Leeds Method: Which Matches Should You Use?
I often get questions about which matches should be used when doing the Leeds Method. For the most part, it depends on what you already know about your family and matches.
Unknown Biological Parent or Grandparent
If you are trying to identify an unknown biological parent or grandparent, you probably don’t know your 2nd and 3rd cousins. In that case, I recommend using matches who share between 90 and 400 cM of DNA with you. This is because we are hoping to find 2nd and 3rd cousins and avoid any first cousins who share two grandparents with you. And, for most people this range works.
(Read this post to understand why we want 2nd cousins but not 1st cousins.)
Known Biological Parents and Grandparents
However, if you know your biological parents and grandparents and are able to identify some of your matches, you should use those matches who share only ONE grandparent with you and avoid those who share TWO grandparents with you.
Above I’ve shared a diagram of me and my first cousin once removed (1C1R), “Dave.” We share 350 cM of DNA. Let’s look at two scenarios:
-
- If I were creating a Leeds Method chart and he was a “key person”—that person who started a color cluster and I used his Shared Matches to create the color cluster—then the chart would be fine. He shares only one of my grandparents with me: Hazel.
- If he was creating a Leeds Method chart and I was a “key person,” then that color cluster would include matches from two of his grandparents: Emil and Myrtle. Instead, he should exclude me on his initial chart.
Why avoid 1st cousins?
When using the Leeds Method, we are hoping to discover four clusters based on our four grandparent lines. First cousins, however, match us on either our dad’s or our mom’s side. If we created clusters based on them, we could potentially only get 2 clusters instead of 4.
Can I add in 4th cousins?
I suggest creating the initial chart first. But then, yes, you can add in 4th cousins. In fact, it’s encouraged!
My question comes when I’m color coding matches through shared matches and I’ll have shared matches, say 20 of them on my blue line, but all of a sudden I get one odd green line person. Would I put both blue and green for the person I’m trying to color code for?
To try and make what I’m asking clearer. I’m looking at George who matches my mom’s DNA (the only DNA tested so far). They have 21 shared matches. 20 of those are from my blue line (Gilpin), but then match 21 is from my green line (Poole). When I color code George am I putting both blue and green or skipping green as an anomaly?
Great question! It sounds like George just “happens” to be related to both color clusters. So I would give him both colors, but I would NOT give his matches both colors. I had one of these and it took me a while to figure out how this match was related to my dad in two different ways. It also made a more distant match appear as a closer match. So keep that in mind!
Best wishes, Dana
At this point I only color code for the matches with a single color, but still different colors. The only time I’ll copy multiple colors for a match is if almost every other match are those exact same multiple colors.
I have a match that I colored teal and shares 28 matches and a little overlap with 2 other colored groups -dark green and light blue.
I have another match that is colored peach and matches only one other person that is between the 400 – 90 CM range. This match is part of the Teal group. Should my peach person be considered a teal match? or no? He has 2 other matches that fall in the 3-4 cousin prediction range that do not match the teal group’s key person.
Hi, Katherine. My best suggestion is to read these two posts: https://www.danaleeds.com/leeds-method-time-to-consolidate/ and https://www.danaleeds.com/consolidating-clusters-example/ And, there’s not necessarily one right answer. You can create clusters and then work with them to figure out how the people in that cluster are related to each other. So it doesn’t really matter if you end up with 3 clusters or 7 clusters – just figure out what the clusters mean.
Hope that helps!
Dana
Dana,
Great example and a picture of this common question. This is an essential concept when first using your method. I still recommend your process to many people before jumping into other tools and also for adding DNA matches from multiple sites into a single combined spreadsheet for easy comparisons and analysis.
Paul Baltzer
Thanks, Paul!
How would one add all of the DNA matches from all sites? Do you sort separately and then figure out which of the four lineages match from each site? And then add them into one spreadsheet? Or is there a way to put all matches together on one from the beginning? I figured that since I don’t have a common match who shows up on each site, it would be hard to sort a big combined list accurately. Maybe I just don’t understand.
So, you are saying, when you click on shared matches, there are none? On ANY of your matches?
Phoebe / Dana,
Was this question ever answered? We are trying to figure out the best way to combine results from Ancestry, 23&Me, and MyHeritage – and incorporate that list using the Leeds Method.
Hi Annette. I usually suggest keeping the spreadsheets for each testing company (and each person) separate. However, if you’ve created 4 groups based on your 4 grandparent lines (or even 8 based on your 8 great grandparents), you might want to create a master spreadsheet. In that case, I would suggest possibly a page for each of these lines. Then you could list all of the matches and have a column for the testing site (or tester). But, I don’t recommend actually trying to create clusters across the different sites.
Hope this helps!
Dana
Agreed. I think too many people try to jump into the more advanced tools before they even know what family groups they have. I also start with the Leeds method for my unknown parentage cases and recommend others do this too. Being able to see all the matches in one place, grouped together and color coded gives me a good starting point for the next steps.
Thanks, Susan. Great points!!
Hi Dana! I’m a DNA SIG coordinator in Australia Our last meeting was on using your method which I think is great for beginners and those not so computer savvy. My comment is abt first cousins removed which happens quite often to us older lot. Often they are under the 400 cM but they make a difference to the shared matches and lines as really they do the same as first cousins – too close!! I’ve told my group if they can identify them – don’t start with them. What are your thoughts??
Hi, Marilyn. Yes, I agree with you! You can always add them back in later. 🙂
Dana
I’m new at this painting. I have 23&me. Only my maternal side. Father and brothers are no longer with us. I do know my parents, grandparents, 1st, 2nd and 3rd great parents on both of my parents side. I do get messages from cousins, usually 2-6th, and are of mixed race that we share a grandparent. How is the easiest way to start painting this? Is there a beginners class? Thanks
I am working on a class and hope to have it out by early summer! 🙂
Dana
I hadn’t heard anything about the class. Is it still in the works?
Hi, Joyce. I am still considering it! The earliest I would offer it would be January. Thanks for checking!
Dana
Hi Dana,
Firstly, thank you for all this information! I have a DNA test on Ancestry which has thrown up lots of matches on my father’s side (also my mother’s, but it’s my father’s line I want to tackle first) but no matches before 3rd cousins relating to his maternal line, which seem to be connected to my 2 x great-grandfather (his mother’s mother’s father). Well, I found a 2nd cousin on my father’s mother’s side, but haven’t been able to make contact.
So should I start the chart with the 3rd cousins? Or should I include my father and grandmother, even though neither ever did a DNA test (passed in 1989 & 1998)? As far as I can tell, none of my cousins on his paternal line have done DNA tests either… so I’ve been trying to follow his maternal line.
And what is probably a really stupid question, where you have Person 1 Person 2, am i right in thinking I’d put Joe Bloggs, Adam Bloggs, Martha Smith, and so on there? Yes, I know it’s a daft question…. (there’s always one, I guess). 🙂
Hi, Elaine. I suggest you start your chart with those 3rd cousins. And, I believe there are no stupid questions! And I really wish I would have used fake names instead of Person 1, Person 2, etc. Yes, you would put your match’s names like Joe Bloggs, Adam Bloggs, Martha Smith, etc.
Dana
Brilliant, thank you Dana! 🙂
Elaine hi.Can you clarify for me how do you know it is your father’s mother father line?Im in a situation like that trying to find my father mother father line who is unknown.
Robert
Did you get an answer to this question? I would love to know it as well.
Dana,
I have a Brick wall and have been diligently trying to use your method for the last week. My Grandmother’s Father is my brick wall. (I was lucky enough to find her mother w/ original Social Security application and verify it with data on Ancestry) I’ve completely loaded from 400 down to 90 cM from my Uncle – her son who did DNA test for me a few years before he passed. He has 125 people between 400 cM and 90 cM and 285 people if I go down to 50 cM! I am only showing (2) Grandparents. The first group comprises my uncle’s Andrews side. The second set is My Grandmother’s Parent’s side. Both have many Surname’s.
I would say the prominent names on Uncle John’s Paternal side: Andrews (31) & Perry(20), 2nd group is Jones (15) and Reddick at (12)
The Prominent Names on Uncle John’s Maternal’s side: FREEMAN (21) /PENLEY (18), 2nd group is Lackley/McClellan at (15) each
How do I determine the correct tier for the father for Sadie Mae Duncan my Grandmother? I know her mother is a Penley. But, I don’t know who her father was.
I did the complete Leeds Method, with 125 names for just from 400cM to 90 cM which is very large. So to see it better I then took the Andrews/Perry’s group on it’s own Excel sheet to evaluate the family breakdown, and the Freeman’s/Penley group on their own excell spread sheet to evaluate the family spread breakdowns. This is how I came down to the Main 4 families for each group. I could just stay with the (2) top but as you see the numbers are very close. I would love any input from you or guidance. Thank you so much. Dona Strong
Hi, Dona. Sorry for the delay! 125 matches between 400 and 90 cM is (currently) on the high side. It’s possible you have pedigree collapse OR that you just have large families. Large families are great, but pedigree collapse can cause difficulties when working with DNA matches. My next step would be to diagram the clusters if the clusters are clear. (https://www.danaleeds.com/visualizing_clusters_2nd_3rd_cousins/) If they’re not clear, then I would suggest using an automated clustering tool like AutoCluster or Collins’ Leeds Method depending on where you tested. (See under “Leeds Method” on my website.)
