How does a paternity test work?
-A curious adult from California
June 1, 2004
Did I find my real dad? Is that really my son? Did Michael Jackson really father Billie Jean's kid? Questions like these used to be very hard to answer. In the past, people used a blood test. This might rule out that you were the father but couldn't prove that you were.
Nowadays, DNA technology is used to figure out who is the father of a child. DNA paternity testing makes it possible to determine a child's biological father to a very high degree of certainty.
Everyone, except identical twins, has a unique set of DNA. DNA is made up of 4 bases or letters, A, C, G, and T. These 4 letters form the written code that makes up the DNA sequence.
Now when someone says that everyone's DNA is unique, what they mean is that occasionally one of these letters is different for different people. On average, two people at random have a different base every thousand bases or so. This is where the statistic that says that everyone's DNA is 99.9% the same comes from.
Since you get half your DNA from your father and half from your mother, your DNA is more than 99.9% the same as your parents. Your DNA is also more similar to that of your grandparents or cousins than to that of a random stranger. Paternity tests use this greater similarity to figure out who the parents are.
So how do you figure out someone's DNA is more similar to another's? There are lots of ways but we'll focus on the simplest, DNA restriction analysis or DNA fingerprinting. DNA fingerprinting uses special proteins called restriction enzymes. Restriction enzymes cut DNA but only at a certain combination of A, G, T, and C. Different restriction enzymes cut DNA at different places -- each has a unique sequence it recognizes. For example, the restriction enzyme EcoRI cuts DNA at the sequence GAATTC and will cut only at that sequence. It will not, for example, cut at GACTTC.
OK, so what DNA fingerprinting does is it looks for differences in the DNA that change where these restriction enzymes can cut DNA. The pattern of DNA fragments is then compared and if the child's DNA looks like a combination of the two parents' DNA, then the child is theirs.
Let's look at an example of how this might be done. Suppose we have three people: Bob, Larry, and Mary. If we take the same stretch of DNA from the three of them, small differences might mean that EcoRI will cut them differently (see Figure 1). In Bob, the sequence GAATTC occurs once in this stretch of DNA. That is, in this stretch of DNA, Bob has one EcoR I site. Now suppose Mary has no EcoR I sites and Larry has two EcoR I sites in this stretch of DNA. You can see that EcoR I will cut this stretch of Bob's DNA into two fragments, Larry's into three fragments, and Mary's will not cut.
When we cut the DNA with EcoR I and separate the cut fragments on an agarose gel, the gel might look something like in Figure 2. In an agarose gel, smaller fragments run faster so you get separation based on size -- the bigger fragments are near the top, the smaller are near the bottom.
Now, suppose Mary has a child and she wants to determine which of two men, Bob or Larry, is the biological father of her child. She consults a paternity testing expert. The expert collects a certain stretch of DNA from Mary, Bob, Larry, and the child, and cuts the DNA with EcoR I. When the expert separates the cut DNA fragments on an agarose gel, the pattern looks like the one in Figure 3. The child's DNA must be a combination of Mary's DNA plus one of the men's DNA. The agarose gel indicates that the child's DNA is a combination of Mary's DNA (top band) plus Larry's DNA (bottom three bands). Thus, Larry is the biological father of the child.
What happens if Larry and Bob have the identical sequence in this stretch of DNA? The answer is that you wouldn't be able to distinguish, based on looking at differences in this stretch of DNA, between the two men. So what do you do? You simply search for other stretches of DNA in which there is a difference between these two men. This is why in real life, multiple stretches of DNA must be examined to ensure that the results are statistically significant.