When DNA Evidence Doesn’t Prove Much

Some criminal defense bloggers are skeptical of the supposedly high reliability of DNA evidence. Since almost every other form of forensic testing is based on shaky scientific foundations, they figure it’s only a matter of time until DNA falls too.

I don’t share their skepticism, at least not about the basic principles of DNA profiling. Other forensic sciences — blood spray patterns, bite mark analysis, psychological profiling, ballistics matching, and even fingerprinting — were invented for use by the criminal justice system. The science and technology behind DNA profiling, however, were developed as part of the general study of medicine and biology, using traditional methods of scentific investigation, and the conclusions have been validated many times over by subsequent research. I have little reason to doubt the basic validity of today’s standard DNA identity matching.

My more careful readers will have noticed I slipped a few weasel words — “today’s” and “standard” — into that last sentence. That’s because I still expect to see some problems in the courtroom. I’m not just talking about lab errors, but about the use of DNA in ways it was never intended, or in ways that have not been validated.

Case in point, the trial of Cleveland Barrett:

…Assistant State’s Attorneys Krista Peterson and Jane Sack told jurors in closing arguments that the DNA obtained from the victim after the alleged incident in July 2010 was a match to Barrett’s genetic profile and evidence that corroborated the victim’s trial testimony.

“Who is the major profile in the DNA that’s found?” Sack asked the jury, according to a transcript from the trial. “The defendant.”

But this DNA was different. It was not from semen, as is often the case in rapes; instead it came from male cells found on the girl’s lips. What’s more, the uniqueness of the genetic link between the DNA and Barrett was not of the 1-in-several-billion sort that crime lab analysts often testify to in trials with DNA evidence. Instead, when Illinois State Police crime lab analyst Lisa Fallara explained the statistical probabilities, she testified the genetic profile from the cells matched 1 in 4 African-American males, 1 in 8 Hispanic males and 1 in 9 Caucasian males.

Barrett was acquitted after spending a year locked up pending trial.

This case illustrates a couple of potential ways to misuse or abuse DNA evidence. (I can’t tell from the news story what other testimoney there was against Barrett or how any of the evidence was found, so I’m sticking with the word “potential.”)

First of all, there’s the fact that DNA was recovered from an unknown source on the outside of the victim’s body instead of from something more incriminating, such as semen in an orifice. Given that all of us shed microscopic skin cells all day long, I wouldn’t be surprised if everyone has someone else’s DNA on them somewhere. As DNA recovery techniques get better, forensic labs will probably recover more and more of it from crime victims, but it won’t necessarily prove anything.

As for that great DNA science I mentioned earlier, that was about the DNA itself, not about the methods for recovering it as evidence. All those biologists and doctors never had any reason to investigate how DNA traces are transfered between people. The probabiliy of a given allele at a given loci is well known. The probability that something like shaking hands or touching the same lightswitch will tranfer detectable amounts of one person’s DNA to another…it might happen a lot, or it might not happen at all. In either case, I’m pretty sure it hasn’t been validated nearly as thoroughly as basic DNA science.

(If you think cops and prosecutors wouldn’t stoop to misusing DNA evidence that might not prove anything, let me remind you that they use the presence of traces of cocaine on money as probable cause for searches and forfeitures, even though it has been known for decades that most U.S. currency has cocaine on it.)

The second potential misuse of DNA evidence is the lowering of the statistical standards. Normal DNA profiles are based on 13 locations on the human DNA strand. With a suspect in custody, or his DNA on file, the lab will determine what alleles (DNA patterns) he has at each location. Then they will attempt to make the same determination for the evidentiary DNA from the crime scene.

The problem is that the DNA from the crime scene is often degraded, and not all 13 genetic loci can be determined. Fortunately, not all 13 DNA loci have to match. Even if 3 or 4 of the evidenciary loci cannot be typed, the remaining loci are still good enough for a match at millions-to-1 odds.

On the other hand, if there’s a mismatch at even a single loci, that excludes the suspect completely: There’s no way he could have left behind DNA with alleles he doesn’t have. So the limiting factor for using incomplete DNA profiles is the unknown possibilities in the degraded sample. If 4 loci cannot be typed, that is 4 places where the DNA might not have matched if the lab had had a complete sample. The more unknown loci, the worse the evidence.

(Although it doesn’t appear to have happened here, partial DNA samples can also lead to coincidental matches in DNA database searches. Do a search with a partial sample and find five possible matches? Well, you know those are coincidental. But do a search with the same partial sample and get a single hit? That’s got to be the perp, right?)

So how did the Barrett prosecution end up using DNA evidence that offered only 9-to-1 odds? They got a match at one DNA location. This was little better than an old fashioned blood type match or having a witness who could only say that the attacker was “tall with black hair.”

Fortunately for Barrett, he had a vigilant lawyer:

Barrett’s case, in fact, involved considerable back-and-forth testimony over the meaning of the DNA. At one point, Fallara testified that she could not determine whether the cells on the victim’s lips came from saliva or skin or even a hair. She did acknowledge, though, that Barrett’s DNA matched the DNA recovered from the victim’s lips at only one of 11 locations on a chromosome – meaning she could not get enough information at the other 10 locations, a distinction that was drawn out by Assistant Public Defender Scott Kozicki in cross-examination.

Fallara even acknowledged in testimony that under a different reading of the data, Barrett could have been excluded as the source of the DNA recovered on the victim’s lips.

“It’s new. It’s highly unusual. And it’s very concerning,” said Gregory O’Reilly, chief of the forensic science division for the Cook County public defender’s office. “There’s a terrific power in the phrase ‘DNA match.’ And there’s a great risk that the jury will put great significance on this when it’s not significant at all.”

(I’m not sure why they talk about only 11 loci. Perhaps it was a non-standard test.)

In theory, there’s nothing wrong with using weaker DNA evidence as long as the weaknesses are explained to the jury. In this case the defense was apparently successful in that respect. It makes makes me wonder, though, how many other people will plead guilty to crimes based on DNA evidence that was no better than this? I’ll bet it’s already happened. We’ll know soon enough when they start filing appeals for ineffective assistance of counsel.

(By the way, next time you hear a private lawyer bashing public defenders, ask him if his law offices have a forensic science division.) 

One Response to When DNA Evidence Doesn’t Prove Much

Leave a reply