Thursday, January 02, 2014

CSI on the fly

A few days ago, Steve linked to RD Miksa's post on the reliability of eyewitness testimony. It's well worth the read. I agree with and appreciate most of what Miksa said on eyewitness testimonial evidence.

I'd also like to add what little I can to what Miksa said about eyewitness testimony often trumping scientific evidence.

People generally think of DNA analysis in forensics as highly reliable. Of course, reliable can be relative depending on what we're comparing.

Generally speaking, I'd agree DNA analysis is quite reliable. However, there are also limitations that aren't often discussed.

  1. History.

    a. First, a bit of history. This history isn't really necessary, but it might make this post slightly more interesting to read for some people.

    b. People often think James Watson and Francis Crick discovered DNA. Actually, Friedrich Miescher discovered DNA back in the 19th century, although he called it nuclein.

    Nor did Watson and Crick figure out DNA contains genetic information. That was already known in their day.

    What Watson and Crick did was figure out what DNA actually looked like. They figured out its double helical structure and so forth. Their discovery occurred in 1953.

    However, Watson and Crick stood on the shoulders of giants in their discovery. And I think far more so than Newton did, to whom the phrase is usually attributed. I've already mentioned Miescher. But Watson and Crick also relied on the work of Albrecht Kossel, Phoebus Levene, Oswald Avery, Colin MacLeod, Maclyn McCarty, Erwin Chargaff, Linus Pauling, Rosalind Franklin, and others.

    In fact, Franklin made significant contributions that led to Watson and Crick discovering the structure of DNA, including helping Watson and Crick eliminate their original idea that DNA might be a triple helix. She deserved the Nobel Prize, but since she passed away at the young age of 37, and Nobel Prizes aren't awarded posthumously, she never received it.

    c. Fast forward to the Human Genome Project (HGP). The HGP was declared completed in April 2003. This bulk was done under teams headed by Francis Collins (public) and Craig Venter (private).

    The first DNA sequence was not from a single individual but compiled from several different people. Although to date, only a very few individual human genomes have been fully sequenced.

    At any rate, the HGP accomplished its stated goals. These included: identifying all 20,000 to 30,000 genes; determining the sequence of the approximately 3 billion base-pairs of DNA; and storing all this information in databases.

  2. What is DNA?

    a. A simple definition is that DNA is information. Genetic information. DNA is like the source code in software. It is the template for how to build a human being. Other molecules "read" DNA, copy DNA, and so forth. But DNA is the blueprint out of which we're built.

    b. A more technical definition focusing on its structure is DNA stands for deoxyribose nucleic acid. DNA is a double helical structure consisting of two anti-parallel (running 5'-3' and opposite) sugar-phosphate backbones held together by ester bonds on the outside and nitrogenous base-pairs held together by hydrogen bonds on the inside. The base-pairs are in the keto tautomeric form. The base-pairs are adenine, guanine, cytosine, and thymine. Adenine and guanine are purines whereas cytosine and thymine are pyrimidines. The base-pairs lie flat like plates and are spaced .34nm apart, with 10 base-pairs per turn and each turn 3.4nm long. The diameter is 2nm.

    c. An adult human being typically has trillion of cells. Estimates vary, and sometimes vary significantly, but we can say there are around 50 trillion cells in an adult human being. Most of our cells contain DNA. So we have trillions of DNA in us, so to speak, out of which forensic scientists can collect DNA samples.

    d. DNA typically exists in the nucleus of a cell. But there's also DNA which exists in the mitochondria of a cell known as mitochondrial DNA (mtDNA).

  3. What is a gene?

    a. A gene is basically a section of DNA with instructions to make particular molecules (proteins). If DNA is the entire code, then a gene is a specific piece of the code. If DNA is a cooking book, say a cooking book on how to bake a cake, filled with a series of elaborate and complicated instructions to bake the best cake in the world, then a gene might be something like a single step or one of those instructions.

    b. Human DNA contains approximately 20,000 to 25,000 genes. And approximately 3 billion base-pairs.

    c. However, DNA also contains sections that aren't genes. In fact, the vast majority of DNA is non-coding, which is sometimes dubbed "junk DNA" by the media. However, the phrase "junk DNA" reflects media sensationalism more so than it does scientific knowledge and understanding. But that's another topic.

    The reason I bring up other parts of DNA is because we can also compare these other parts of DNA in one person with the same in another person.

  4. Forensics.

    a. There are several different ways to undertake DNA analysis. What I'll offer is one of the simpler and more generic versions.

    b. The first step in DNA analysis in forensics is, of course, to obtain a DNA sample. This can be done from many different parts of a human body. Common sites are a person's saliva, semen (sperm), blood, nails, bones, teeth, hair.

    But a couple of qualifications are in order. Generally speaking, there's no DNA in the actual piece of hair. Rather, the DNA needs to be gathered from the root of the hair, i.e., the hair follicle. Similarly, blood contains red blood cells as well as white blood cells, among many other things. Mature red blood cells normally do not have a nucleus, and thus no DNA. So the DNA from blood is typically from white blood cells.

    c. Next, DNA is extracted from the sample, and then purified.

    d. Usually more DNA is needed. So a process known as polymerase chain reaction (PCR) is used to produce more DNA from a small sample. In fact, PCR can produce millions of copies of DNA (or a section of DNA) from a single DNA (or a section of DNA) in a matter of hours. It's quite inexpensive too.

    e. I think the following is important to note, and I'll try to keep it simple, although it might still be a bit too technical. But people can skip this part without losing the main point.

    It's important to at least mention variable number of tandem repeats (VNTRs), and in particular a type of VNTR known as a simple tandem repeat (STR). This is because scientists need to have genetic markers in order to compare DNA samples. STRs are probably the most commonly used genetic markers. Others include single-nucleotide polymorphisms (SNPs) and mitochondrial sequences in mtDNA.

    STRs are constituent parts of DNA. They're DNA sequences that repeat, which in turn helps determine the length of the DNA. Plus, STRs can vary in length between individuals. So PCRs which target STRs can be used in DNA analysis by amplifying different and variable regions in a person's genome.

    f. Once we have millions of copies of the same DNA, we can begin to create a DNA or genetic profile. This starts by a process known as gel electrophoresis. What gel electrophoresis primarily does is separate larger and heavier molecules from smaller and lighter ones.

    g. Now let's put all this together.

    Say there was a murder.

    Say all the DNA is from blood samples.

    Say we have a sample of DNA from a crime scene. Say we have the DNA from the victim. We have the DNA of two suspects. Thus we have four DNAs in total: the crime scene DNA; victim's DNA; suspect 1's DNA; and suspect 2's DNA.

    The DNA is extracted, purified, and PCR performed on the relevant parts of each DNA sample.

    Then, when we use electrophoresis on each of these three DNAs, each of the DNAs is pushed through a gel via the use of an electrical current. The shorter bits of DNA will be able to move through the gel at a greater rate, while those that are larger will not. This enables the separation of the DNA according to size.

    The result will be something like this:

    The question is, which of the two suspect's DNAs match the crime scene DNA? As we can see by comparing each of the suspect's DNAs with the crime scene DNA, suspect 1's DNA matches the crime scene DNA. Hence, suspect 1 is determined to be the culprit.

    To reiterate, all this is just a basic and simplified version of the process.

  5. Limitations of DNA testing.

    a. On the face of it, DNA analysis seems highly reliable. And it is quite reliable. However, there are some limitations which can potentially impact its reliability. I'll quickly run through common ones.

    b. Sample quality.

    It's possible the DNA sample has been contaminated - unintentionally or intentionally. Contamination can occur at any point during this process. Like when the DNA sample is found, collected, stored, transported, handed over to a forensics lab, analyzed, etc. As for intentional contamination, given methods like PCR which aren't exactly difficult to perform (e.g. a high school lab could have all the necessary equipment and a motivated high school student with average intelligence could learn to do it), it's possible for someone to replicate another person's DNA and plant it at a crime scence, or at a later point, thereby falsifying evidence.

    It's possible for a DNA sample to have degraded. Scientists could run PCR on a degraded DNA sample, but it will still reflect its degradation. If this occurs, then the police will likely hope a scientist(s) can accurately interpret the results. But this then would depend in large part on the acumen of the scientist in interpreting poor quality results. Not all scientists are equally adept at interpreting results.

    c. Laboratory effectiveness.

    Oftentimes police departments rely on a local forensics lab(s) to run their DNA tests for them. But what's the quality control at the lab like?

    It's possible there are laboratory errors. Mistakes by scientists. Such as faulty lab procedures, mislabeled specimens, transcription issues, etc.

    It's possible DNA samples from different criminal investigations could be mixed up at various points. If DNA samples are mixed up, it could also generate irrelevant DNA, which again could be misinterpreted by scientists.

    d. Test reliability.

    Every test, including DNA testing, has what's known as sensitivity and specificity measures. Basically, these are measures that indicate how reliable a test is based on true and false positives and true and false negatives. The ideal test will be 100% sensitive and 100% specific.

    These in turn are related to positive and negative predictive values.

    DNA testing in an ideal setting can have 98% or greater sensitivity and specificity. But one problem is not all labs that run DNA testing reach such high rates.

    The fact that DNA testing doesn't produce 100% sensitivity and specificity even in an ideal environment means there will always be a small percentage of false positives and false negatives.

    e. Genes.

    Of course, family members sharing the same parents are far more likely to have a similar genetic profile than complete strangers. Identical twins can have identical STR patterns for example.

    f. Chance.

    It's possible two unrelated individuals may have the same DNA profile sheerly by random chance or coincidence. This is more true if DNA samples compared only compare a small number of parts of DNA (i.e. loci). The more of DNA that's compared, the less likely this will occur. But how much DNA we have to compare can depend on many uncontrollable factors.

No comments:

Post a Comment