Using DNA testing and Gworks to Solve Unknown Parentage Cases


All my slides are at

Step 1: TEST your DNA:

at Ancestry and 23andme then transfer to the others

Upload the Ancestry DNA results to other sites for free to get more matches:

  • MyHeritage*
  • Family Tree DNA **
  • GEDmatch
  • LivingDNA (primarily British, not many matches yet)

**$19 for the full tools

*  $29 for the full tools

How does testing your DNA help?

You get half your DNA from each parent, about a quarter from each grandparent and so forth if enough biological cousins are tested you can figure this out

More likely you will need to figure it out from cousin matches

Chart from the ISOGG wiki  (courtesy Dimario, Wikimedia Commons)

The "G" trick

The cousin level = the number of "G"s , else the greats plus one

If you are in different generations take the shorter path and the other is removed by the generation difference

If your roots are deeply American, you may only need to test at Ancestry

Any of these matches may resolve the case for you but proceed carefully

Enter the cMs and get probable relationships

This online calculator that gives you the probable relationships is an invaluable aid

The basic methodology involves analyzing and building the trees of your closest matches ...

Look through the family trees of 2nd, 3rd, and 4th cousin DNA matches for common ancestral couples.

Build private, unsearchable family trees down from each common couple to find where they intersect and produce someone in the right place at the right time.

This may only get to the grandparents  or great grandparents

Clustering can be a quick shortcut to finding the common ancestors. The idea is to display your matches who match each other in an easy to understand visual format

You can sign up to cluster at or you can use the clustering tool from

The result might look like this, each colored box is expected to consist of matches descended from one set of your great grandparents

On some versions of clustering, all the names are clickable to the match (others have a list below)

Here is the cluster image for Jack who was looking for his unknown Dad

The top two boxes were each paternal and the top one was almost all one surname

The clustering methodology was developed by Dana Leeds who has much advice on her site -

There was an unusual surname several times in the 2nd cluster so I searched the box one tree for that name and found one couple in recent times where surname 1 married surname 2.

They had 2 sons, one was in the right place at the right time!

A 2nd cousin match from cluster one had a huge tree

Once you have a candidate parent build their pedigree tree of back to at least 1800

Make sure this tree is private and unsearchable initially but once you are close you can make it searchable so that ThruLines can use it for you

Add a fake child to that parent and then connect your DNA to that fake person

Ancestry will use auto complete in the name field

Select the pedigree tree you created and link your DNA test to the fake child

Now wait a day or three for the hints and see if they are consistent with your hypothesis

If you have the new Beta Ancestry match view, those shared hints do not update, you have to look for shared ancestors on the match page

The match page in the Beta Ancestry view

Check the cMs of each match and compare them to the charts

Do they fit your theory?

Or you can use the "What Are The Odds" tool at DNApainter to check your theory

Is there automation for comparing trees to find common ancestral couples?

Collects and compares the trees of your matches into your own private database at



Find the common ancestors from your Ancestry matches and builds a tree!

DNA2tree currently only works on iPhones and iPads and is a paid subscription

sample Ancestry tree built by DNA2tree

For this to work well, you need many matches with good trees

In order to use the tree gatherer you need to be a paid member

GWORKS is available at

Find the app called DNAgedcom Client on your computer or look for this icon

Once you are logged in to your account at DNAgedcom you will see this screen

Every person you are working on needs their own account at DNAgedcom but you can use your one paid account to collect the data
Use the GWorks Client (DGC) to gather your Ancestry match data

Enter your ancestry username and password

Fastest is to check Quicker and Skip Distant Cousins

Once you are logged in, the name(s) of people who have shared DNA with you show up in the profile box, select the desired one in the dropdown

Click Gather Matches, when it finishes click Gather Trees, when done You do not need Gather ICW for GWorks it is used for clustering

Now your database is ready!

For adoption cases, I usually start by looking for common ancestors born after 1800

by putting 18 in the box at the top of the Birth Date column and clicking StartsWith

Another useful search is to look for a surname, since spelling varies I often start with the first 3 or 4 letters

Click the arrow next to the name so it points down and all the matches who have this person in their tree will show

Clicking on the tree name under the ancestor in Compare All Trees takes you to the DNA match
Having the same ancestor in 3 or 4 trees may not be significant if the trees are all from the same user

You can use the "Show Trees " on the Gworks page to delete all but one and then rerun the matching

Jane had a second cousin match at Ancestry with a good tree

Can  GWorks can identify which line from the second cousin match is the one they are related on?

Top ten ancestors in GWorks

Can you find any GWorks listed names in the far right of the pedigree?

The Pratt Cranford marriage has 8 children

Three of them have DNA tested descendants, check the abbreviated McGuire diagram for clues

3 children


Surname frequency can help you decide which children's trees to build down in an unknown parentage case

So the next step was to look for the surname of Pratt B's wife and then for child D's spouse

Targeted Testing

Child D's spouse's surname is in the database many times, now to get a child of D, a possible parent to test

These are the names of the original Spanish settlers of New Mexico
These numbers for the tree matches scream out ENDOGAMY

Top ancestors for an unknown father case ...

People with endogamous roots do best to wait for close matches, 2nd cousins (or better) ... Tessa had three, one at each site. She also had been told a surname for the father: PADILLA

23andme match

The 23andme match was the grandchild of (Jose) Pasqual Padilla, sharing 5.1%  or 368 cMs.


Based on other matches the 23andme match was confident that he was the half 1st cousin of the Padilla father we were looking for; that Tessa was descended from the 2nd wife of Pasqual,

thus a half first cousin once removed -1C1R 

The Ancestry and MyHeritage matches shared common ancestors: surname MARTINEZ

23andme match

The Ancestry match - 287 cM - had a tree with just parents. Building that tree in our research tree we found the great grandparents of the My Heritage match -282 cM

Eugenio and Maria Martinez

but they were the great great grandparents of the Ancestry match. The son of that couple who was the Ancestry match's grandad had several wives and many children - so perhaps another  half 1st1R?

Obituaries are a great source for finding the names and parents of the living. This one from the Albuquerque Journal found the Martinez-Padilla marriage we were looking for

Next task was to see if we could find a daughter of (Jose) Trinidad Martinez married to a son of (Jose) Pasqual Padilla

So did Pasqual have a son named Luis?

23andme match
Ancestry match
9 Sons

Yes Pasqual's son Luis married Teresa Martinez! the daughter of (Jose) Trinidad Martinez

Now which of those 9 sons was in California at the right time?

Only 2 of them 

Meanwhile Julie, a daughter of one of the possible half brothers, tested independantly at 23andme and shared 774cM (about 10%)

with Tessa

Now Tessa showed off her sleuthing skills, she called the librarian of the small town that most of these Padillas were from and got copies of their obituaries and lots of information. Neither possible father was still alive but she tracked down a son of each one of them and talked them into testing

So Julie is either a half niece or a 1C1R

image from Blaine Bettinger's chart 2017

Here are the tests results for her possible half  brothers

Tessa was welcomed into her father's family; here she is with her half brother and half niece

French Canadian!

Another example of endogamy, recognize this one?

You can use a half sibling or first cousin match to isolate one side by starring those matches

The Genetic Affairs clustering tool lets you cluster JUST YOUR STARRED MATCHES

DNAgedcom's client and GEDmatch both have clustering now too

The profile CSV file from clustering can be used for tracking your work

The m_ file is a useful spreadsheet for tracking your research
You can upload GEDcoms from other sources and add them to your database at DNAgedcom

Some adoption searches end happily - new dad Billy has sold his house and has moved near to daughter Cheryl and his sister in Florida

To learn more take a class at