(Caveat to all below. I’m learning about this. So info. could be wrong.)
Probabilistic Genotyping in some ways seems the easiest DNA analysis to understand. But it’s still kind of wonky.
DNA in general is complex and baffling and much of it is counterintuitive. The simplest way that I can think to explain what True Allele does is by way of the Wheel of Fortune letter board. But I’ll get to that later.
We are all genetic literature, comprised of words that form into sentences and then, paragraphs, and finally there is this novel that tells the genetic story of who we are. But most of the words in our novels are identical to the words in the novels of others, down to the order in which they appear: We are more alike than we are different.
What science has done is figure out certain types of words that distinguish us from each other. And they found about 23 of these kind of words that when strung together define who we are as individuals: This is our DNA Profile.
And so when identifying a person, or matching us to DNA found at a crime scene, that list of words is compared to the list of words found at the crime scene, and if they are a match, good chance we were there-or our DNA was transferred there by way of contact with another person or object.
In the actual science of DNA, the “words” or profile is comprised of pairs of numbers at what are called locus “markers” (plural is loci). Two numbers, one from each parent are identified at each of these locus markers. And these pairs of numbers are located together under a heading or type.
But let’s pretend that rather than numbers, we are talking about words that need to form a 15 to 23 word (marker) sentence.
With DNA from only one person (source), all the words of the sentence are present and are easy to read, and identifying who that sentence describes is a piece of cake. But as the science has progressed DNA testing has become so much more sensitive, revealing fragments of sentences (allele drop out from degraded DNA samples), or numerous sentences that have to separated out from one another in order to read them correctly (multiple contributors).
What a DNA analyst now often finds themselves analyzing are similar to a partially filled Wheel of Fortune letter board. There will be a few letters, or perhaps a few words (allele peaks-pairs at a marker) that can be viewed, but the complete sentence is not apparent (allele dropout). Or you can have a letter board where there is clearly more than one sentence, and this has to be unscrambled (multiple contributors).
All DNA analysis has always dealt with probabilities. For example, if a match is one in a trillion, that means you would have to meet a trillion more people before you’d find someone with that same DNA profile.
What probabilistic genotyping does is to fill in the blanks, or unscramble DNA where there is more than one sentence, or more than one person’s DNA in the mix utilizing known probabilities. This process is in some ways similar to what we do when we try and figure out the Wheel of Fortune puzzle answer.
Using our limited brain capacity, we scan, much like a computer, for all the possible word combinations that could solve the puzzle. We do this based on our knowledge of the probabilities of what that complete sentence has to look like. We work off of the words that we can view, and then eventually do fill in the blanks, and find the sentence that best fits.
With probabilistic genotyping, the possibilities that solve a puzzle where either there are alleles missing OR one that contains so many alleles that more than one person is being represented, are enormous. It’s a process of either filling in the blanks, or unscrambling and distinguishing each individual profile from the others. And without computers, we would be lost. There are simply too many variables for us to process this in our tiny little brains. It just isn’t humanly possible.
But the computer can do this. And when loaded up with statistics that allow it to do the math, it will navigate through gazillions of possibilities, based on the known probability of allele pairs and how they fit together in humans throughout the world.
The probabilistic genotyping software is programed to solve this extremely complex jigsaw. A puzzle that again deals with DNA profiles that are missing alleles or the sample has too many alleles, indicating that there is more than one person’s DNA present in the sample.
And though it sounds complicated, in some ways it operates exactly as we do when we approach more basic puzzles. We have a set of criteria. Many possible answers. Known factors. And limited number of solutions that can be correct. Obviously the more information we start with, the less experimenting or “guessing” has to be done to finally get to the correct answer.
These programs have been tested and validated. And though there are concerns, they do seem to get accurate results. Like any science, there has to be constant checks and balances. But people are currently being convicted by way of this type of DNA analysis, so why then can’t people be exonerated by way of it, as well?