The alleged falsifying of DNA testing data by a Montgomery County-based laboratory employee has exposed a vulnerability in computerized forensic analysis previously unknown to some experts, raising questions about the reliability of the most widely used and trusted method of testing.
The Orchid Cellmark employee electronically manipulated the analysis in 20 tests, the company says. Though she did not alter the outcome of the tests, she overrode procedures designed to ensure the accuracy of the tests by substituting data in the known specimen, or control samples, according to Cellmark.
"I have not heard of anything like this before," said Lawrence Kobilinsky, an associate provost at John Jay College of Criminal Justice in New York.
Analyzing DNA through computer software emerged in the mid-1990s and has become the most trusted and reliable method of forensic analysis, Kobilinsky said: "I would find it surprising if anyone has ever done this or if anything like this has been discovered before."
Cellmark - the world's largest private DNA testing firm - fired Sarah Blair in September for "professional misconduct" involving the cases, performed for the FBI and the Los Angeles Police Department.
Blair, who was based at Cellmark's Germantown lab, has denied the allegations.
The incident has prompted the public defenders in Los Angeles County to examine all pending cases in which Cellmark analyzed evidence.
"What had previously been represented was that you could not electronically manipulate the raw data. ... This wasn't just our belief, it was the belief of the relevant DNA experts in the field," said Patrick Kent, chief of the forensics division for the Maryland public defender's office, also reviewing cases in which Cellmark was used. "But that clearly is not the case."
Kent said Blair was using GeneScan software, produced by Applied Biosystems of Foster City, Calif. A spokeswoman for that company, Lori Murray, said it is looking into the matter.
Dr. Paul Kelly, the CEO of Cellmark's Princeton, N.J.-based parent company, Orchid BioSciences Inc., said yesterday that the company quickly identified the errors and its customers are pleased with its actions.
"Whatever program [Blair] was using, these sorts of events can happen in any testing situation if someone really wants to get involved in such a level of professional misconduct," said Kelly, who said Blair worked for the company about a year. "The important thing is we have the process in place to identify that."
The Los Angeles police cases involving Blair have not gone to trial, and the Los Angeles County district attorney's office anticipates that there should be no damage to them.
It is unclear where the FBI cases originated. Special Agent Ann Todd, a spokeswoman for the FBI lab at Quantico, Va., declined to comment on that.
In a typical forensic analysis, a DNA sample is placed into a vial and a robotic needle sucks the substance into the testing machine, where the computer software analyzes the material, Kobilinsky said.
After an analyst matches a DNA sample to evidence from a crime scene, tests with a known specimen, or "control" sample, are performed to ensure that the experiment worked properly and is free from contamination.
A positive control test analyzing a specimen known to contain DNA should produce a result showing DNA, or deoxyribonucleic acid, Kobilinsky said. A negative control test that examines a substance not containing DNA should show no signs of the genetic material, he said.
In a September letter to the Los Angeles Police Department crime laboratory Robin W. Cotton, a director of technical forensic science for Cellmark, wrote that when a control test showed an "imperfect" profile and should have been reanalyzed, Blair instead substituted a control profile with no imperfections.
The letter stated that all of the affected cases have been re-tested, and the results of the original analysis remained the same.
Presumption of ethics
However, the incident shows that the objective method of computerized testing can be compromised by the people conducting the analysis, Kobilinsky said.
"Underlying everything is a presumption of ethics," he said. "Electronic tampering of the results - it's just unheard of."
As a result of the incident, Lutherville defense attorney Joseph Murtha said he will challenge the lab's credibility in future cases. He has a pending case in which Cellmark will perform the DNA analysis, and if the results are connected to his client, he will call for an independent review of the analysis.
"It's incredibly disappointing, because Cellmark had a very good reputation," Murtha said. Blair's firing "has called into question whether or not you can rely upon the test results."
Louis P. Willemin, deputy public defender for Maryland's Howard and Carroll counties, said an incident of alleged lab misconduct is "the defense's worst nightmare because it's not the easiest thing to find out."