Recently I gave an updated talk about GEDmatch.com for my local DNA special interest group, DIG, here in San Diego. GEDmatch.com is a DNA geek’s playground, but many less computer inclined folk find it difficult at first.
It is the only place for those who have tested at Ancestry DNA to compare their results to a possible relative, chromosome by chromosome. It also has many tools that are unique such as ancestry composition calculators with more recent breakdowns and more categories than the main companies. I covered those in detail in my original talk about GEDmatch tools. Those slides are at http://slides.com/kittycooper/gedmatch#/5
The new talk – http://slides.com/kittycooper/gedmatch-10#/ – covered uploading your data, how to manage your kits and mark a kit for research, and much detail on the one-to-many function as well as all my other favorite tools (starred in the image to the left).
There is a new 23andme upload which is nice and fast as it uses the API so you actually log into your account there rather than uploading a file.
It makes sense to upload all your kits when you have tested at more than one company but please mark all but one kit as research only, so DNA relatives are not confused by seeing so many versions of the same person.
To mark your kit as research only you click the link EDIT or DELETE your DNA resource profiles at the bottom of your list of kits. Then select a kit to edit. On the next screen you check a box to mark it for research only. Keep the kit with the most SNPs as the public one. You can use the DNA File Diagnostic Utility to check which that is. Probably that is your Ancestry kit unless you tested at 23andme before December 2013
Why have more than one kit? Well they each test slightly different SNPs so it makes sense to compare apples to apples not oranges. In other words, if you have a new DNA match that has tested at a specific company you want to compare them in the one-to-one with your kit from that same company for slightly more accurate results.
The one-to-many uses lower thresholds than the defaults for the one-to-one comparison, so I usually lower the SNPs to 500 and the cMs to 6 when I do a one-to-one, but take the smaller matches that show up with a grain of salt …
The one-to-many is the most important tool of all. Use it to regularly check for new close matches. I went over what each column is and discussed the generations listing. Note that you can sort any column by clicking the blue arrow at the top of that column. Normally it is sorted by total cMs. Here are the columns explained.
|Kit Nbr||The kit number: Ancestry kits start with A, 23andMe with M, Family Tree DNA with T, MyHeritage with H, and others with a Z|
|Type||The companies change chips occasionally so this tells you which version, V4 is the current version for 23andme|
|List||Not what I expected, clicking this gives you the one to many for that kit|
|Select||You can select a group of kits with this box, then click submit at the top and have many comparison options|
|Sex||If the user did not put it in you see a U, else F or M|
|Haplogroup||there is a column for the mtDNA haplogroup and the Y haplogroup, self reported.|
|Details||Click the A to get the one-to-one autosomal report|
|Total cM||For close family, total cM will follow Blaine’s charts at ISOGG well, up to 3rd cousins|
|largest cM||For endogamous groups I recommend sorting by this column (the blue arrow at top)|
|Gen||Subtract one to get the cousin level. So 4 means a 3rd cousin. A nephew or uncle will be 1.5|
|Details||Click the X to get the one-to-one X report.|
|Total cM||Total matching cMs on the X.|
|largest cM||Largest segment on the X, as X can reach very far back in time so only look at the large matches initially|
|Name||the name or pseudonym of the person|
|email address to use for contact|
In the generation listing your parent or child is 1, an aunt or niece is about 1.5, a grandparent is 2, for cousins subtract one from the number so if it shows generations of 4, the suggestion is 3rd cousin. My recommendation is to follow up with anyone who is 4.5 or less. Start with the closer ones. Accuracy in autosomal matching declines after 3rd cousins. I showed Blaine Bettinger’s chart for estimating a relationship from total shared cM (this has been updated recently see http://thegeneticgenealogist.com/2016/06/26/update-to-the-shared-cm-project ).
You can do a one-to-one from the one-to-many by clicking the A. I included a number of examples including my Dad at two different companies (note the small breaks!). Also two full siblings will have many green bars, places where they are fully identical because they inherited the same DNA as each other on both paired chromosomes from each parent.
A very useful feature is to select a group of people from the one-to-many and then look at all of them together in various chromosome browsers. Also you can get a CSV of their matching segments that compares everyone in the group to everyone else. Since it is sorted by person this is a quick way to add match information to your master spreadsheet or to start a spreadsheet if you tested at ancestry.
I went over the 2D chromosome browser, also covered in my blog post here which compares everyone to just the first person in the list (you?). Next we looked at the 3D browser which compares everyone in the list to each other. You have to select each chromosome one by one.
After the one-to-many and the one-to-one, I find the ”People who match one or both of 2 kits” most useful. It is a quick way to see which family line a new relative is part of by who else they match.
Finally at slide 36 I got to the paid, tier 1, utilities.
I use the matching segment search when I am interested in a specific segment. You used to be able to specify the segment but so few people used it that way that they turned into a listing of all your overlapping segments.
The relationship tree only works if you have X and it was an experiment.
Of course, my favorite in tier 1 is the triangulation tool(see my post on that). I suggest selecting both options the first time you use it. Then run it again with just the sort by kit number because you can cut and paste that to a spreadsheet and then sort by chromosome and start point to see the triangulation segments
I mentioned the lazarus tool where you can construct a kit or partial kit for a dead ancestor if you have enough descendants and relatives tested but did not have time to go into its details.