Scientist remains unsatisfied with re-testing of evidence

BY FAYE FLAM

PHILADELPHIA - The scientist who did the initial DNA tests for the Roger Keith Coleman case said he always expected new tests would back him up - but he's still not satisified.

Edward Blake, who runs Forensic Science Associates in Richmond, Calif., is still raising questions about the investigation. For him, the re-testing of DNA evidence using the latest technology hardly nails shut the 25 year-old murder case. "I am concerned as a scientist," he said, calling the retesting, "a cynical exercise in manipulating a scientific investigation."

Blake's original test concluded that Coleman's DNA and sperm taken from victim Wanda McCoy's body shared a genetic type present in one in 50 people, or 2 percent of the population. The odds were slim that a more accurate test would show the wrong man was sent to his death.

The technique he used in 1990 was state of the art, he said. Back then, DNA testing for forensics was still relatively new. Just four years earlier, biologist Alec Jeffreys used DNA for the first time to catch a serial killer-rapist in a small English town.

Blake said he was the first forensic scientist to employ a procedure called polymerase chain reaction, or PCR, which is now common practice for making many copies of a strand of DNA. PCR allows scientists to extract good information from a tiny sample of blood, semen, saliva, skin or other tissues.

Early PCR testing was an improvement over what was done before - checking whether blood was type A, B, AB or O. But it was a far cry from anything that could be called fingerprinting. Blake carried out what was to be the standard test for several years - analysis of a gene called DQ alpha, which comes in a handful of distinct varieties.

Samples extracted from McCoy's body showed sperm from two different men, said Blake. One presumably came from her husband, he said. The other was linked to Coleman, but not conclusively.

This type of evidence was very good for exclusion, said Larry Presley, a forensic scientist from National Medical Services in Willow Grove, Pa. That is, if Coleman's DNA type did not match either type found on the victim it could have helped clear him. But the FBI, Presley said, didn't consider this type of match sufficient to demonstrate guilt.

Over the next 10 years, DNA typing improved, which is why Blake said he originally requested the Coleman samples get retested in 2000.

Current technology relies on a technique called Short Tandem Repeats, or STR. DNA carries its code in strings of four different chemical blocks, or bases, usually abbreviated as A,T,C, and G. But we are 99.9 percent identical to one another.

Forensics uses long strings of DNA called non-coding, or "junk" DNA, that don't make up part of the genetic code. These strings of DNA vary from person to person much more so than the DNA in the genes.

Within these stretches, scientists have identified places where some short bit repeats over and over, say, ATG-ATG-ATG-ATG. In some people this might repeat 12 times, in others, 10, or seven.

Presley has compared STR analysis to counting the number of Smiths in different phone directories. All towns have Smiths, he said, but some have more than others.

The power of modern STR comes from examining up to 13 different spots where such repeats occur - sort of like counting not only the Smiths but the Andersons, Joneses, Macdonalds and a slew of other names. At some point you can be pretty sure if two phone books match this way they go with the same town.

Scientists say the chance someone other than an identical twin matches your DNA on a modern STR analysis amounts to less than one in a trillion.

Before results from the STR came back Thursday, Blake said he regretted that the case had become political. The purpose of the retesting should have been to back up the historical record.

Instead, he said, the case was hijacked as part of the death penalty debate. He said he suspected the hype would backfire and hurt the cause of the anti-death penalty activists who played it up.

Blake said the DNA results do not prove that Coleman killed McCoy, and the initial investigation left open critical questions. Normally DNA samples taken from rape victims come on a swab, but Blake said he was given just the wooden stick from which he had to scrape DNA. No one ever explained what happened to the rest of the swab, he said.

Then, he learned that Coleman had blood spattered on his coveralls, but no one sent him that DNA for testing to see if it matched the victim. He was never told why, he said.

Blake asked for DNA testing of the blood samples along with a retesting of the semen but said the governor's office first told him they'd try to find them, then reported the samples were destroyed, he said. All the evidence, he said, was destroyed except for that tiny stick he kept frozen all these years.

This latest DNA testing shows Coleman had sex with the victim, said Blake, but not that he killed her. "They could have been having an affair," he said. "I haven't heard anyone ask that question and answer it."

"I've had cases like that," said Presley, "where the DNA had nothing whatever to do with the homicide." A recent one involved the murder of a part-time prostitute, he said. One of her last customers stood accused, though he was later cleared and someone else charged.

"Blake is a good scientist," said Presley. DNA testing is now capable of near perfect matching, he said, "But bottom line - there's no substitute for a good investigation."