One of the most interesting sessions I attended at RootsTech was a panel on genetic genealogy chaired by CeCe Moore with Tim Jansen (advisory panel member, 23andMe), Catherine Ball (Ancestry DNA), and Bennett Greenspan (Family Tree DNA). Here's my summary.
Q. How do you persuade someone not interested in family history and DNA to contribute a sample?
A. Tell them you're interested in learning about your family history and you think you're related. Ask if they are interested in learning about your common family history and if you can collaborate?
Some people won't want to do it. Accept it, after you've explained the benefits to them in a way that they can understand. Many people don't want all the details, some just want a 15 minutes summary - then a brownie. In one case a cousin wasn't being cooperative, but his wife was able to persuade him!
Q. What do you tell people when they don't get good (high resolution) matches?
A. Be patient. As the databases get bigger it enriches the experience for everyone. Encourage others to test. Holes are being plugged all the time.
Q. What should I do if I get contacted by someone who is adopted and matches me on an autosomal DNA test but we don't know where the match might be?
A. He should join the Adoption DNA Yahoo Group. Make sure you've DNA tests from older generations so that you can limited the match to the mother's or father's side. Take it further by testing first and second cousins which will also help limit where the match occurs.
Q. If I have a DNA sample from a deceased person should I get an upgraded test now or wait until the technology advances?
A. Costs are coming down; technology is improving. Likely in 10 years everyone will be getting a complete DNA analysis so if you have limited sample wait. On the other hand despite the best preservation efforts by companies swab and spittle samples deteriorate. Although they'll always be good for Y- and mt- DNA tests they may not be satisfactory for the next generation of tests.
Q. Why is genetic genealogy still back in the old days of genealogy databases when each had its own incompatible format? If you and a cousin had your data on different databases they couldn't be brought together without reentering the data manually. Then came gedcom. When will the same thing happen for DNA data?
A. From each of the major companies you can download all of your own data. For Y-DNA you can place your data on YSearch and for mt-DNA on mitosearch both of which accept data from each of the major companies. For autosomal tests you can download your data from 23andMe and upload it to FamilyTreeDNA, for a price, and will soon be able to do the same for AncestryDNA results. Then you can use FTDNA's matching capabilities. Gedmatch, a private initiative, will accept the autosomal data for free. To ensure privacy you have to take the positive step of moving the data, just as you do with a gedcom. The database matching, maintenance and analysis provided is a value added commercial service.
Q. How can data from a whole genome medical DNA test be used for family history. How can it be used to help other medical studies?
A. From the medical test results it is possible to extract the SNPs used in genealogy. It requires a bit of programming; contact the people who did the medical testing. GenBank will take mt and complete DNA sequences for medical research purposes. You can also search the National Institutes of Health website and search for research on the disease of interest.
Q. How can I know whether there's enough sample available for future tests for someone who is elderly?
A. Family Tree DNA will respond to a request from clients and let you know how much sample remains. They will also send a new test kit so you can collect an additional sample if there is insufficient, without cost for existing clients.
Q. What is the limit of how many degrees of cousin-hood we can establish and are going to be able to to establish through an autosomal DNA test as technology improves?
A. At present the companies test pretty much the same SNPs, about 700,000 with 23andMe adding an additional 200,000 which are considered medically significant.
AncestryDNA is very confident (95%) in its prediction of a 4th or 5th cousin relationship. As you go further out relationship through multiple lines can muddy the waters. AncestryDNA has examples of correct predictions of 7th and 8th cousins.
There are examples of sections of autosomal DNA matching from ancestors with a common ancestor 500 years ago. But DNA inheritance is random so there may well be closer cousins that you don't match. Given the criteria established for matching by the testing companies 99% of the time you can identify a 2nd cousin, 90% of the time for 3rd cousins, 50% for 4th cousins, 10% for 5th cousins, 2% for 6th cousins.
Outside this session CeCe Moore had discussions with AncestryDNA which she reports at http://www.yourgeneticgenealogist.com/2013/03/ancestrydna-raw-data-and-rootstech.html
Sadly there's still no immediate prospect of AncestryDNA offering their service outside the US. I was told perhaps by the end of the year in Canada. Unlike Family Tree DNA which sold a large number of tests at WDYTYA Live last month because they had the test kits on hand, so the sample could be taken right there, AncestryDNA had no kits available at RootsTech!