Combine Kits into One Superkit on GEDmatch Genesis - Sat, 6 Apr 2019

Today GEDmatch Genesis added a new Tier 1 application. They state:

I did that myself manually with 5 kits about 6 months ago, uploaded my combined raw data to GEDmatch Genesis, and reported the results in my post: The Benefits of Combining Your DNA Raw Data.

I thought I’d try the new GEDmatch Genesis application to see if it produces essentially the same result.

I selected the Tier 1 “Combine mupltiple kits into 1 superkit” application and it gave the the option to select up to 4 kits that are already uploaded. I had all my 5 kits uploaded and I selected FTDNA, 23andMe, Ancestry and LivingDNA. I left out MyHeritage which uses includes almost the same SNPs as my FTDNA file does.

I pressed the “Generate” button and within a second, I got my combined kit:

Comparing my kits using the GEDmatch Diagnostic utility gives:

When I manually combined the kits, I got 1,389,750 SNPs, but GEDmatch only combines the 1,123,247 SNPs it wants to combine that it knows it is going to use. Slimmed SNPs are what GEDmatch actually uses for comparisons with other kits. I’m surprised that GEDmatch’s 834,457 slimmed SNPs are over 20,000 more than my manually combined kits. I have no explanation for that.

I’ve included my Whole Genome kit from Dante, that GEDmatch only loads the SNPs in the VCF file. Those SNPs are the ones where I am different from the human reference genome. The SNPs where I am the same as the human reference genome are not included. The GEDmatch people still have to fix the upload of VCF files so that human reference genomes are added when the SNP is not included in the file.

The one to one comparison was possible immediately, so I compared the GEDmatch combined kit to each of my individual kits, and to my manually created All-5 kit.

All of the comparisons indicate that I match myself at least 99.210%. It’s not important that there are some small breaks in the matching segments which results in more than 22 shared segments. I expect that when the one-to-many comparisons become available, the overlaps will improve just as they did with my manually combined file.

The Bottom Line

If you’ve tested with multiple companies and you subscribe to Tier 1, you should combine your kits to get better comparisons at GEDmatch Genesis. Make sure you make this combined kit the kit for yourself that you use for matching, and change all the others to Research so that you show up only once in other people’s match list.

The only unfortunate thing is that you don’t have access to your raw data at GEDmatch. So you won’t know exactly what they did and you won’t have the raw data for yourself to look at or use for other purposes.