some feedback if you email me an image of your spreadsheet: drleeds@sbcglobal.net
Dana
Thank you for sharing!
Hi Dana,
I am very excited about this method, as my ggp’s on my mother’s maternal side were both foundlings, and we therefore do not know their family names. I have been using my dna matches on Ancestry to try and trace back the trees of my matches to a common ancestor, but it is tedious and I often can’t find the documents I need to extend the trees. My question for you is, can this be used when there is only one 3rd cousin match, and my 1st and 2nd matches are all known to me? I have numerous 4th cousin matches who come from my ggp’s town of origin, but only one 3rd. I also had my mother’s dna done on 23andMe, and she had numerous 3rd-4th cousin matches, but of course on 23 there are no trees, so dead end there. How would you suggest I proceed in this case? Is this method not going to help me uncover my maternal family names?
Hi, Deb. I do think this method can help you. You were first talking about Ancestry – is this with your DNA or your mom’s? If you haven’t tested your mom at Ancestry, that would be my first recommendation. If you’ve already tested her, I’d probably recommend using AutoCluster and running down to maybe 50 cM. Then, you can work with each cluster and start to look for the shared ancestor of that cluster. It sounds like you haven’t tested your mom at Ancestry, so I would do this first! You have only half of her DNA, so she should have a lot better matches.
Thank you! She suffers from dementia so I’m not sure I can get her to cooperate, but it’s worth a try!
Hi, Debra. My apologies… it looks like I somehow missed a few comments. Depending on how many 2nd cousins you have – and where they fall in your tree – it should work well!
Hi Dana,
I’m still struggling with this because I have so little to work with. I had my mom’s dna done on Ancestry. My objective in doing this is to uncover the family names on her maternal side; I have done research and have a tree for her paternal side. On the maternal side, both of my mother’s grandparents were foundlings, so their surnames were not the names of their birth families. So I constructed the table using only those matches from the area of Sicily where her grandparents were born. The problem is, other than a 2nd cousin who doesn’t have any more information than I have, the closest match is a 4th cousin with only 70 cM. I set up the table using only those above 50 cM. The columns seem to sort into 4 groups (I am looking for ggp’s on my mother’s maternal side only, so 4 families), but some of the surnames only appear in one tree, and at least one surname shows up in 2 columns. I’m stumped as to how to interpret this. I’m still unsure about which could be family names. Should I construct a table using lower matches?
Hi! Sorry your post got lost in the holiday schedule. Yes, is your mom still living? If so, I’d have her DNA at all of the sites. You have to pay at Ancestry and 23andMe, but can transfer for free to MyHeritage and FamilyTreeDNA. Hopefully, you’ll find some closer matches that way!
Hello Dana,
I was just introduced to your method at a virtual genealogy conference this month. I’m just wondering if my family dynamics on my maternal side is going to be an issue. I am an only child as was my mother. My maternal grandfather had only one sibling and one niece who had children (we’ve lost touch). His mother’s family is the one I’m trying to trace. She was one of 10 children BUT was the ONLY one to have had children who lived to adulthood. I know who all of my matches are through the 3rd cousin level (all on my dads side). I know two of my mom’s side matches for sure – but they are at a 41 and 43 cM match. There are many in the 40’s and 50’s that I have no idea how we connect. If I try your method from 40 cM on up, any insight as to how reliable / unreliable it will be?
Hi, Janice. Having such small families can make it more difficult, but you might have enough information to work with. If you go back another generation or two, you might have a lot more children. If you’re going down to 40 cM, unless you just don’t have that many matches, I recommend an automated method like AutoCluster. You can find a “getting started” post under my “The Leeds Method” tab. Hope this helps!
Dana, for starting at 3rd cousins, what cM do you recommend starting the first cluster? ancestry’s marked 3rd cm or dna painter’s definition. Thank you for any help. Several of the people I am helping with birth parents are latino and also indegenous people. It is really really hard.
Hi, Mindy. To create this initial chart, I just use the relationships Ancestry is suggesting. I’m not sure if you’re saying the Latino and indigenous populations are with one adoptee or two different ones. If one, it does help separate them out! But, yes, anytime you work with endogamous populations it is really difficult. And, the Leeds Method can help you separate an endogamous group from the non-endogamous parts of a family, but it cannot really sort within a specific endogamous group.
Just heard about Leeds Method on Genealogy Guys Podcast and found it fascinating!! Since I’ve tested with multiple sites and uploaded my DNA to others, should I do one master chart using all the sites or use each sites’ data separately? Would combining them be more confusing or lead to better insights?
Hi, Barbara. I’m glad you found the method and hope you find it useful! I would keep each sites’ data separately for the actual clustering. But, in the end you could create a master chart possibly showing your matches from each site and how they are related to your 4 grandparents (or maybe even 8 great grandparents).
Hi Dana,
Is it possible to use the Leeds method with a LOT of endogamy in tree? I am getting so frustrated trying to sort out my matches. It appears that all my dna matches are related on both sides of my tree.
Thank you for your time and help if possible
Hi, Marika. Unfortunately the Leeds Method does not work well with endogamy. With endogamy “everyone is related to everyone else,” so the DNA doesn’t really help us separate out our matches into different parts of our families.
You might read this post: https://www.danaleeds.com/can-the-leeds-method-help-me/
Also, if you have a subscription to Legacy Family Tree Webinars I suggest this presentation: https://familytreewebinars.com/download.php?webinar_id=1292 (It’s titled “Dealing with Endogamy” and is presented by Paul Woodbury.)
Dana
I saw your recommendations for when you have 2 clusters with a lot overlap. My question is, if you are able to identify the root** or I guess potential root for the 2 overlapping clusters, does it then make sense to split them out again? Essentially, at this point I am trying to work down to my paternal grandfather from my most recent common ancestor I identified for my clustered column.
Hi, Lauren. Although I often consolidate, you definitely don’t have to! And if you are seeing a reason why the 2 clusters formed separately, I’d keep them separate.
Best wishes,
Dana
I have columns as well for MRCA; name in my tree; actual relationship; notes; and pondering age when I can get it.
Those are all great additions!
Hi-a little off topic-but thought I would ask anyway! So, I filled out my x matches on a match chart, or who they should be-and out of the 32 spots-I have only 12 differant last names! And it’s full, no more spots, no more possibilities. My question is-what do you think the x-inherited from my father would look like?
Hi, Sandra. I’d love to see more of what you’re doing. Can you email me at drleeds@sbcglobal.net?
Thanks!
Dana
So I understand using 2nd and 3rd cousins but what about all the 3-4th cousins? Do we use those?
Dianne
Hi, Dianne. I don’t use them for the initial chart. But once it’s created you can add in those lower matches.
Hope this helps!
Dana