Saturday, November 26, 2005

The False Hope of Sibling DNA Testing

I have updated this discussion in 2011 with a new posting discussing more current technology.  The reader would do well to read this essay to gain some insight into the issues of sibling testing, and then the essay dealing with current and improved technologies.

___________________

I have had the opportunity of researching the orphanage histories of several girls that were shown to be "sisters" by DNA testing. In all of the cases the girls were found at locations far apart, and were given different birth and finding dates. In other words, if there weren’t a DNA link, no one would have suspected any relationship. All of the relationships were discovered by the adoptive parents finding orphanage sisters through the newsgroups that looked like their daughters, and, after making contact, having a sibling DNA test performed.

But as I traveled to the various finding locations for these "sisters," I began to believe that there was another explanation for these discoveries. It seemed to me unlikely, for example, that a birth mother would leave one girl at the orphanage and another in a small village 8 miles away. As the number of "found" sisters began to multiply, I became highly suspicious. It is of course possible that these matches are true sibling matches, but the analyst in me pushed for a more likely explanation.

So I took a trip to Genetree in Salt Lake City. I sat down with one of their geneticists, and explained my experiences. In summary, this is what I learned about sibling tests:

Genetree (and most other DNA labs) tests 27 genetic markers ("Locations") when doing sibling tests. They will draw a DNA sample from both children, and compare the 27 tested markers from each child against each other, as well as against a DNA database. The database is made up of DNA samples from all over Asia, including Japan, Korea, Vietnam, etc. If the two tested children are not identical twins (sharing exactly the same DNA), the results are compared to the lab's DNA database, and a probability quotient is calculated. In the case of Genetree, my contact admitted that the DNA sampling in their database from China is very small.

Each of the 27 markers has a probability quotient assigned to it. For simplicity's sake, I will limit this discussion to just one location, or marker. Let's assume "Marker 4" has four possible outcomes, 12, 13, 16, and 18. The probability of each result occurring ( as determined from the lab's DNA database) is 5% (value 12), 10% (value 13), 25% (value 16) and 60% (value 18) in a given population. If two children are tested, and both carry "value 12", then there is a very high probability that the two children are related (I will leave the exact probability to a Statistician). These probabilities are compiled from all 27 markers, and a sibling probability percentage is applied.

Without DNA from the birthparents, the tests will either confirm a relationship (Composite probability greater than 90%) or preclude a relationship (probability less than 15%).

I asked the technician if it was possible that the DNA pool from a small city or village in China might differ statistically from the database DNA, and thus allow false positives to occur. He admitted that this was possible.

In other words, it is possible that due to the limited movement of people in China, that genetic “abnormalities” develop in cities and villages. Instead of a 5% probability of “value 12” being in a given population, for example, it might be 20% or 50% or 75%. Thus, two children might be tested and given a high probability of being siblings, when in fact that is not the case – a false positive.

I believe this is happening in the adoption community. I believe that if we took a significant sample of DNA from Jiangmen, Fengcheng, or many other Chinese cities, and tested it against the Asian DNA database of the U.S. testing labs, we would "discover” many sibling pairs. But if we tested these same children against a much larger DNA sample drawn from their own cities or villages, the tests would show no relationship.

At this point I am very reluctant to put stock in the results of these tests because of the limited DNA samples that are from China, and more specifically from the many regions of China. Unless the children are determined to be identical twins, I would be cautious about forming sibling relationships until further research is done to verify those results.

20 comments:

Anonymous said...

I can't agree with your findings, at least in our case. Our daughter and her bio sister were in the same travel group, we did not stumble upon their similarities on a group site. I think a 99% chance of twins is proof enough for us.

Anonymous said...

It seems to be a common trend these days; all these "siblings" found and some are not even from the same orphanage or city!! I think people are placing too much importance on this anyway. Where does this leave adopted siblings? I think there are about half a dozen "half sisters" from my daughters' province alone and I think people are really reaching when they interpret these results.

Anonymous said...

How I wish I could believe your findings and beliefs. My daughter did find her twin sister, and if I had to do it all over again, I would have stopped at the "similarities" said "hmmm, that's interesting", and moved on. It's added a complication to our lives that I wish wasn't there. But the results came back quite conclusive - in the high 90's. Had they come back any less, I would now happily say it's just a coincidence and move on.

