Monday, August 4, 2014

My First Look at GEDMatch Autosomal DNA Analysis

I have had autosomal DNA testing done at 23andMe, FamilyTreeDNA and AncestryDNA, and reported on the results.  They all told me my ethnicity estimates, and they all differed one from the other.  On 23andMe and FamilyTreeDNA, I can see a chromosome map showing my chromosomes compared to other testers on the same autosomal system.

Unfortunately, very few testers on 23andMe and FamilyTreeDNA have uploaded a GEDCOM file to those systems, so I cannot compare my GEDCOM file with those of many of my matches.

There are other DNA analysis tools that I am just starting to investigate.  Several months ago, I downloaded a spreadsheet file of my autosomal DNA test results from 23andMe.  FamilyTreeDNA and AncestryDNA have this feature also.  I then uploaded it to GEDMatch (www.gedmatch.com) and waited several months for the results.  I checked it again today, and saw that I could use the site.

Here is the GEDMatch.com home page - I signed in and saw:


This page tells my GEDMatch kit number (in this case, it's my 23andMe kit number) in the left-hand column.  In the right-hand column, there is a section to upload my DNA raw file (which I've already done).  Below that is the "Analyze Your Data" box with links to:

*  'One to many' matches
*  'One-to-one' compare
*  X 'One-to-one'
*  Admixture (heritage)
*  Admixture/Oracle with population search (New)
*  Phasing
*  DNA File Diagnostic Utility

On the right side of the right-hand column is a section for GEDCOM files (I haven't uploaded one yet) and Miscellaneous utilities.

For this post, i'm going to look at the 'One-to-many' DNA comparison and the 'One-to-one' comparison.

Here is the screen for the 'One-to-many' comparison - I needed to enter my kit number:


After clicking the "Display results" button, I saw the top of the page for "DNA Matches to Kt No. M......):


On the screen above, there are several buttons - one to see the Chromosome Browser comparison, another to see an Autosomal Matrix comparison, and a third for an X-DNA comparison.

Further down the screen is a table of autosmal DNA kits that have some matching segments (at least 7 centi-Morgans).

The columns show the kit number, a "select" check box, sex, Mt and Y haplogroups, Autosomal DNA (total cM, largest cM, and generations), X DNA (total cM and largest cM), kit owner's name, and the kit owner's email.

When the user first sees this table, it is ranked with the highest Autosomal Total centi-Morgans.  The table can be ranked by choosing any of the blue up or down arrows in the columns:


Note:  I truncated the graphic above to hide most of the kit owner names and email addresses.

The user can select (check) at least two of the profiles to be compared to his/her own profile.  I chose the first ten on the Total autosomal list above (two were duplicates), and scrolled up and clicked to see the "Chromosome Browser" button.  Here's the top of the page:


The graphic tells me the person(s) that I match on each chromosome, and defines the matching segments, and any overlap with a previous match on the list.

Further down is the persons that I match on Chromosome 15:


As you can see, five of the kits match me exactly on this chromosome with over 50 cM.  While that sounds like a lot, it's typical of only 4-5 generations back - in other words, it's 2nd-3rd great-grandparents (I think)!  I've been in contact with the kit owner on FamilyTreeDNA and we think we match on the Hill family in Wiltshire, England.  My third-great-grandmother is Rebecca Hill (1790-1862).

 I looked at the other buttons on the DNA Matches page above, but haven't shown the results.

I wrote down one of the kit numbers above, and my own kit number, and went back to the first screen above, and clicked on the "One-on-one" comparison.  That opened with this screen:


On the screen above, I entered the two kit numbers, checked the "Show graphic bar for each Chromsome?" Yes button, and clicked on the "Submit" button.  The top of the resulting screen looked like this:


The green bars indicate a full match, the yellow bars a half match, the red bars are no match.  There are other indicators.  I'm not sure what all of that means...but I think that green bars are desirable and yellow pairs are good, but red pairs are bad.

Apparently, if I find a lot of green and yellow together on a chromosome, that means that there is a significant matching segment on that chromosome.

Further down the screen, I saw:


Look at Chromosome 15 - there is lots of green and yellow, all in a row!  The small table above the color bar tells me that there is 54.4 centi-Morgans in that segment.

I can do a similar exercise on the X-chromosome results from the DNA Match table.

Now, I think what I want to do next is to make an Excel spreadsheet of the matching chromosome beginning and ending locations, along with the kit number, the person's name, email address, and number of generations, family tree location, and probable matching common ancestors.

There is a lot more on this website, and I'm still exploring it. I haven't looked at all of my matches, but I think I see at least one other genealogy blogger.

I know I've missed some of the website features also, and here, perhaps, my readers can help me out.  What other features should I be using?   Should I add a GEDCOM file? What should I put in my Excel spreadsheet?  What other third-party websites should I be using?  Please help me with links to other blog posts or articles about these analysis tools.

The URL for this post is:  http://www.geneamusings.com/2014/08/my-first-look-at-gedmatch-autosomal-dna.html

Copyright (c ) 2014, Randall J. Seaver

7 comments:

Anonymous said...

1) Currently GedMatch is still recovering from moving server so many features are not available -- many of us are patiently awaiting their return, such as the segment triangulation.

2) Rather than build your own spreadsheet, consider using Genomemate (www.genomemate.org) -- a user-created program/database that can collate data from all the testing companies and Gedmmatch and help with statistical analysis as well ass tracking email contacts with matches. I am a proud user of it.

3) Yes, upload a Gedcom file. In the past you used to be able to go to a linked Gedcom from the one-to-many report -- hopefully that will return when they fully restore the website.

Elizabeth
(in charge of kits A131521, A393287, A501757, A658108, and F278130)

Fax said...

I agree with Elizabeth's suggestions, but there is another especially useful feature: DNA GEDCOM matching. It produces a list of people you match who also have matching people in your GEDCOM and theirs. It's a relatively "short list" that is a great starting point. I used it yesterday but then it disappeared when I returned later, so it's one of the features they're obviously working on.
Doris

Jen A said...

Thank you for sharing this Randy! I have been slowly dabbling in DNA analysis and learning to use Gedmatch was my next step. I look forward to all the great comments that will come from those who use it regularly. You should also know that there is a Facebook group for users that is quite helpful. Thanks!

Unknown said...

Thanks for posting this. I just did my DNA through AncestryDNA and am looking to learn more about my results than what AncestryDNA shows. I've also just started using GedMatch so I'm looking forward to walking through this process with you :)

Anonymous said...

This article is so helpful, and along with the comments it will help me so much with using GedMatch. I do not always take the time to comment, but I read all your blogs and usually find 2 or 3 "gems" weekly. The contribution you make to the genealogical community is invaluable. Thank-You!

Kendra Schmidt said...

Thank you for clarifying to some degree how to use Gedmatch. I just followed the instructions and am slowly learning how to analyze the data. At this point I don't have any close matches but am working on having siblings and my mother tested to enhance comparisons.

Anonymous said...

I have used my Ancetry DNA on FTDNA and GEDMatch as well, You can get the ancestor distribution pie charts on all 3 (GEDmatch calls this "admixture"), and GEDmatch will also show you how much of each ethnic piece is on each Chromosome graphically. FTDNA and GEDmatch both have chromosome tools showing one or more matches graphically on each strand, and GEDmatch also allows upload of a selective GEDCOM file which it can match against other subitted files for similar names and locations. So many nee tools to go with these have started popping up... I'll be busy for a while!