How do you know if DNA sample A matches DNA sample B? Aren’t there all those T and C and A and G things in lots of long complicated combinations? How can you find out if DNA matches or not?
In: Biology
Essentially they dye the DNA bright blue and put it in small wells carved into this kind of gel slab and then send an electric shock through it in one direction. The shock moves the pieces of DNA down the well in large and small chunks, like this
I I I I I
Or a similar pattern. Different people’s DNA strand break and move differently
Then they do the same thing to the dna of suspects, and whatever lines line up with the one they were testing is the match
It’s kind of the old fashioned way but it’s tried and true
Edit okay posting is really messing up the spacing of the bars and making it looks even but they’re not they’re staggered sorry
They usually have specific fragments/areas of interest that they look for, and the more of these fragments match the closer of a match it is. They don’t usually sequence the entire DNA just to match samples within the same species.
I answered a similar question a while ago, which I’ll copy here. Hope it helps!
>The vast majority of your DNA is identical when compared to all other humans, but some sections are not. In order to be able to identify individual people, we have to use the parts of our DNA that are most highly variable and can change relatively quickly. However, there are a few different types of DNA that can be used for this.
>
>For some time now, both forensic and paternity DNA tests have mostly used sections of DNA called short tandem repeats (STRs, also called microsatellites). These are composed of the same DNA sequence repeated several times in a row. Since they aren’t part of actual genes, their exact sequence doesn’t really affect anything, and the number of repeats can easily [increase or decrease by chance](http://www.web-books.com/MoBio/Free/images/Ch7F3.gif) when DNA is copied. One example is [TPOX](https://strbase.nist.gov/str_TPOX.htm), which is found on chromosome 2. As you can see there, this region can have anywhere from 4 to 16 repeats of the four-letter sequence AATG. All people have two copies of this region (one from each parent), which may be the same length or different lengths. We can easily measure the number of copies by running DNA on [a gel which sorts it by size](https://d2vlcm61l7u1fs.cloudfront.net/media%2Fb51%2Fb5148b21-e053-4458-90c2-23f7088c00b5%2Fphpy8aPjA.png) and comparing it to other pieces of DNA we already know the size of. If both of someone’s copies are the same length, then only one band will appear, but there will be two bands if their copies are of different lengths.
>
>Of course, it’s possible for two people to have matching TPOX regions just due to random chance, so we look at multiple different pieces of DNA that are unconnected. The FBI uses a system called [CODIS](https://www.fbi.gov/services/laboratory/biometric-analysis/codis), which originally consisted of 13 regions but has recently been upgraded to 20. Since we can measure how common different repeat numbers are in the population for all of these regions, we can estimate the rarity of any given combination. Based on the original set of 13, a single person’s particular set of copy numbers will typically be somewhere around a 1 in a billion combination, and the chances of any two random people perfectly matching are something like 1 in 500 trillion ([source](https://www.nature.com/scitable/topicpage/forensics-dna-fingerprinting-and-codis-736/)). The upgrade from 13 to 20 markers significantly pushes the odds of false conclusions down even further. So basically, if we have two samples that match at all regions, it’s almost definitely the same person (or their identical twin), and similarly matches at half of the regions would strongly support a close family relationship (e.g. parent-child or sibling).
>
>More recently, various companies have begun examining larger amounts of DNA for the purposes of ancestry and familial information. This technique works slightly differently. Instead of using just a handful of repetitive regions, these techniques look at a much larger number of individual DNA letters which are variable in humans (often called single nucleotide polymorphisms, or SNPs). The exact number varies between DNA tests, but as an example 23andMe uses between 500,000 and 1,000,000 depending on the kit version ([source](https://isogg.org/wiki/23andMe)). Unlike with STRs, where there are many possible states (e.g. anywhere from 4-16 copies for TPOX as discussed above), for SNPs there are typically only two options (e.g., an “A” or a “G”). This means that for any single SNP, there is a pretty good chance that two people could just match randomly instead of as a result of being related, but by using a huge number of them that are evenly distributed across all chromosomes, it quickly becomes very easy to tell if two given individuals are relatives.
Latest Answers