THE ACHILLES HEEL IN DNA TESTING Series: DNA TESTS: A Reason for Doubt

This is the second of three editorials on the reliability of DNA results from the state’s forensics lab.

When Lawrence Napper went on trial in November 2001, accused of the kidnapping and sexual assault of a 6-year-old Houston boy, DNA evidence cemented his conviction.

Now, due to an exhaustive probe of the Houston police DNA lab, the results that helped sentence Napper to life in prison are unraveling.

According to the retests, the match between Napper and the DNA found on the abused boy is extremely weak. When the DNA was rechecked, there was a possible match at two points on the DNA chain, but not at 11 others. The jury was never told that the so- called “match” was statistically insignificant.

The tenuousness of the evidence against Napper, which has prompted his attorney to call for a new trial, exposes the little- known Achilles heel of DNA testing.

Widely viewed as an almost infallible forensic science tool, DNA evidence is only as good as the individuals who handle and interpret it. Particularly when DNA from more than one person is mixed in the same sample, which is often the case in sex crimes, analysis is an art as well as a science.

That fact crystallized in Virginia this summer when three nationally respected DNA experts told The Virginian-Pilot editorial page that the Virginia Division of Forensic Science apparently erred in its analysis of DNA evidence in one of the highest profile murder cases in state history.

Flawed interpretation appears to have trumped science in analyzing DNA taken from the body of Rebecca Williams, a Culpeper housewife whose murder led to the false conviction of Earl Washington Jr.

Washington, who spent 9? years on Virginia’s death row, was pardoned in October 2000 on the basis of state lab work. The lab identified convicted rapist Kenneth Tinsley as the source of DNA on a blanket at the crime scene. But, in the now-disputed result, the lab also excluded Tinsley as the source of DNA on Williams’ body, leaving the murder unresolved.

A scientist hired by Washington’s attorneys says that the disputed DNA is Tinsley’s. The experts say that the state’s different conclusion seems to have been based on an erroneous reading of a contaminated slide.

The state’s refusal to acknowledge error or allow for a third- party review in the case raises troubling questions about procedures at a lab that has previously enjoyed a stellar national reputation.

Two additional Virginia cases join with a growing body of evidence from Texas and other states to suggest that the nation needs a better system of monitoring its DNA labs.

“Our internal policing is proving to be inadequate,” argued Betty Layne DesPortes, a Richmond attorney who chairs the jurisprudence section of the American Academy of Forensic Sciences.

The case of Reginald Howard, a Richmond man accused of raping his girlfriend’s teenage daughter, illustrates the power afforded state forensic scientists in interpreting DNA results.

In a Feb. 6, 2003 analysis, a state forensic scientist wrote that the DNA profile obtained from a stain on the victim’s underpants was 2 billion times more likely to have originated from a mixture of the victim’s and Howard’s DNA than from the victim and a random Caucasian.

Only in a parenthetical phrase, easily overlooked, did the scientist note that the stunning statistics were arrived at by excluding the results at 8 of 16 points of comparison between Howard’s DNA and the stain. That’s because Howard didn’t fit the profile at the eight discarded points.

The report conveyed a certainty about the match that was not justified by the data. Nor was it mentioned that the analysis failed to find the “Y” chromosome indicating the presence of a male sperm cell, which would seem to be necessary for Howard to be guilty beyond a shadow of a doubt.

Fortunately for Howard, his attorney was sufficiently suspicious to get the report into the hands of a DNA expert, University of California-Irvine criminologist William C. Thompson. His testimony resulted in appointment of an expert witness and ultimately in Howard’s exoneration.

Also worth contemplating are the DNA results from the state lab in the case of Karl Michael Roush, who was erroneously charged with the 1996 abduction and slaying of Spotsylvania County teenager Sofia Silva.

Roush was freed after FBI investigators noticed similarities in the killings of Silva and the later slayings of young sisters Kristin and Kati Lisk. Roush could not have committed these murders because he was in jail at the time. Investigating what went wrong, the FBI laboratory discovered a mistake in the state’s forensic analysis of key fiber evidence.

The discovery resulted in the resignation of the scientist who performed the work.

Less noted at the time was the fact that a DNA test had also linked Roush to the Silva murder, even though it was later proved he was innocent of the crime.

In light of Roush’s exoneration, that report linking him to Silva’s murder deserved independent review to ensure the integrity of the state’s scientific conclusions. There’s no public record this occurred.

That sort of public accountability doesn’t happen in the current world of DNA laboratory oversight. Only if a lab concludes that it made a mistake and sets up its own corrective action plan do outside forces, such as the FBI or accrediting agencies, typically take a look. Even then, the results of the review are usually secret.

As law enforcement puts more and more emphasis on DNA testing, the nation’s DNA labs need a higher standard of oversight. Forensic scientists must not interpret results to fit the prosecution’s theory of a crime. And prosecutors, defense attorneys and judges need to stop accepting DNA test results as infallible. Science may be certain, but human interpretation is not.

Tomorrow: Suggestions for reform.