Tuesday, July 8, 2014

DNA News of the Week: Solving Problems With Segments

I had not been able to finish watching the last half hour of the Jamboree Webinar I purchased until last week. For some reason (maybe the fact I use Wifi?) I kept losing the video signal half an hour into it, and could never advance passed that. I purchased the session "TECHNIQUES FOR USE OF AUTOSOMAL DNA TESTS TO BREAK THROUGH GENEALOGICAL BRICK WALLS" (link to syllabus1, syllabus 2).

The techniques used by Dr. Janzen focus on establishing a good foundation for drawing conclusions using shared segments. Having your parents tested allows you to do phasing which is extremely useful. Even if you only have one parent test it's very useful. The more close cousins you test or compare with, such as 1st through 3rd, puts you in a better position when it comes to making connections with more distant cousins. The segments shared at these cousin levels tend to be large, and more conclusive than the much smaller segments shared by distant cousins. You can extend the length of your ancestral segments if your matches shared segment extends beyond yours. When you find others sharing smaller segments at 4th cousin and beyond you can then more confidently label those segments. Unfortunately, 23andme and Family Tree DNA are on different builds, with 23andme on the 37 build and Family Tree DNA is on the 36 build. This just means more positions have been added to the chromosomes by 23andme, making them longer. This can lead to problems when comparing between companies. Family Tree DNA may eventually move to the 37 build or advance to a 38 build.

Dr. Janzen emphasized the need to take great care before assuming you received a segment from a particular ancestor. You need to look for all possible connections going back as far as possible. His example was the surname Broshears (which coincidentally has a connection with my family). This name isn't very common so he assumed that was his connection with one of his matches. Doing more research he discovered it was not their shared ancestral line. Instead it was Alexander.

Our Mullens lived 15 miles from one
of my matches Mullens
As an aside, interestingly this week I received an email from a representative of an adoptee. He concluded the same thing I did. This person is related to me through my Irish line due to the fact several others with the same roots match me on that segment. He also collected gedcoms from the other matches and found out that we all shared ancestors with the surname Mullen. Three of us do. Mullen being very common, the population of Ireland not being very mobile, and a paper trail going back only around 150 years makes, confirming our connection is through Mullen nearly impossible. We are hoping advances in segment identification will solve this problem. He also concluded we were likely 3rd cousins based on the amount of DNA we share in common. This can be misleading due to the fact the Irish population is fairly endogamous.

Use with caution (from ISSOGG site)
Dr Janzen also spoke briefly about endogamous populations. Isolated populations, such as those living on islands, tend to marry blood relatives. This leads to even distant relatives sharing large amounts of DNA. Autosomal DNA testing as it currently stands is not very helpful for those with ancestors in these populations.   The French Canadian population is like that,  since everyone is descended from a small population of founders. The founding population of the US was much larger and tended to be always on the move (and receiving new infusions of DNA from immigrants). My ancestors, anyway, were constantly on a westward move until we hit the Pacific Ocean. So far, along my US line, I've only found one match sharing with me through more than one line.

Someone asked a question at the end directed toward the medical side of testing since Dr. Janzen is a medical doctor. They wanted to know how useful 23andme type testing is. He said it can provide useful information, but isn't really that informative at this time. As he said a majority of illness is due to factors not related to DNA. Lifestyle is the number one contributor to disease and illness.

Dr. Janzen's session points up the need for easy access to segment information. Without it we can't confidently draw conclusions. He stated we need to put pressure on the testing companies to release this information.
(Another aside, I just finished listening to Dr. Janzen speak again about the use of segments. He said he feels testing 2nd cousins offers more bang for your buck).


Kalani said...

Since last year, it was after July I noticed that when I want to download the raw data from FTDNA, it gives me two choices, 37 build or 36 build. Doesn't that mean then that they have 37 build?

Annette said...

Thanks! Oops. I don't know? Maybe Dr. Janzen's information is out of date? Or I misunderstood him? I'll see if I can access his video again.