98.75% positive = negative? DNA paternity testing

  1. I'm trying to prove that my daughter is my daughter. We just did DNA testing, but either me or my daughter has a mutation in our DNA and the lab is having a hard time coming up with an answer. They said that they are 98.75% sure that I'mno the father, but they won't say I'm the father unless they are 99% sure. So they will tell the courts I am excluded from being the father unless they can come up with another 0.25%. They are going to test the mother now and the doctors think the number will go up when they do, hopefully.
    What I'm looking for is some information on exactly what tests are done and how they are done. I don't want some kindergarten fluff answers, I'd really like to know the actual processes. Thanks for anything you come up with.
    P.S. If you have a child, make sure your name is on the birth certificate, you have no rights to see your child unless you are listed on the birth certificate and it is hell trying to get your rights restored.
  2. jcsd
  3. I dunno whether this qualifies for kindergarten fluff or not....

    The way it is done is that they extract a sample of your DNA (and the daughter's), and then replicate it many times using a process called PCR. Then, with these more reasonable amounts of DNA, they use a type of electrophoresis to seperate out the components of the DNA to for a genetic "fingerprint".

    The way this works is to use enzymes call restriction endonucleases to break up the DNA into certain segments, each of which leaves a specific trace with PAGE (polyacrylamide gel electrophoresis), in terms of how well they form ions, and their relative sizes. Basically, the cut segments of DNA are put on the centreline of a gel (which as a seive) and a current applied and then switched off. In that interval, the ions would migrate towards the positive or negative terminals with different speeds(due to different sizes), and hence get separated. A staining agent is used to make the pattern appear.

    By cross referencing particular groups of patterns on the dna fingerprint that are know to vary across families, they can identify a genetic relation. The mother of the child is useful in sorting out which signature came from the mother and which from the father, hence sorting out anomalies.
  4. Okay, that was good. How about the percentages. 98.75% sure seems pretty sure to me. Does it sound reasonable that by testing the mother I can gain that 0.25%? How do they come up with a percentage? Any idea what they meant by mutation? I would ask the lab, but the doctor went on vacation and the lady I spoke to didn't know.
  5. russ_watters

    Staff: Mentor

    From what I have heard, DNA testing is generally far more accurate than that. If they are used to 99.9999% (one in a million) then 98.75% (roughly one in 100) would seem low.
  6. lol, one in a hundred? call me crazy but isn't 98.75% by definition 98.75 in a hundred?
  7. Yes but I think the point that was being made was that for every sample of 100 men from the population, statistically one would be a match for your child. In short, not particulary accurate.
  8. Monique

    Monique 4,445
    Staff Emeritus
    Science Advisor
    Gold Member

    I'm sorry FZ, I really don't think it is done that way.

    Here is what happens: DNA is isolated from child and alleged father, preferentially also from the mother. The genetic fingerprint is generated by amplifying 16 polymorphic loci on the genome (16 regions that are known to vary in length from one person to another with high heterozygosity). Now: each fragment will have a different size range and is labeled with a different color. These fragments are put on a cappilary and are seperated on size and detected by color.

    The idea is: if the child has a state x at a loci, and the mother doesn't have that state x, the allele HAS to have come from the father. If the alleged father has state x, he MIGHT be the real father, but you are not sure.. you need the information of all the other loci. If the father DOESN'T have state x, you would think you are sure that the father CAN'T be the father. True, but these length repeats in the genome are unstable, the length can change from one generation to the other - mutate. So you still need the information of the other markers. That is why you can not reach 100% certainty, but the more loci, markers are typed, the more certain one can be.

    The problem is: the technique itself is not fail-proof eather, it can happen that some loci, markers didn't amplify. An option would be to type 7 more markers and use a different technique. The first one I described is based on fluoresent detection, the second one could use silverstaining to visualize the results. If THIS didn't yield enough certainty either, they can do a 3rd test, restriction fragment length polymorphism with up to 8 markers.. I guess this is the one FZ described.. the problem with this one is that a different type of marker is used: single nucleotide polymorphism, which only has two states (compared to the other marker which has many states), less states is less information.

    Important in my opinion, I don't know if it is routinely done, is to also get the same information from the mother so that it is easy to track which allele came from where in the child.

    I hope this helped more than it confused you, Tribdog, at least you know that they can do these three different tests to be absolutely certain of the answer. Good luck.
  9. Monique

    Monique 4,445
    Staff Emeritus
    Science Advisor
    Gold Member

    Thank you for pointing that out damgo, this is what I found, the two approaches compared:


    So basically the PCR method is much faster (3-9 vs 1-14 days) and easier to perform, but the RFLP method is more accurate (4 RFLP locations = 6-9 DNA locations). That is surprising, since I would consider RFLP old technology..
  10. LURCH

    LURCH 2,507
    Science Advisor

    The mother and daughter must not live nearby (in Arizona), right? For legal matters such as child-support and custody/visitation, I heard that AZ will accept anything over 95% probability as proof. Here in Michigan (and most states, I believe) require 99%.

  11. 99% here too. 95% would have been great.

    I wish distance was all that separated mom and daughter. It's a long sad story, but you asked for it and I want to get it off my chest anyway.
    Mom was someone I never should have slept with. She got pregnant and one day left my daughter with me and didn't return. That was fine with me. Everything was great for almost two years. I wasn't named on my daughter's birth certificate, but it didn't matter because she was with me. I should have taken care of it, but I put it off and now I regret it. One day we learned that the mom was in prison, but that didn't affect us because we hadn't seen her in so long anyway. The next day the mom's parents showed up and grabbed my daughter, threw her in a car and took off. The police wouldn't do anything because I couldn't prove that I was the father, but they could prove that they were the grandparents. I haven't seen my daughter since October 6th, 2002 and the courts are so slow I'm not scheduled to see a judge until October 6th, 2003. There is nothing I can do but sit and wait and pray that the labs come up with another 0.25% so that I can have my daughter back. The grandparents have already raised 6 kids, only one graduated high school and 3 are in prison. I don't want them raising my daughter.
    On a side note, I'm a cynical, apathetic sort of guy. Logical, not emotional, most of the time, but not when it comes to my daughter. I love her more than I ever thought I was capable. I have not been happy for 1 second since she's been gone and I won't be again until I get her back.
  12. russ_watters

    Staff: Mentor

    I was rounding to make it even for the basis of comparison.
  13. Yes.

    But I have a feeling that the % is a measure of the similarity of your DNA to your daughters. I don't think it is a measure of the likelyhood that you are the father, that no doubt comes from determining how likely it is that your daughters DNA could match yours simply by chance. I'm not sure how they'd measure that, as I'm not a biochemist, but I have a feeling that it was this measurement Russ-Waters had in mind when he was talking about 1 in a million and so on.
  14. I bow to the wise one... :smile:
  15. There is no DNA Lab in the World that can guarantee 99% accuracy.

    It's miraculous that they can approach two standard deviations (97.5%).

    It's not the precision of the result itself, but all of the prospective errors that are built into the analysis.

    I know it's no help, but any State that would require >97.5% results on a biological model, is just plain stupid.
  16. LURCH

    LURCH 2,507
    Science Advisor

    Did you check that through a lawyer? I did some snooping around, and the info I found no this chart says AZ only rquires 95%.
  17. The lab doing the testing has said that 99% is their cutoff point. They will return an "excluded" rating for anything lower and will not send the exact % to the courts. I already tried talking to them about it.
Know someone interested in this topic? Share this thead via email, Google+, Twitter, or Facebook

Have something to add?