Tuesday, September 20, 2011

DNA Technology Improving for Sibling Testing

Everyone knows the stories of two families searching their child's orphanage adoption group and finding another family's child that bears an uncanny resemblance to their own child.  At times, such matches seem possible, with the two children sharing common characteristics such as birth dates and finding locations.  Sometimes it borders on the absurd, such as the adoptive mother who thought the child on the Fisher Price Little People Sonya Lee box looked just like her own daughter.

A few years ago I wrote an article on what I felt were significant weaknesses in then-current sibling DNA testing technology, cautioning adoptive families not to put to much faith in their accuracy.  The reason was simple:  Using only 27 genetic markers, the tests were possibly susceptible to "genetic drift", a problem with small, inbred populations, which many Chinese towns and villages are.  Additionally, these tests were often (usually) conducted against databases with few actual native Chinese DNA in them.  Rather, they consisted of diaspora Chinese, Korean, Japanese, and other Asian samples.  Obtaining a 95% "probability" result simply meant that the two children were more closely matched in DNA than 95% of the database.  With a database of millions of DNA samples, 5% of the DNA database's samples would produce a higher "probability". Usually, that margin of error allowed for thousands, or even tens of thousands, of possible matches ("False positives").

I remain of the opinion that many of the most well-known stories of reunited siblings in the Chinese adoption community are more than likely not really siblings at all.

Technology has improved significantly in the intervening six years, and today many DNA labs don't test only  27 markers, or even 1,000 markers, but currently a million markers or more are compared when a DNA test is done.  With so many genetic comparisons being done, previous problems of genetic drift and general uncertainty of a sibling match are eliminated.  With modern testing, the need for parental DNA to perform sibling matches is no longer needed.  When two people's DNA are compared with one million markers, the result is either a positive or negative.  The ambiguity is gone.

This is, of course, of significant importance to the adoption community, where parental DNA is usually lacking.  With current technology, one can now achieve the level of confidence one could only obtain with parental DNA five years ago.  Combined with falling testing costs, and it is now possible for every child from an orphanage to submit DNA and for sibling matches to be made across a wide number of submissions. 

One lab that employs this new technology is 23andMe.com, located in Mountain View, CA.  For $99 (plus $9 per month for 12 months) 23andMe.com will analyze over one million genetic "genomic variations" on a person's DNA.   Of interest to adoptive families, the lab will then cross-analyze the submitted DNA against the DNA from every other person in the company's database, and alert you of any sibling, half-sibling, first cousin or parental matches.  They also allow you to be alerted if a match is made in the future.  Thus, two DNA samples can be independently submitted by interested adoptive families, and 23andMe will provide information (if both parties agree) that allow matched individuals to share information. 

But 23andMe's test goes way beyond DNA matching.  As a result of their huge database of DNA samples and the results of studies done on specific genetic markers, 23andMe can provide you with ancestral information on where your child's ancestors originated -- did her ancestors originate in northern China or Southern?  Did an ancestor migrate into China from another country?  With female children, this information is only available for the maternal lines, but it is nevertheless fascinating reading.  Meikina's DNA indicates that some of her ancestors originated outside China, most likely in Vietnam.  Meigon's ancestors were the same people that migrated over the land-bridge and settled North America. 

Additionally, and perhaps the most important practical information, 23andMe's report will detail possible medical risks that may be found in one's DNA.  For example, my daughter Meigon's test indicates she has a lower than average risk of Parkinson's and type 2 diabetes, but a higher risk of high blood pressure.  Meilan's DNA suggests she may be at significantly higher risk for breast cancer.  These assessments may have important ramifications for Meigon and Meilan's futures. 

Families would be well advised to utilize the current technology in the future for any sibling testing, or to confirm previous tests conducted with the old technology.  Not only will you be given a definitive answer to your sibling suspicions, but you will be given an amazing array of useful information about your child, some of which may have important implications to their lives.
 ________________________

After I posted the above article, I contacted 23andMe.com to have them provide a more detailed explanation of their sibling testing.  They provided me with the following answer:

We can determine a relationship between two siblings by comparing the amount of DNA that they share. We look for segments of DNA between two people that are identical-by-descent (IBD); the more IBD segments two individuals share, and the longer those segments are, the closer their relationship. Siblings, half-siblings, cousins, parents/children all share a certain range of IBD segments. We can assign a relationship based on the segments. 

