MODULE 2: Grouping your AncestryDNA matches - The Leeds Method and my "twisted" Leeds Method

These notes are for participants in the Society of Australian Genealogists' program, Analysing your AncestryDNA results. Others are welcome to use these notes for their personal research. Please contact me at chrisw9953[at]gmail[dot]com for other uses. I'll strive to update the notes as there are further developments or my understanding of the area grows.

September 2024 - There have been recent updates at AncestryDNA including the optional subscription to ProTools. While my approach to grouping remains the same, I am now trialling slight changes to numbering and colouring to take advantage of these changes. For examples, groups are numbered as 1P and 1M rather than P1 and M1 as the number and colour of the group signifies the ancestral line. This post will be updated prior to the Summer 2025 program.

Introduction to the Leeds Method

Click HERE to watch a twenty five minute video I've recorded on the Leeds Method and my "twisted" Leeds Method.

Remember that many of us have two predigrees we research - our social/known family and our biological/genetic family. In 2018, a US study estimated that one in four people who take a DNA test discover an unexpected parentage event between themselves and their great grandparents ie 25% between you and your closest fourteen ancestors. Genetic genealogy is an important tool for every family historian and cannot be ignored. Knowing your tree and knowing your matches is the key to breaking down brick walls and, for many of us, there are many surprises along the way.

The Leeds Method is a process for grouping your DNA matches by using your second to fourth cousins to assign them to one or more of your four biological grandparents. Dana suggest that you commence with the first match below 400cM and work down to 70cM. I take it down to 40cM, at least, and suggest you do the same.

Grouping your matches using the Leeds Method is not 100% accurate but it’s a good start. Endogamy, close relationships (like two brothers marrying two sisters) and small matches are hard to correctly categorise. However, I’ll guarantee you that you’ll always learn something from using the Leeds Method.

The Leeds Method is easy to prepare and is a visual guide that can help everyone who takes a DNA test. I’ve used it for adoptees, people searching for biological fathers, grandparents and as far back as 2nd great-grandparents with success. If your can't get your four grandparents now, keep grouping your matches as they come along for breakthroughs or find people who are prepared to test. Once you have your grandparents, you can group further back. I group to my 2nd great-grandparents.

You may have matches or groups of matches that you can't fit into an ancestral line. You can then investigate these matches to see if you can find MRCAs amongst groups of matches and how they might fit into your ancestral line. Remember that you must prove your pedigree at each generation and there could be cases of misattributed parentage for you and/or the matches. Always be openminded as we know DNA doesn't lie but we need to know how to interpret it.

The Leeds Method works well when you're unsure of the relationship with any of you DNA matches. So it's an ideal tool for adoptees. What if, thought your family relationships and traditional research, you know some of your DNA matches?

This is where my "twisted" Leeds Method is used.

I'll present the approach and grouping method I've used successfully but you can use any grouping system that suits you provided you're systematic in it's use.

Setting up AncestryDNA for grouping

To allow Thrulines and Common Ancestors to operate-

upload or build a searchable public or private tree at Ancestry,
take your tree back to your grandparents, at least, if known,
record any unknown ancestors in your pedigree tree as "Unknown" or "_____" and
attach your AncestryDNA test to the tree.

If you have access to other AncestryDNA tests as a manager or collaborator, you can attach them to a tree on your account. More than one AncestryDNA test can be attached to an Ancestry tree but make sure they're attached to the right person.

Why use AncestryDNA groupings?

Using AncestryDNA grouping to cluster your matches builds on your known matches to the possible relationship with unknown matches and hopefully breaks down those brick walls. My three prong approach to achieve this:-

Grouping AncestryDNA matches
Building private, unsearchable "quick and dirty" research trees at Ancestry looking for Most Recent Common Ancestors (MRCAs) for your matches
Developing "What are the Odds?" (WATO) trees at DNAPainter to record your findings

Developing your groups

Think about how you want to set up your groupings
Different goals require different groupings
Groups can be edited or deleted at any time
Trial and error is a good way to learn
There’s always someone to give you a hand
Primary groupings are the ancestral P and M groups
Additional research or focus groups can be used up to a total of 24

My preferred ancestral groups

If you are new to grouping your AncestryDNA matches, my suggestion is that you first group to paternal P0 and maternal M0 and then move to divide them into your four grandparent lines - P1/4, P5/8, M1/4 and M5/8. You can then divide these groups into your great-grandparent and then second great-grandparents.

If you're comfortable with grouping and know your ancestral tree and all (or most) of your second great grandparents, you can start grouping to the P1 to P8 and M1 to M8 groupings shown above. A process to do this is covered in Scenario 3. Or you could attempt to set up the 16 groups and see how successful you are in grouping your matches to those groups. Remember, that you can always delete your groups and start again!

Key message: ORGANISE YOUR GROUPS SYSTEMATICALLY!

If you're using the P0 to P8 and M0 to M8 approach, you can use any combination above BUT make sure that all are covered even if you label a group UNKNOWN rather than with a family name.

AncestryDNA's SideView

In late 2022, AncestryDNA commenced using its impressive SideView technology to divide your matches into Parent 1, Parent 2, Both and Unassigned "buckets". Read more HERE. See if you can use your known matches to edit Parent 1 and Parent 2 to Paternal and Maternal. You can also review the larger Unallocated matches to assign them to a side based on their shared matches.

Scenario 1 - Who are my biological parents?

Goal - Have a name in each box in the pedigree and matches grouped to ancestor/s

Go to the first match below 400cM, assign the match to a group and assign all the shared matches to the same group - don’t worry if matches end up with multiple dots
Then do the same for the next ungrouped match down to 40cM - don’t worry if matches are in more than one group
Now go back to the first group, examine trees and see if you can find MRCAs using “quick and dirty” research tree
Prepare a WATO tracking matches to MRCAs
Repeat the process for each group while looking for cross over of groups
Can be a long process to resolve the pedigree……

Scenario 2 - Who's my father?

