5
$\begingroup$

How do we know that everybody's DNA fingerprint is unique?

I know, I know, everybody's DNA is unique.

But when we do DNA fingerprinting, we're looking at very specific regions of high variability.

How do we know that just by chance, two people's DNA could be the exact same in the spot(s) we're observing.

With real fingerprints, uniqueness is (almost) guaranteed (I believe it's a 1 in 64 billion chance) because they are developed by physical stresses on the fetus in the womb.

Is there any similar mechanism/affect that acts on these non-coding/highly variable sections of DNA?

Thanks!

evamvid

Improve this question
$\endgroup$
4
  • 1
    $\begingroup$ Yes, the fact that they are highly variable. I don't have the numbers at hand but while it is theoretically possible for two DNA fingerprints to be identical, I would expect a similar or smaller probability of this than for "real" (let's call them digital) fingerprints. $\endgroup$
    – terdon
    Commented Feb 27, 2014 at 2:20
  • $\begingroup$ @terdon one meaning of the word digit is finger :) $\endgroup$
    – user13107
    Commented Feb 27, 2014 at 4:58
  • $\begingroup$ @user13107 yes, that's why I used it. $\endgroup$
    – terdon
    Commented Feb 27, 2014 at 5:48
  • 1
    $\begingroup$ Essentially, for the same reason we "know" that no one is going to be struck by lightning 100 times. It is possible but highly unlikely. $\endgroup$
    – Bitwise
    Commented Feb 27, 2014 at 13:50

2 Answers 2

Reset to default
5
$\begingroup$

First, remember that identical twins actually have the same genotype. So its not exactly true that everyone has a unique genome.

But to get at the heart of it, you're asking how I can be sure that I have a different genome than you, or even than say my brother. And moreover you're asking how these differences are obvious enough that they can be detected without whole genome sequencing.

The answer is that there is no mechanism to ensure differences, but that the differences arise from the enormous variability of our genome. Look at the major histocompatibility complex, or MHC, a cluster of genes that affect our immune system.

Just to simplify things, lets focus in on four genes in this cluster: HLA-A, HLA-B, HLA-C and HLA-DRB1. According to this site there are 2579, 3285, 2133 and 1411 alleles for each, respectively. If we make a naive assumption that each of us received a random assortment of these genes, that makes 2579 x 3285 x 2133 x 1411 = 25,549,791,000,000 possible combinations. The chance that we both have the same combination is then about 1 in 25 trillion, just considering 4 genes.

But wait, you say, surely some of these genes are not equally likely and surely they segregate together, making this chance much smaller. You would be right. But this consideration stops mattering very much when you account for the rest of the variable regions. They combine multiplicatively, so considering a fifth gene with only 2 possible alleles brings the possible combinations to 50 trillion.

If you consider just a few more genes, the numbers get unthinkably high. This pretty much ensures that, even considering a small sequence, we can be sure that the genetic fingerprint is unique.

Improve this answer
$\endgroup$
3
  • $\begingroup$ That said, you would need to compare one very large number to another to evaluate overall uniqueness. For example, although there are only 6 or so billion people in the world (i.e., approximately 10^9), The relevant divisor for overall uniqueness would be the number of possible pairwise comparisons between people (e.g., around 10^17; which is much larger than a trillion). $\endgroup$
    – Jeromy Anglim
    Commented Feb 27, 2014 at 6:32
  • $\begingroup$ Identical twins don't have identical genotypes. They have very similar genotypes, but there will typically be a few dozen base pairs that differ due to mutations. These usually won't be in the regions used for DNA fingerprinting, though, and will require full-genome sequencing to spot. A complicating factor is that the error rate for sequencing is similar to the mutation rate, so you 'll need to sequence multiple times to be sure you're seeing the actual mutations. $\endgroup$
    – Mark
    Commented Feb 27, 2014 at 7:55
  • $\begingroup$ Thats not true. Twins do not have exactly the same genotype, since they occured some individual variations. Its possible though, that the standards used for genetic fingerprinting in crime scenes is not precise enough to differentiate twins. This has happened in Germany and led to the release of a subject from prison as he couldn't be distinguished from his twin brother. With more test points (which was not allowed at this time) the differentiation woud have been possible. $\endgroup$
    – Chris ♦
    Commented Feb 27, 2014 at 8:23
3
$\begingroup$

Possibly partly offtopic, but I'd like to add that when using dna fingerprint as evidence of identity in a legal investigation, the odds become smaller due to combination of human factors and reduced accuracy. There is the twin issue already mentioned, but in addition there are laboratory errors and the possibility of contamination with other evidence. Samples may be damaged in other ways as well, and dna samples can be planted.

In the analysis done by Jonathan Koehler, it was found that around 1.2% of tests used in this context gave out an incorrect match. Additionally, in this other study by him it was found that the way results are presented has a considerable impact as well.

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .