Ever since 1953, when the DNA double helix structure was discovered by American scientist James Watson and British scientist Francis Crick, working together at Cambridge University, the world of molecular genetics has never been the same. DNA turned out to be a polymer (a large chain of repeating molecules), and a particular type of polymer at that, called a nucleotide (sugar and phosphate in a nitrogen base), and furthermore, the nitrogen-containing base (connecting the two sides) was arranged in the form of a palindrome (a series of letters or numbers that reverse themselves; ABCDEEDCBA or 1234554321). It therefore wasn't necessary to map the whole length of the double helix, which if stretched out would be six to nine feet long. It was only necessary to slice off and look at fragments or strands with one side of the helix where palindromes began and ended. Any one of these fragments, in theory, would be polymorphic enough (contain enough variation in form) to unlock the genetic code (genome) in its entirety. A few single strands should define how the other strands would look because each side of the helix was an exact complement of the other side, held together by what is called base pairing, the predetermined, palindromic fashion in which base molecules bonded it all together. It didn't take long (1970 actually) for scientists to develop enzymes (called restriction enzymes) that would slice off fragments of a base pair at selected points. They then discovered that under electrophoresis, smaller fragments migrated toward the positive electrode, producing an autorad (short for autoradiograph) with fragments lined up from small to large.

DNA typing produces what is called a random probability match -- sometimes as high as one in several million, sometimes as low as one in a hundred. It is defined as the probability of a match between a sample left at the crime scene and a suspect. Different methods of calculation produce different results, but it all depends on comparing sets of bands or spots on autorads. That is to say, there are two separate, but related issues: (1) the statistical methodology; and (2) the matching criteria. The latter is the closest thing to a margin of error, and it's usually stated in terms of requirements that the analyst observe "no more than a 5% difference in bands on the autorad" before declaring a match. Some experts have said that a 2.5% criteria would make a better cutoff point Humans have 46 chromosomes in their DNA, each one containing about 550 genes, and each gene containing as many as 35 alleles. A chromosome contains two complete sets of all genetic material. We're not talking about the two sides of a DNA molecule, just the chromosomes on a DNA fragment. At certain locations, there will be genes containing the same genetic code as the other side. These things we call "genes" are really locations (loci) on chromosomes. A thing called an allele is a specific form of a gene, an alternative form of a genetic locus. Alleles at locus points are inherited from parents, and there as many alleles as there is diversity in your family background. Chromosomes have the same loci all the way along their length, but may have different alleles at some of the loci. It's the alleles, not the chromosomes or genes, that are dominant or recessive, and give us things like eye color and blood type. The genes just give us medical diseases and other functionality. Alleles are characterized by slightly different nucleotide sequences and are distinguished by their different phenotypic effects. There's no known laboratory technique for determining a person's complete genotype (their entire genome). It can only be estimated by studying their family history or the outward characteristics they share with others in the subpopulation they belong (their phenotype).

Juries are intellectually overwhelmed by DNA evidence, and understandably so. Nothing in forensic science, except maybe mass spectrography, is as complex and complicated. Certainly, nothing has ever before been capable of producing odds like one in seventy thousand trillion (which might be the population of our solar system if all the planets were densely populated). Except for identical twins, the chances of similarity in DNA are astronomical or infinitesimal. As the O.J. Simpson trial demonstrated, the average American citizen doesn't respond well to advanced topics in molecular genetics. The issue is what jurors (and the general public) should be exposed to, and that debate properly centers less on the practical matters of jury instruction (about reliability depending on quality) or how many visual aids (pieces of demonstrative evidence) are too many, but instead on the whole matter of forensic science ethics For a moment, let's ignore the idealistic task of keeping things within the competence of the jury or laymen, and focus in the "whistleblowing" clause (all Code of Ethics have them). What would motivate a forensic scientist to "disclose any...shortcomings indicative of innocence"? Does this mean that any expert put on the stand should be forced to testify about how vulnerable their lab is to contamination, mismanagement, and incompetence? Well, the answer is yes, and that is precisely what defense lawyers have been doing whenever they get the chance. In the post-Simpson era we now live in, juries are regularly exposed to this rather than a short course in molecular genetics. Badmouthing America's crime labs has become the way to attack DNA evidence.

Discovery is a pretrial process where each side (prosecution and defense) shares their lineup of witnesses, what physical evidence they have, and what trial strategies they plan to pursue. It differs from plea bargaining because charges and penalties are not discussed. Discovery is informal, and depends on the working relationship between prosecutor and defense (typically, it occurs over lunch). By law, the only thing that must be shared by the prosecutor is anything that might shed light on the defendant's innocence (exculpatory evidence). This is called the Brady doctrine. Discovery is not a constitutional right, but a privilege. Defense should not rely upon a fishing expedition for "all Brady material", and if the process becomes uncooperative, the judge orders something called disclosure, a type of court-ordered discovery. There's lots of other similar pretrial processes such as suppression hearings and, most importantly, Daubert hearings (for scientific evidence). DNA evidence doesn't lend itself well to informal discovery processes because even though it has the potential to be exculpatory, DNA lab reports aren't exactly the easiest thing to discuss, challenge, or replicate. The reports either say "inclusion", "exclusion", or "inconclusive". They bring an unaccustomed degree of certitude to the courts. It may very well turn out to be the technology that overpowers the criminal justice system.