Technology never stands still. The latest change affecting all of us who love using DNA for genealogy is a new chip from Illumina. The past six or so years of autosomal DNA testing have shown that the current chip is great for testers with European ancestry, but does not have enough SNP coverage to figure out the details of the ethnic make up for people from other parts of the world. Many more and different SNPs are tested in this new GSA chip.
All the 23andMe tests done since this past July use that chip, as does Living DNA (highly recommended if you have British ancestry since it does local regional breakdowns). I imagine eventually the others will follow along. The bad news is that there is not that much overlap between this chip and the previous ones, which affects cousin matching.
Debbie Kennet wrote a blog post describing the new chip at https://cruwys.blogspot.com/2017/08/23andme-launch-new-v5-chip-and-revise.html
Because the SNPs are so different the DNA results from these kits cannot be uploaded to GEDmatch, however our friends there have built another site to handle these new kits called GENESIS. They have come up with a whole new algorithm for relative matching that works with lower SNP counts.
At the recent i4GG.org conference (videos coming in February), I gave a presentation on what’s new at GEDmatch, the second part of it went into much detail about GENESIS, starting with this slide:
The functions available at GEDmatch are being gradually implemented at GENESIS. Most of the key ones are there now. Plus there is some new functionality. One major addition is the showing of the number of SNPs actually overlapping between kits. Very important to know since the overlaps can be as small as 108,000 SNPs or as large as 580.000.
The thing I like best is the new multi kit analysis selection page which lets you add kits easily, including dropdowns of your own kits, to the ones that you already checked at the one-to-many or kits that match two kits. But I miss the tag groups.
The other new feature that is great is the ability to look just at the FIRs (fully identical regions) on the one to one comparisons. Only full siblings and other close doubly related people share long blocks of FIRs. It also gives the total number of cMs in those FIR blocks, which is very helpful for figuring those ¾ relationships. UPDATE 16-Dec-2017: I have received emails asking about the statement they see under a one to one match like “80.8 Pct SNPs are full identical” – these are individual SNPs, normal when you come from similar population groups. This is not the long shared regions of double relatives.
Missing is the connection to trees from the one-to-many but that will come. The current one-to-many is not the final version I am told.
I recommend you look through the screenshots from my presentation or upload a kit or two and take a look around. In a few months the GEDmatch database will be moved over but why wait?