What’s in Your Toolbox? — DNA Painter Shared cM Project Tool

Source: DNA Painter Shared cM Project Tool
Tool Type:
website
Purpose: rank possible relationships given the shared amount of DNA
Cost:
free
Company Compatibility:
 all
Operating System(s): all

 

DNA Painter is a website with some essential tools for genetic genealogy.  In fact, they are so valuable that each one deserves its own time in the spotlight. The simplest is the Shared cM Project Tool.  It’s an elegant, visual way to evaluate possible relationships for an unknown match based on how much DNA they share with you.

 

Features

The Shared cM Project Tool has two main features:  it visualizes which relationships are possible given how much DNA two people share, and it tells you which of those relationships are most likely.

 

Installation

No installation is necessary.  Simply go to the website:  https://dnapainter.com/tools/sharedcmv4.

 

The Layout

The DNA Painter Shared cM Project Tool has three main areas:  (1) A field in which you enter a shared amount of DNA in centimorgans (see below for percent DNA), (2) a table showing which relationship categories are most probable, and (3) a relationship chart.  The probability table will appear after a value is entered in the field.

The graphic above shows the tool before you’ve entered a value in the field (labeled ① above).  The colored chart is imposing at first, but it’s easy to read with a little background information. 

You are shown in the white block in the center marked “Self”.  Moving down are your own descendants (child, grandchild, etc.) in blue, and moving upward are color-coded blocks for your parent (grey), grandparent (green), etc.  Extending down from those blocks are the relationships to you through that particular ancestor, with half relationships on the left and full relationships on the right (except for half-siblings and full siblings.  I don’t know why they’re both on the left.  I wasn’t consulted!)  Everything is color-matched, so your descendants are in blue, your parent and their descendants are in grey, your grandparent their descendants are in green, and so on.

Each box includes the name of the relationship, the average amount of shared DNA, and the know range of shared DNA.

 

Using the Shared cM Project Tool

Using the tool is simple.  Simply enter the amount of DNA you share with a match into the field (labeled ①), and the tool will do the rest.  Let’s consider an example of a mystery match who shares 157 cM.  Start by entering “157” (without the quotation marks) in the field.

The screen will change in two ways.  First, a table with probabilities will appear where previously there was an empty area (labeled ② in the figure above).

If your match is at 23andMe, which reports shared DNA in percent rather than centimorgans, click the teeny green text that says “show %” below the filter field.  A new field will appear in which you can enter the percent value (without the percent sign) as reported by 23andMe.  The tool will convert to centimorgans for you.

 

The second change in the appearance of the tool will be in the relationship graphic:  the boxes for relationships that are ruled out for a match who shares 157 cM will fade, making it easy to see at a glance which relationships are possible.

 

The DNA Painter Tools Is More Versatile than a Table

You could get some of the same information from a table, of course, like the one provided by Blaine Bettinger’s Shared cM Project.  Data from this project are crowd-sourced from tens of thousands of self-reported values.  (You can submit your own data here.)  In fact, DNA Painter’s Shared cM Project Tool is based largely on numbers from Blaine Bettinger’s project.

Clusters of relationships that are genetically similar (like a half first cousin and a first cousin once removed) are grouped together, with the average shared amount (440 cM for Cluster 4), 95th percentile range (235–665 cM), the 99th percentile range (131–851 cM), and the theoretically expected amount (425 cM).  The 95th percentile includes 95% of the reported values, and the 99th percentile includes all but the most extreme 1% of outliers. 

Tables like this show the average and range of shared centimorgans for given sets of relationships.  The interpretation is not always straightforward, though.

Consider our example from above.  A shared amount of 157 cM is closest to the average for Cluster 6, but it is within the 95th percentiles of Clusters 5, 6, and 7 and the 99th percentiles for Clusters 4, 5, 6, 7, and 8.  That means the DNA match could be anything from a half first cousin (half 1C) to a third cousin once removed (3C1R), with a number of possibilities in between.

