"There were two separate DNA samples that could have led to Avery's release. The first one proved inconclusive. As the Innocence Project, which took on Avery's case, reported, a fingernail scraping sample showed that there was unknown DNA underneath Beerntsen's nails. But as the Neflix series explained, DNA in that situation can only be broken down into groups called "alleles." And though there were alleles that matched Avery, there were also alleles that could not have belonged to him. However, because he was a match for some of the DNA, the Wisconsin Court of Appeals found that 'the DNA evidence does not make it any more or less probable that Avery assaulted P.B.'"
Link.
Oh for the love of Benji. This looks like the writing of someone who doesn't have the faintest clue about what DNA testing looks like.
DNA is not broken down into groups called alleles, alleles are variants of genes we all carry. DNA testing looks for which variants of certain genes you carry and determines what is the probability there is another person in the population with the exact same makeup. Forensic testing doesn't actually look at genes, it looks at tandem repeat sequences in front or behind a coding region. The numbers in the result are the number of repeats in that allele of your genome. We know, from testing, how common a certain number of repeats is and can infer what is the probability another person has the exact same combination as you do. The usual goal is to go to above 10
-13, i.e. you'd expect fewer one person in every 10,000 billion people to have that genetic make-up, so it is incredibly unlikely you have two people with that particular make-up. The only exception is identical twins, even inbreeding for several generations doesn't result in identical profiles with the kits we have - we test for a sufficient number of alleles, the population will be genetically unviable before you get people with identical profiles.
In theory one allele that shouldn't be there excludes your suspect as the culprit, although this is not strictly the case in rare cases of heteroplasmia or if you have a weak sample where non-specific binding became important. This is where those 300 RFU we talked about previously come in, this is the limit where the manufacturer guarantees the results had to come from specific binding. The other possibility is that you're dealing with a mixture of DNA from two or more people.
If the DNA profile is a match for Steven Avery but there are other alleles also present that he does not carry it is possible the sample is a mixture of DNA of Steven Avery and another person. Statistical tools exist to calculate the odds Steven Avery is indeed the donor of some of the DNA, you can get them above 10
-10 in many cases, which should be enough to go beyond reasonable doubt. If all of his alleles are present and there are other alleles also present which implicate a mixture of three people or less, the calculation will show he was a likely donor. If none of his alleles are missing the DNA can not exclude him as a donor.
If, however, a
single allele Steven Avery has is missing from the DNA profile, he is all but excluded as the donor. Two missing alleles exclude him beyond any reasonable doubt. This is only true if other alleles in the portion of the sample are above the 300 RFU or whatever other limit the manufacturer specified.
In other words if you want to exclude Steven Avery you need to look for what is missing, not what is present.
It would be nice to see the DNA data for ourselves, but the court's reasoning looks...dubious.
I hate to break it to you, but my take is the court knows a whole lot more about DNA testing than whomever wrote that article. It could well be the courts' decision is scientifically sound and the fingernail scrapings incriminated him more than they exonerated him.
McHrozni