Anonymous said...

You wrote: "All of the relationships were discovered by the adoptive parents finding orphanage sisters through the newsgroups that looked like their daughters, and, after making contact, having a sibling DNA test performed."

Perhaps this is the case in all of the found siblings whose histories YOU researched, but in many, if not most of the cases, it was not as simple as noticing similarities on orphanage newsgroups. In our situation we traveled with the other family. We knew the other child in China and watched her growth in photos over many years. Her distinctive look is shared with our daughter, one that we have not seen in any other child from their city, SWI or province. I do think in many cases there is a regional or village "look" that makes it harder to distinguish true biological relationships. In our case we are confident that the DNA, their history, and the girls' distinctive characteristics are enough to consider them biologically related.

Many people believe that families like ours are combing website "searching" for siblings. This is not the case. We did not go looking for this.

Anonymous said...

I do not agree. I think that, in case of twins, family members might travel because they don't want people to notice that a twin was abandoned.This out of fear.If you look around in china you will almost never see people with small baby's.Image a person with two baby's, that will draw too much attention. If you don't want to be discovered, you 'd better separate the twins. In our case, 5 years after they were adopted, the SWI told us they thought our girls ( adopted by 2 families,found 2 days apart) were twins. We did a DNA test and the result was 99%. If the SWI would not have told us, we would not have them tested, althought there is a big resemblance. It did complicate our lifes, but i think it will be of great importance when the girls grow up.

Anonymous said...

I think everyone is missing Brian's point. The numbers the DNA testing companies are providing are unreliable. It dosn't matter if the quoted number is 90% or 99%. If the underlying DNA database does not reflect the true gene frequency in the appropriate population, the numbers are artificially high. Perhaps way, way too high.

Anonymous said...

In response to "ruidh", my daughter was tested with another child from our travel group and the result came back as over 95% probability of siblings. However, I did not take this on faith. I found a large DNA database that included our daughter's region and compared the results against the lab's results. After much research I learned how to calculate our probability and it came out a bit HIGHER than the lab indicated.

The results depend on the database the lab uses. Regardless, we had our own numbers checked against a very large CHINESE database and our results confirm what the lab report says.

If you are a genetics or DNA expert, fine. Otherwise, give us some credit for doing more than taking the word of the lab.

Remember, too, that the results could come out artifically low. MUCH LOWER, with a lab using the wrong database.

Anonymous said...

I whole heartedly agree with you in regards to placing too much faith on DNA test results which are not reflective of the communities from which our children are from. I too would be very reluctant to allow for DNA test to be done and an introduction of a twin to my child, only to set her up for more future letdown when DNA tests become more accurate and she would then find out her sibling is no more than a person from a simular gene pool. Why this desperate need to find siblings for our children? People who are just entering into the homestudy process are now saying their hope is to find a sibling one day or that they will have their child tested to see if she is sisters with certain children. Come on people! We are talking millions of children and yet these miracles are happening more and more all the time! Are people determined for this result and therefore making it happen? for the sake of the child, think hard before you enter into this unreliable source. It could spell disaster in the end.

Anonymous said...

Isn't searching the one and only right of our adopted children themselves?

Anonymous said...

If you would be willing to do some real good research, you all would find out that a lot of twins were adopted at the same time from the same orphanage. When there is a big resemblance, it's not so hard to find out. The DNA test is just the last part, but before that, most parents knew for sure.
yes i always thought that it was the right of my kids to do that searching, but in our case, we were told by the SWI after years. Why waste more years, kids only start to bond when they meet frequently from a young age.
Most of us did not find a sibling by joining one of those sibling finding groups.

Anonymous said...

Anonymous said...

"Isn't searching the one and only right of our adopted children themselves? "

Who said anything about "searching"? The families who have discovered a sibling connection did not SEARCH. It fell into their laps. If you happened to see a child from your travel group or your child's SWI who looked identical to your child, and your child had very unique features, would you just put it aside? What would you tell your child when she's 16? 18? "By the way, when you were six, I saw a child that I strongly believed might be a biological relative of yours but I did nothing about it" ?? Until you've walked in our shoes you have no right to judge. We do what we feel is best for our children.