Goal is to find biological father-

by dividing DNA matches into potential paternal and maternal lines
using known maternal matches and their shared matches
to infer paternal matches.

My approach to grouping matches in this situation-

Set up one group for the potential paternal matches (inferred as not maternal) and four groups for each of the maternal great-grandparents - P0, M1/2, M3/4, M5/6 and M7/8.
For all matches with “Common ancestors”, verify relationship and tag to group/s - M1/2, M3/4, M5/6 and M7/8.
For matches only tagged to one group, tag shared matches to the same group.
For untagged matches over 70cM with trees (linked or unlinked), see if you can quickly establish maternal MRCA and tag and, if only one group, tag all shared matches to the same group.
For untagged 3rd cousin and close matches, start at the highest untagged match and tag that person and shared matches to unknown “potential” paternal groups.
Look at the 3rd cousin and closer “potential” paternal matches with a tree and see if you can find common family names to concentrate on that branch for the MRCA.
When you’ve found the MRCA build them into the Q&D tree, track in WATO and tag match to MRCA group.
Iterative process until you crack it and, if not, plan the way forward.

Scenario 3 - I know my grandparents but how am I related to all these matches?

Goal is to-

"prove" your genetic pedigree and
identify MRCAs between your matches
to break down brick walls
by using an iterative process
that results in your matches being grouped into your sixteen 2nd great-grandparent lines.

My approach when commencing to group matches-

See if Parent 1 and Parent 2 can be edited to Paternal and Maternal and group to P0, M0 and UK
Set up groups for 4 grandparents (P1/4, P5/8, M1/4, M5/8)
For 4th cousin and closer matches with “Common ancestors”, verify relationship and tag to group/s
For matches only tagged to one group, tag shared matches to the same group
For untagged matches with trees (linked or unlinked), see if you can establish MRCA and tag and, if only one group, tag all shared matches to the same group
Examine the shared matches of all untagged matches of 40cM and above to see if a group can be established and, if not, organise them into unknown groups.

From here, you can divide the groups further until you get to the sixteen 2nd great-grandparent groups.
Once you're confident with this method of grouping and "know" the pedigree, you can break straight to the eight great-grandparent or sixteen 2nd great-grandparent groups.
Remember, if you don't have any tagged matches on one or more lines, you may have misattributed parentage in your pedigree.
Also matches over 40cM that don't fit into one of your P1 to P8 and M1 to M8 groups should be explored as you may have misattributed parentage in your pedigree.

Here's a summary of the step-by-step process that I generally follow to group AncestryDNA matches directly to the four grandparents-

Important note - Steps 6 to 9 are for the allocation of matches that are UNASSIGNED to P0, M0 or UK and you need to select the UNASSIGNED group at the start of each of these steps. Slide will be updated shortly.

Here's a summary of the step-by-step process that I generally follow to group AncestryDNA matches directly to the 2nd great-grandparents-

Here are my groupings where you'll see that the paternal and maternal groups all appear from 0 to 8. I use special symbols like plus signs and underscores to bring the additional groups together. Unless you group in such a systematic manner, you will almost certainly miss that vital match. If you have unknown groups, name then _MUK#, _PUK# or UK#.

Chrome extension - Click all checkboxes

As demonstrated, this is very useful for tagging or untagging long lists of shared matches to one or more groups. You can get the extension HERE.

The eureka moment!

You've grouped your AncestryDNA matches, built trees for your matches looking for MRCAs between them and recorded the matches in WATO trees and suddenly you see two lines coming together as shown in this chart...EUREKA!

Example of that eureka moment!

Sue's looking for her birth father but the process is equally applicable to all scenarios and back in your ancestral line. I'm still trying to work out if this chart was presented by Leah Larkin or Roberta Estes but I've had it in my toolkit for a while now.

Sue has matches with H1, H2, Abbie, Charlie and Bert. Grouping, researching and recording reveals-

H1 and H2 are in the same group. When we build our "quick and dirty" tree, we find the green couple are their MRCAs. They are first cousins 1x removed on the green line.
H1 and Abbie are in the same group. When we build our "quick and dirty" tree, we find the blue couple are their MRCAs. They are second cousins 1x removed on the blue line.
H2 is in the same group with Charlie and Bret. When we build our "quick and dirty" tree, we find the pink couple are their MRCAs. H2 and Charlie are first cousins 2x removed on the pink line. H2 and Bert are third cousins on the pink line.

The blue and pink lines come together when the son of the blue couple has children with the daughter of the pink couple. Provided we can't find any other possible common ancestors, it's looking like Sue's descended from one of the children of that couple. In this example, I'm showing her as descended from a son of the green couple.

Suggested groupings for close relatives

If you collaborate or manage the AncestryDNA test of a parent, sibling or other close relatives, you can also sort their matches to P1 to P8 and M1 to M8 groups. Some suggestions are shown below. Also, remember if you have tested a parent, you should tag their matches to the recommended P1 to P8 and M1 to M8 groups and only tag the matches that you don't share with that parent. You might be interested in using the tagging function at Genetic.Family which does carry tags across AncestryDNA tests that you can access through your Ancestry account and also across testing sites and GEDmatch.com. Please note that the cross site tagging is not 100% reliable but can be very useful.

Start from your highest unknown match or group of matches to make those discoveries!

My "twisted" Leeds Method groupings

Here are some groupings that were initially developed for the February 2021 presentations. You should develop your groupings based on your DNA dilemma and the AncestryDNA tests you have access to as a collaborator or manager.