How The Test Works

Comparing Father With Child

How Does Paternity Testing Work?


  • DNA samples are collected using simple mouth swabs (known as buccal swabs) and submitted to a testing lab that conducts a test on many different STR locations found in your DNA.
  • Children inherit their DNA from their parents and a child’s DNA STR types originate from a combination its mother’s and father’s STR types.
  • By comparing STR results from the child and known mother we can tell which of the child’s STR types are absent in the mother and so must be from the father, that is, we identify the child’s paternally inherited STR types .
  • If any of the child’s paternal STR results are absent in the questioned father’s results, he cannot be the true biological father.
  • If the questioned father matches all of the child’s paternal results he could be the true father and a statistical calculation is carried out.
  • Modern STR methods analyse many different STR locations in your DNA simultaneously, often as many as 15 STRs all at once.
  • When the questioned father matches all 15 of the child’s paternal STR types you can expect the report to quote a probability of paternity of around 99.99%.
  • If he doesn’t match all 15 then he is not the true biological father.
  • It is possible to conduct the test without a sample from the known mother but this is not recommended because stronger statistics and more reliable results are obtained when mother is tested.

Read on for a more complete explanation…

In the home page article Amazing DNA I describe how a person’s complete genetic code is packed into 23 chromosomes in the nucleus of every living cell. I also said that every cell, other than the sperm and egg, has a double copy of these chromosomes and so each has 23 chromosome pairs. The reason that there is a double copy is that one chromosome is inherited from the father and the other from the mother.

Sperm and egg cells are the only cells that have a single copy of the chromosomes, the full, paired set is only formed when the sperm cell fuses with the egg during fertilization at the point when a child is conceived and pregnancy begins. A fetus develops as the fertilized egg cell divides and duplicates many, many times. On each cell division a full copy of the 23 chromosome pairs is made and passed on to the new cells. As more cells are copied they start to differentiate into the different tissues and structures needed to create a new child, but every cell regardless of what type of tissue it belongs to, retains a complete and accurate copy of the full DNA sequence throughout its life.

The fact that every child’s cell has a paternal chromosome paired with a maternal chromosome means that when DNA testing is conducted a double result is produced, one from each chromosome.  Therefore, during any DNA test, if we know a child’s pair of chromosomal types, it gives us information about the DNA types of the parents; we know that half of the child’s types will match with its mother and the other half will match with its father. This fact leads to the scientific basis of paternity testing. Look at this illustration depicting a mother, father and four children where the chromosomal types, called alleles, for a DNA test are given as 3, 6, 7 and 9.

Paternity Test Diagram

Because the mother must pass on one of her alleles to her children, all of her children must have either a 6 or a 9. Similarly, the father must pass on one of his alleles, 3 or 7. The children of this couple, therefore, can only be one of the four types shown. When a child is conceived it is completely random which of the maternal and paternal alleles he or she will inherit, it is like tossing a coin.

In paternity testing the tester will simply look to see if the supposed father matches an allele present in the child’s result. If he does not then that provides proof that he is not the true biological father and another man is. If the supposed father does match a child’s allele then this result provides positive evidence supporting the belief that he is the true father and a statistical calculation can be carried out. The statistical calculation looks at the probability that any unrelated man, picked at random, will have the child’s paternal allele by pure coincidence. This is calculated from the frequency with which the child’s paternal allele is found in the general population. If it is a rare allele then this leads to strong support for the view that the supposed father is the true father, if the allele is relatively common in the population then the support is weaker but still provides some positive evidence. If you are considering a paternity test many laboratories advise that you should submit a test sample from the mother as well as from the supposed father and child. You should comply with this request if at all possible because comparison of the child’s result with the known mother establishes, with certainty, which of the child’s alleles is the paternal allele. This, in turn, will lead to a much stronger statistical result and a clearer outcome. Usually you have to pay no additional charge to test the mother as good companies want to give you the best result possible.

For paternity, STR tests are most often used because there is a lot of variation between unrelated individuals. In other words the individual alleles tend to be relatively rare and STRs tend to provide strong statistical results when the supposed father matches all the paternal alleles and cannot be ruled out. In addition, STR tests are accurate, quick and easy. They work just like in the illustration above except that many different chromosomal DNA sections (called loci) each containing an STR will be targeted and tested simultaneously to increase the power of the test. Typically testers will analyze as many as 15 separate STR loci, and, if there is a match, quote a probability of paternity often greater than 99.99%. In general terms the more STR loci tested the better and the less room there is for doubt. The image below is part of a test for one individual and shows results at 4 different STR loci (these loci are called: D19S433, vWA, TPOX and D18S51). The numbers in the boxes give the number of repeats for each STR allele. At TPOX this person only has a single result because he inherited an allele with 11 repeats from both his mother and his father.

STR Profile B&W ed2

STR testing works well for paternity and is very reliable most of the time, however, occasionally, a natural phenomenon known as mutation changes the DNA allele type and has an impact on the interpretation of the result. DNA mutation, if it occurs, is a rare, but naturally occurring, mistake that takes place when DNA is being copied during cell division. It is a copying error. A mutation can change an STR allele, for example, it could change a type 3 allele into a type 4 allele. If such a mutation occurs when a new sperm cell is being made and if the mutated sperm goes on to fertilize an egg, then the resulting child’s paternal allele will be type 4 when the father’s type is 3. This would be a real problem if mutations were common but fortunately they are not. The precise mutation rates are different for different STR loci but in many hundreds of tests we’d only expect to see one or two mutations. If one, or more, mutation does arise it will either reduce the power of the statistical result or render the test inconclusive. It is extremely unlikely that the true father will be wrongly excluded because of mutation provided you use competent testers. If a mutation is suspected then best practice is to conduct more DNA tests to confirm one way or the other. The best testing companies have many back-up DNA STR tests that they can draw upon if their routine test needs more power because of a mutation. They usually conduct extra tests at no extra charge to ensure you get a clear result one way or the other (see How to Choose A Test Laboratory).