People seem to think this is happening on a daily basis. Far from it. There is an incredibly small number of children who have been tested (less than one-quarter of 1% of China adoptions).

Anonymous said...

I am sure that my daughter has a sibling. He is the brother that waited for her and traveled to China with us to bring her home. He is the brother that laughs with her, fights with her, and plays with her. We chose to adopt because we believe that love, not biology, makes people kin. Even another Chinese adoptee with a 99% match to our daughter would essentially be a stranger to her. Yes, they would share some common history, like she does with other China adoptees from her province and SWI, but not history based on a relationship. (The importance of this common history and heritage is why my children take Chinese lessons and know other families through FCC). To me, a DNA test - no matter how accurate - cannot indicate that kids are "real siblings." After all, aren't we "real parents" and "real families"? As a mother with both biological and adopted children, it is obvious to me that the love and committment we have for each other is what makes related. Genetic similarities are simply interesting, but should not be so elevated.

Anonymous said...

I travelled with a couple who discovered their daughters twin quite by accident.They where not in our travel group. They are now learning to work out a new relationship with the other family to keep the girls close.

There where only two adopting families in our group and I could just as easily been the mother of the twin. I have to say that I see a tough future ahead...and alot of work. To raise these girls as twins in different states. These parents are in unchartered seas and we need to support their decisions as a community, not fuss over what the believe or how they choose to handle their lives.

My own daughter has a sibling as the previous poster said, one that was chosen to also be a part of our family. Both of my girls are from China. Biology is good, but it is not what makes a family.

On another note as my older daughter grieves over the loss of her birth Mom and worries over the other children she may have kept while giving her away, I don't feel the need to look for confirmation. At this point in time it would truly wound her to know.

I will support her in the future if she wants to go back to China and try to find out more.

Someday in 10 years or less I believe that a large group of our Chinese Daughters and Sons will take China by storm and demand the answers THEY need or want.

After all China is a very large place made up of very, very small ones. Secrets are hard to keep forever...

Anonymous said...

I think the point is being missed by many of the comments.

Theoretically siblings could have 0% DNA markers in common if each got a different "1/2" of the birth parent's DNA. The odds of this happening is not high but it is possible. When villages stay together for generations in the same location (eg little migration) then the genetic background of the village becomes more similar as many villagers are 5th cousins tiwce removed (just an example) or whatever since there would be more intermarriage than there would be, on average, for a bigger city.

This doesn't mean that true siblings/twins aren't found however it does mean that with such a limited number of markers being looked at it is highly likely that there are also false positives and false negatives - *especially* if children come from villages where historically there has been little marriage outside the village.

It isn't that DNA comparisons are *junk sicence* persay, rather when you don't have enough markers for a good comparison and don't know an accurate baserate for the underlying markers used (meaning the naturally occuring frequency of the markers in a population) it means that the results could change as the knowledge of the baserate becomes more accurate. The results could change as the number of markers used when making these comparisons increases. This is a basic statitical principle that underlies matching.

As a result it is possible to get false negatives and false positives. Kids may look very similar - after all if they share some genetic markers there will be similarities. Take a look at the celeberty look alike contests where some people are dead ringers for the celeb and they are NOT related to them.

The problem occurs when assumptions are made using a science that is still developing, with a database that is inadaquate and possibly with the wrong choice of markers for a particular population or village, and the number of markers is small.

Anonymous said...

I too found out by accident. Wasn't looking. Birthdates were the same, same orphanage but one left in the town the orphanage is (near the river dock) and the other 1 week later in front of the orphange. 3 and 4 weeks old. Early pix were very very similar. We did sa DNA test and all they could confirm was they were related. So we sya they are fraternal twins. We live in different countries, have met and will keep the relationship. We will see how they react, but whether or not they really are twins or jsut related, they will always share something. They were also adopted 1 day apart. It will be up to them to determine what kind of relationship they will have. But it's nice to know you have a bio relative somewhere when you were abandoned at birth. As parents, we keep in touch. We have photo albums of each other's families. It's nice. And if the DNA gave us false positives, so what? If DNA tests become more effective in the future, I'll leave it up tot he girls to decide.

Anonymous said...