Dana,
Thank you, I finally know how to start looking at my DNA matches! I got my DNA done two years ago and had no understanding of what to do next other than email people. I have my 90-400 cM matches in a chart and only have 6 colors. Few questions:
1. When you get your class done, I want to take it, send me an email!
2. I have six colors, 36 matches, and 8 overlaps. I think that means, I probably have a column for each of my four great-grand parent couples (looks like it, based on Surnames I’m seeing on trees they have posted). The overlap (2 colors) probably means that the match is coming from either the paternal or maternal side of those great-grand parent couples. I hope that makes sense the way I’m saying it.
Thanks, Stacey
Hello, Im trying to use DNA to help my father in-law find who his dad was. Through DNA testing, we have discovered another man who looks like he might be a half brother. Unfortunately, he too does not know his father and is also searching for information. I have used their ‘shared matches’ to get a list of cousins. There are 2 people who are 2ns-3rd cousins (103cm and 93cm). There is one person listed as a 3rd-4th cousin (75cm) and the rest are 4th to 6th cousins (60cm and less). I’ve tried colour coding and I have 4 people overlapping. The rest are not overlapping. That is all that comes up. What should I do next? Is any of this information useful for finding a common grandparent?
Thanks
Hi, Alwyn. Since you’ve already identified matches you need to work with, you probably don’t need to use the Leeds Method for this research question. Do the 103, 93, and 75 cM matches have trees? Or can you build a tree for them? If so, you then can look for repeating surnames and/or people. You will often have to build their trees down (meaning back towards the present) to try to locate a potential father. You can also use those smaller matches, though it gets harder since they are more distantly related. You also might want to join the Facebook group called DNA Detectives where you can ask questions, read posts about what others have done or what was recommended to them, and possibly get a Search Angel who can help you if you’d like help.
Best wishes!
Do you have any suggestions on configurations to look for clusters that indicate great grandparents?
Hi, Curtis. For the most part, we will not get clusters based on great grandparents. Instead, if you find a cluster for a specific grandparent, then you can look at the Shared Matches (or In Common With) of the people in that group. It is very likely that some will be descendants of the great grandparent couple, but others will be descendants of his ancestors. And still others will be descendants of her ancestors. In this way, you can likely isolate the matches you need to work with your research questions.
Hope that helps!
Dana
Does it matter which match I start with? I am trying to determine an unknown father, and there are many people who I suspect are 1st. Cousins 1/2x removed. The tester has 6 2nd cousins and 14 3rd cousins on paternal side, yet when I use 90-400cm I only get 3 groups and they overlap in too many ways. Parents are not related. Can I start with 90cm matches- working backwards? I think I should get 8 paternal groups for 3rd cousins,which I can deal with, but will it solve the problem of 1st cousins- removed? Or will the result be the same? Thank you for all the great articles!
Hi, Dee. Yes, you can try starting at 90 and working up! Or you can start in the middle. 🙂 It could be that the highest match is actually a 1st cousin of some type. You might also try a Collins Leeds Method chart which can include all of your matches.
Dana
Hello! Thank you for providing information about your method of analyzing matches. I followed your instructions and came up with 9 clusters for my half 1st cousin 3 times removed, 4 of which belong to the ancestor in question (her grandfather). It appears that 32 of 55 matches cluster under her grandfather. What can I infer from this about the matches’ relationships? I suspect that the matches’ ancestors are his children by different women. Is that too much of a leap? Thanks!
Hi, Tee. First of all, some of the clusters might need to be consolidated. (You can search my website using the search feature for “consolidate” or “consolidating.” And your hypothesis might be correct. It is also possible that the key person – or the person who started each cluster – is a 3rd cousin so the people in the clusters are related through one of your great, great grandparents.
Any time you create clusters, the next step is to figure out how the people in each cluster are related to each other. Then you can figure out how the test taker is related to that group.
Best wishes!
Hi, is there anyway I can correct the typos in that post I just sent you?
It should have read: “…. I know the full set of GGG’s I share with him. He is a 55cM match, and Ancestry labels him as a 4th- 6th cousin. He matches to the correct clusters I am trying to workd down to.
So would he be a good research subject to begin with? Further down from him: comes 3, 4th-6th cousins who range from 37cM’s to 44cM.
Does this seem like a workable family grouping for the Leeds method, if I want to get to the 3 DNA clusters who mostly sport 4.4th cousins, that I only share 21-27cM’s with?