But which of those relationships is most likely?  And how much more likely are they than the other options?  That’s where the probability table of the Shared cM Project Tool really shines.

 

Interpreting the Probability Numbers

In our example, a match who shares 157 cM could be any of the 21 relationships that are highlighted in the graphic.  That’s too many choices!  Fortunately, not all of them are equally likely.

More than half (≈53%) of matches who share 157 cM will fall into the group that includes half 2C, 2C1R, half 1C2R, and 1C3R.  That’s the most likely relationship cluster, but others are possible: roughly 27% of such matches will belong to the group that includes 2C, 15% to the one that includes 3C, 5% to the one that includes 3C1R, and less than 0.2% to the one that includes 1C1R.

It’s important to remember that “most likely” is not the same thing as “definitely”.  After all, someone’s got to be in the 27% of 157-cM matches who are 2C range and the 5% who are 3C1R range.  Also, factors like endogamy can affect the probabilities (see below).

To explore the connection to a match like this, I’d start by looking at the most likely category while maintaining a lot of mental flexibility about the other options.  I would use shared matches/ICW to narrow down which branch of my tree to focus on, build out the other person’s tree, and contact them to ask for their insight.

 

Bonus Feature

A nice bonus feature for those of us who assist and educate others is the green link directly above the probability table.

When you click that link, the URL (web address) for the tool will update so that you can send someone a link to the page with the numbers already filled in.  For the example above, the link is https://dnapainter.com/tools/sharedcmv4/1700.  I use this feature a lot when helping people understand their DNA results.

 

Limitations

A key limitation of the Shared cM Project Tool is that it’s not meant for cases with pedigree collapse or endogamy. In those situations, the amount of DNA two people share is often higher than expected based on their closest known relationship, because they are related in more than one way.  As a result, the tool will often favor closer relationships over the real ones.

If you know you are from an endogamous population, you can adjust your expectations when using the Shared cM Project Tool toward more distantly related cousin categories.  For example, I have a match at AncestryDNA, B.G., who shares 149 cM with me.  The most likely relationships are the 2C1R group (53.9% chance), the 2C group (23%), or the 3C group (17%).  Only 5.5% of such matches are in the 3C1R range, but that is, in fact, how B.G. is related to me.  She is also a 4C, a 4C1R, a 4C2R, a 5C2R, a 6C1R, and a 7C, all of which can add to the amount of shared DNA.  B.G. is not an outlier; I see this pattern all the time on my Cajun (endogamous) side.  When I use the tool for my maternal matches, I mentally note that they’re probably more distantly related than the tool might suggest.

One other limitation is specific to Family Tree DNA (FTDNA), but fortunately, there’s a workaround.  FTDNA includes small segments—as small as 1 cM—in their matching algorithm.  Such segments are likely to be false positives that are not reflective of shared relationships, and they can increase the apparent amount of shared DNA by 20% or more.  To circumvent this bias, you can view the “Detailed Segment Data” for that match in the chromosome browser at FTDNA and manually subtract out the segments that are less than 7 cM.  Use the adjusted figure in the tool.

 

Posts in the Series (so far):

7 thoughts on “What’s in Your Toolbox? — DNA Painter Shared cM Project Tool”

  1. I am a bit confused.. 23 and me says half sibling (1903 cm, 25.6%) including a shared X chromosome. On the DNA painter it says grandparent/aunt/halfsibling..how can you determine which one it could be?

    1. If you are both female and share the entire X chromosome, chances are that you share a father. Technically, it’s still possible for the relationship to be grandparent or aunt, though. You can usually rule out grandparent based on age. Aunt/uncle can often be ruled out by comparing how much DNA you both share with your mutual matches. If you are half siblings, you should share the same amount on average, whereas if one of you is in an older generation (e.g., aunt), you would share on average, twice as much with your mutual matches.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.