Elisabeth, just because two children were given the same birthdate by the orphanage, it does not make them related. Aren't you lying to your child if you don't KNOW if your child is really related to the other child? You say "if it's a false positive 'so what'? Well, you are telling your child that she is related to this other child, even calling them twins, but you make it seem like it's no big deal if it's a false positive. Where's the trust your child will have in you as her parent? She'll grow up believing she has a biological twin in another country because you had a test done saying "they might be related". And you say "so what?" if it's not accurate? You just assume they are fraternal twins without any proof? I'm horrified.

Anonymous said...

I am a Asian woman, born in China. My daughter is an orphan born in China, adopted from China early in 2006. I have an Asian co-worker where I work and she has a bilogical daughter born in the US. Her daughter and mine are two weeks different in age. The two girls look almost identical to one another (even to myself and my co-worker who can tell small differences in Asian features easily). From an appearance perspective they could easily be identical twin sisters (and certainly could be fraternal by looks).

BUT THEY ARE NOT. They are NOT releated by blood. They only share a common genome by being Asian.

Let me repeat that, THEY ARE NOT RELATED BY BLOOD, even though they look like they could be twin sisters.

Common look, common declared birth dates, etc. mean nothing Michelle and Elisabeth. And with the problematic aspects of DNA type matching that currently exists with the Asian genome, DNA matching answers nothing.

Parents need to stop fantisizing and then looking for convenient ways to validate such fantasies. It's not healthy for you or your child to be doing this kind of thing. Why do you do this!? What does it validate for you personally?

Anonymous said...

Just an observation: the angriest email comments here come not from parents who believe they have found a biological sibling and who are upset by the Research-China.org post. They are by adoptive parents who are dismissive of other adoptive parents having perhaps found a bio sibling. Haven't we all done the reading that shows that it is highly likely that our children will yearn, at some point in their lives, to see a face that looks like theirs and that is biologically related to them? Not ALL adopted children will but many will. I have two girls, one bio and one adopted, and they are most definitely sisters. But if I had any reason to suspect that my adopted child had a bio sibling, I would definitely consider it a loving parental duty to follow up on it. If I found a bio sibling, this would not in any way detract from the sibling relationship my daughter has with my other daughter. I guess I'm just wondering why some of the adoptive parents who dismiss this possibility here are so angry about it. At the same time, I am grateful for the research done on Research-China.Org and the reasonable tone assumed on the website itself.

Anonymous said...

Jennifer,
Thank you for your genuine and objective input. It's true that there is so much anger in the comments of those who agree with the author of the article. How many times and ways has it been published that most adopted children grow up and want to know all they can about their family? How can any of these parents NOT want their children to have someone close to them who can identify with and relate to their child's unique circumstance? How can they judge others for wanting that? We did find our daughter's sister through our orphanage Yahoo group. We are searching now for the right DNA test, although there is no doubt. Our girls are from the same orphanage, look like identicle twins even tho they're 1 yr apart, and have a rare congentital condition(1/21,000) that only occurs if both parents have the gene. We've met twice so far and shared many emails and photos. When this family put her in my arms I found she even has the same voice as my daughter. The only complication is that we live 16 hours away and still we've managed to meet up twice since we met them less then a year ago. We all agree that because our girls have so much in common, we consider them sisters anyway so how I ask can this hurt them? How many of us had/have a special friend that we loved and called our sister or brother because we were so close? As long as both families agree and there isn't some huge difference in moralistic standards, it really shouldn't be very complicated. Instead it should be celebrated and all this can be done while still being honest with your child when they are old enough to understand. Even if the families never end up getting together, at least you found someone who could possibly be related and when your daughter/son is an adult, and want's to know what you did to help them find their family, you can look them in the eye and say you tried because you knew it would be important to them someday.
I wish the best to all of our adopted children to find as many answers that they want and need regardless of how validated they already feel because their parents raised them to understand that their true value comes from God and from within. Brenda

Research-China.Org said...

The DNA technology has improved considerably in the last few years. One lab that I have a lot of confidence in is 23andMe.com. Rather than testing 27 alleles like the tests did of a few years ago, it now compares 500,000 markers. With that many genetic comparisons, genetic drift is no longer an issue. 23andMe.com is the lab I would use for a sibling test.

Good luck!

Brian