Hi, A. (I deleted the other.)
Yes, that would be a good research subject to begin with! And I would work with any of his Shared Matches to try to figure out this cluster.
I don’t quite understand the last part, but you can work with any cluster of people who have Shared Matches in common to try to find out their connection to each other and then your connection to them.
Dana
I am trying to confirm my grandmother’s biological father using my deceased maternal aunt’s DNA. Having used your Leeds method successfully to sort cousins and their shared matches into the 4 colour coded groups to represent my deceased grandmother’s grandparents, I am at a loss at to how to proceed to the next step. Here’s the problem: while her maiden name is prevalent in 1 group, I am unable to discover the other 3 of the 4 the names that would represent my grandmother’s grandparents. My aunt has only one second/third cousin with 149 cM but many possibly 4th cousins. Despite building a tree for the most common colour coded group through shared matches, I am unable to name that group. Only one of the four colour coded groups has a name (her maiden name). How do I find the other group names?
Hi, Elebea. I apologize for missing your comment. My suggestion would be to start by ignoring the surnames you are expecting to find. Since you have the 4 groups, try to figure out how the people in each group/cluster are related to each other. You can work with trees that are already there or you might have to bulid them out farther. You also might look at the Shared Matches of those in the groups if you aren’t finding enough people or trees. And definitely look at higher matches if you aren’t already.
Best wishes!
Dana
My paternal grandmother and grandfather each had children with other partners before they married. How would this be reflected when I undertake my analysis?
Hi, Leslie. That’s a great question and it’s making me stretch my brain this morning! 🙂
If you have half 1st cousins, they would likely show up around 400 cM so might be on your Leeds Method chart. But they would likely be showing as matches to other people on this grandparent side. So, I think it’s possible it wouldn’t change your results. But it’s also possible that you might get 2 clusters for this grandparent. I think it’ll depend on who has tested, how large these families were, and even your age. (I need to blog about how age can affect these charts!)
Hi Dana!
I just read this page and the comments, but I have a question not already answered on this page and I am hoping you can guide me.
My mother’s father was adopted. I have his original birth certificate and adoption paperwork, so I already sorted all of my grandpa’s maternal matches, easy. The birth father’s name and identifying info was a red-herring search and, since discovering that, I have managed to figure out matches that fall on his paternal side.
Unfortunately, the highest match for that side starts at 58cM. I used their shared matches and built out trees for many of them, and I can see they descend from two lines way up the tree where 3 brothers from one family married three sisters from another, and named most all of their children by names repeated in each family line (so many Mosses and Wesgates,) but the trouble I am having is that I cannot seem to pinpoint with certainty the specific branch I need to be drilling down on to isolate my grandfather’s father. I had my mom do 23&Me and there is exactly one match on that side, one generation down from my mom; she only goes by her initials and will not respond to my messages. She shares 101cM with my mom and has just one other match to her at 90cM to my mom. She holds the key to solving this puzzle. I cannot figure out which family line she belongs to.
I already have my mom’s and my DNA uploaded to all the other sites recommended by DNA Detectives, but the matches on those sites are also really distant (less than 90 cM, and mostly less than 50cM.)
Do you have any suggestions that might help me in this search that I maybe haven’t thought of or tried? My mom’s health is failing (she’s 78 and diabetic) and I don’t know how much time I have left with her, so I am hoping hoping I can solve this puzzle as soon as I possibly can.
Thank you for your time!
Hi, Barbara. 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!
Wow! This is definitely a tough one! If you can find descendants of some of these couples to test, that could be very useful. You also might hire a professional genealogist with DNA experience. Someone might be able to help you identify this match that goes by their initials. Or they might be able to do more research and give you ideas of what you can try next.
Best wishes!
Dana
Hi Dana,
A question if I may please? I am brand new to the Leeds Method, but although I have over 2500 names in my family tree, which I have been working on since about 1985, Ancestry shows that I have only 4 cousins within the recommended 90-400 limits, 2 each on maternal/paternal lines. So does that mean that the method can’t work for me until/unless more turn up?
I am trying to track down an unknown maternal G Grandfather and I think that two of those 4 cousins may be descended from the missing man. I also have another slightly more distant cousin who is DNA related to both of those two and to me, but doesn’t fall within the limits. Geographically they are all in roughly the right area at the right time period.
Hi, Philip. In your case, instead of doing the Leeds Method I would work with those matches that you think are related to your unknown maternal great grandfather. I would also work with their shared matches. Then try to find connections within those groups. It sounds like you’re on the right track!
Best wishes,
Dana
Thanks Dana, thats about what I thought you might say, but worth asking the question. 🙂