In plainer English:  When the DNA strands of the birth mother and father are separated to produce the eggs and sperm, it does not occur like a zipper, with one gene going on way, and the next one going the other.  The egg or sperm contains strands (complete sections) of original DNA.  These "gene clusters" are highly unique, and if a gene cluster appears in two individuals, it is strong evidence that the two people are related.  If two individuals possess many such gene strands in common, it is definitive proof that they are siblings.  

With modern genetic testing such as 23andMe's, population drift and other genetic anomalies are no longer a consideration.  This was the main weakness of the 27 allele tests.  But current technology is based ONLY on the two DNA samples, and are not compared to DNA databases in order to confirm a relationship. 

7 comments:

Roberta said...

Do you have any academic research to back up the claims of this commercial lab?

Research-China.Org said...

The website provides detailed academic studies detailing the various components of the report (migration, sickness, etc.).

Brian

Anonymous said...

I Googled the company and was impressed with what I found. Lots of positive buzz out there.

Cheryl said...

Brian - I'm so glad to see you post on this topic. I'm convinced that this new technology holds the key for our kids! I just want to point out something that I learned when doing genealogy with my own extended family of birth.

There are 2 companies in the forefront of this testing. One is 23andme and the other is familytreedna. There are significant differences in the 2 companies that people need to be aware of before deciding which to use (if they can't do both!).

23andme focuses on genetic testing for health risks. Their "Relative Finder" test is almost an afterthought. Because most people are using the service for health information, it is reported that many don't respond when contacted by others that match them on the Relative Finder test. A match isn't any good if the person won't respond. Also, the cost of the 23andme test is $99 plus $9/month for 12 months (total $207). But the kicker is that if you drop the $9/month subscription after a year then they won't notify you of any future relative matches or new health info. This is important because it might be years before a significant number of our kids decide to be tested.

FamilytreeDNA doesn't do health testing, but focuses on finding genetic relatives. So people tested there are usually open to contact from genetic matches. FTDNA's Family Finder test is on sale right now for $199 (no future costs). For this price, you will be notified of any genetic matches now or at anytime in the future...supposedly indefinately. BUT you don't get any health info from their test.

Personally, I would do both to make sure that my results would be compared to the most people possible.

Hope to hear of matches in the future!

Research-China.Org said...

Cheryl provided me with some additional information on how this DNA technology works, and why it is important for adoptive families:

Taken from:
http://www.familytreedna.com/faq/answers.aspx?id=17

"Basically, they look at 700,000+ data points. But even specific points or genes can be shared by non-related people, esp. people from an inbred/isolated population. So this test is actually looking for "chunks" or "strands" of DNA data points. This is because scientists have discovered that DNA doesn't break apart into individual autosomal DNA points, but into strings of points that are passed on to the next generation. It is statistically impossible for unrelated individuals to share large, identical strings of DNA. But with each passing generation the DNA separates differently so that eventually descendants don't share identical segments with ancestors that are more than 6 or so generations back. So this test is THE test for identifying people related within 5 or so generations. This could be especially useful in cases of trafficking. If an adoptee finds just one distant cousin, they might have a clue on where to start searching for birth family."

Cheryl said...

I just came across this link that compares the 2 companies doing this type of testing, FamilyTreeDNA and 23andme. As the author suggests, adoptees should consider joining both databases to increase their odds of locating birth relatives.

http://www.dna-testing-adviser.com/Autosomal-DNA.html

Mary Turnberg said...

I was contacted several years ago by a woman who said that her mother was a match with my adopted daughter from Kunming, Yunnan. I had originally submitted all three of my adopted Chinese children's DNA to the National Geographic's Genome Project and evidently received access to the Familytreedna's data bank since we all get updates of matches all the time. This woman I spoke with several years ago believed very strongly that my daughter, Madeleine was closely related to her mother (who by the way had an interesting story of being transported to Canada and the USA as a slave/servant and that was how she came to this country) at any rate I didn't put much stock into her story at the time because I didn't fully understand how the Familytreedna's science worked. She was very emotional about this DNA connection because her mother had lost contact with her birth family but I didn't pursue further contact since my daughter was young and it didn't seem to me very likely that there was anything more than a "drift" connection that could be said of thousands of other people from China. After reading this blog now I'm thinking that there was a much closer DNA match/relationship than I first thought. I just don't understand the science of this very well.

Mary