I am genuinely trying to understand your points on this, so I did spend an hour re-watching Dr. Larkin's testimony.
This really has nothing to do with BK--she wasn't trying to say his matches suggested a certain relationship and that's not always reliable so the IGG results are not reliable.
Dr. Larkin wasn't talking specifically about BK because they did not have any DNA evidence to examine at the time. She WAS talking about possibilities that could happen from the processing of DNA and why it was necessary for the defense to be able to examine the evidence. She also spoke about how relationships are determined in IGG which is a combination of science and also experiential judgement, she talked about problems that which can occur with developing an IGG. However, I don't think any of this would have been brought up if AT didn't have questions about the DNA in this case.
She was talking about the percentage of dna shared between matches, which are called centimorgans, or cMs. When you have a match list, it will tell you how many cMs the matches share with the profile you're working with. Some cMs are pretty obvious--if you share in the 3400 cM range, this is almost always going to be parent/child.
But if you share around 800 cMs, this could be a first cousin or a half niece. I have a match that ancestry suggests based on shared cMs is my first cousin once removed or my half first cousin, but I know through research he's actually my second cousin.
And that's all she's really saying. The amount of cMs shared can suggest the likely relationship, but many times you'll share more or less than the usual amount so you need to do your research. But the cMs are useful in genetic genealogy because when you map out the tree you can use the cMs to kind of check your work--if I have this person over here in this branch, do the shared cMs make sense with his match over here in the other branch. Do all the shared cMs on this tree make sense for how they all relate to each other.
I don't think she was suggesting this. We know there was enough dna to do IGG. AT wouldn't even be holding this hearing if there wasn't enough dna to do IGG. There would have been no match list, no tip generated, no use of the genetic genealogy sites, etc. Everything she's asking to look is the product of the IGG.
A microarray is generally used for a variety of reasons, but whole genome sequencing/next generation sequencing can be used for smaller amounts. If the dna produced matches and those matches led to BK, there was enough dna to do IGG.
Are you saying AT thinks LE got the dna illegally through the IGG?
No, I'm saying if Blum was right, there were not enough diploid cells to get a complete STR profile. A partial profile could lead to a false identification. If it was a partial profile, and the identification was false, then the IGG would be wrong, too.
If she thinks LE got the dna illegally via the IGG, then she doesn't think there wasn't enough dna to do IGG.
Yes. Keep in mind neither Dr. Larkin or Gabriella Vargas have seen any details from the investigation. Dr. Larkin testified that there are some workarounds in GEDmatch to be able to see matches that didn't opt-in to be visible to law enforcement. That's a violation of their terms of service. She has seen other genealogists do this and knows how to do it herself. We don't know if LE in this case used that workaround but AT definitely wants to find out in hopes of getting the dna thrown out. Because it's a match to BK and she knows it. And remember also that Gabriella Vargas reneged on her testimony and admitted that she inadvertently agreed to some of what she signed without fully reading it. I'm not sure she will be called again.
Both points are not at issue because one point is the dna on the sheath being a match to BK. This has not been disputed by the defense and they have made no issue of it.
The other point involving dna is the IGG process used to generate the lead that the dna came from this family and likely from BK. The lead itself is not the issue though. The issue is the defense wants to look at the techniques used and sniff out any violations of terms of service.
Yes, before the defense experts were speaking in hypotheticals. They had no knowledge of what actually occurred in this investigation. AT wanted the court to see why it was important for her to examine the materials the FBI believed to be investigative techniques and not subject to discovery.
I have no idea what might or might not be exposed regarding the techniques and processes used by LE to generate the match list--people in general are not happy when LE doesn't follow the rules and most users of genetic genealogy are not happy when something discourages people from using these services.
But if the IGG was wrong--what are we doing here? Why try so hard to get it thrown out if it's wrong? The IGG has nothing to do with the dna on the sheath being a match to BK. The IGG doesn't determine in any way that he is or isn't a match. The IGG could be wrong and wrong again all day long and he would still be a match.
JMO
IMO, IF the DNA was partial due to being degraded by sitting on brass, and the ID was made on the partial DNA, THEN, both the STR profile and the IGG developed from the STR profile could be completely wrong. It would all come down to GIGO (Garbage In, Garbage Out.)
Howard Blum wrote that the sample was "20 skin cells (maybe less.)" in Eyes of a Killer, Chapter 4. All the literature I can find online about how many skin cells are needed to get an STR indicates that normally 80 skin cells are needed. The literature also indicates that some of the cells are destroyed in processing. So I have to really wonder how good is the STR in this case? Was it complete or partial?
Just to put into perspective how tiny the sample found on the snap of the sheath was: "A single square inch of skin has about 19 million cells."
Skin Fun Facts: Did you know…? Your skin is your largest organ and plays a vital role in detecting hot and cold, regulating your body temperature and
pdskin.com
"Currently, the optimum DNA input to the PCR for STR profiling is around 500 pg [[1], [2], [3]], which equates to approximately
80 diploid cells (∼6 pg/cell [4]). If the potential of DNA loss through workflow processing is not considered, any item from which 80 cells are collected should generate a full DNA profile."
If Blum is right they collected only 20 diploid cells or less.
"Many laboratories use commercially available STR amplification kits. Depending on the kit and reaction volume, the optimal concentration of input DNA will be in the range of 0.5ng – 2ng.
Adding too much or too little DNA to the amplification reaction can result in problems in the analysis."
Current forensic DNA analysis uses polymerase chain reaction (PCR) based short tandem repeat (STR) testing
nij.ojp.gov
A further problem with such a minuscule sample is that the entire sample would be used up in the testing process leaving the defense without any option to have another lab run the same testing process on part of the sample to see if the results are identical or not. In this case, we have a prosecutor who initially told the court that they were going to have to "trust them" on the DNA identification. From that point forward he has proceeded to drag his feet about giving the defense any information about the DNA testing process at all. IMO, it's not a good look nor should it be tolerated. A fundamental of the scientific process is the reproducibility of results - that you have to be able to replicate the results to prove if the process worked correctly or not.
Here, the sample is too small to give half of it to the defense, so the next option would be to give complete documentation of the findings and then the defense can run BK's DNA to see if it matches LE's findings.
What is going on in this case is one of the reasons why I don't like non-scientists being in control of scientific evidence. In this case we have one lawyer trying to hide a scientific process from another lawyer and her DNA experts. It is really ridiculous. There should be no reason for that whatsoever. Either the scientific process worked correctly or the results were questionable or it did not work. Regardless of which outcome, it must be completely exposed and examined. Further, the Judge, also a non-scientist, got involved in deciding what parts of a scientific process the defense may view. How would JJ know what is Germaine about the testing? This is certainly not his area of expertise.
IMO, we need to do better. Our court system owes that to all of us.
