Testimony in Wade murder trial centers on DNA
While an Allegheny County crime lab biologist would only say Allen Wade couldn’t be ruled out as a suspect in the slayings of sisters Susan and Sarah Wolfe, a computer program provided more certainty — that there was just a 1 in 6 trillion chance someone else contributed DNA to key evidence.
In the ninth day of Wade’s homicide trial Thursday, the defense argued that the statistics were just as susceptible to human bias as traditional DNA interpretation, and that the computer program lacks transparency.
Dr. Mark Perlin, CEO of Oakland-based Cybergenetics, said his company’s TrueAllele software took DNA data from the county Medical Examiner’s crime lab and calculated the likelihood of all the different combinations in the mix of DNA found under Susan Wolfe’s fingernails following her Feb. 7, 2014 murder.
The Wolfes were found shot to death in the basement of their East Liberty home; surveillance footage and DNA evidence led police to charge Wade.
When compared to Wade’s DNA profile, the TrueAllele program determined Wade was 6.06 trillion times more likely than another, unrelated black man chosen at random to have been a contributor to the sample, Perlin said. In other words, the chance of some other black man being the source of that DNA was 1 in 6.06 trillion.
In testimony this week, crime lab biologist Walter Lorenz could say only that Wade could not be excluded as a potential contributor to the DNA under Susan Wolfe’s fingernails, using traditional methods of interpreting DNA evidence.
Wade’s defense fought and lost before the trial for access to the TrueAllele source code for a greater understanding of how the program works.
“It’s a trade secret, like most source code is for most companies,” Perlin said. “We make the executable program available so it can be tested like other software.”
Defense attorney Aaron Sontz questioned whether the program or its reports were truly objective, given that Cybergenetics used one set of DNA data to make its report, but didn’t report on a second set of data it analyzed when the crime lab was trying to get a better look at the DNA by using enzymes to duplicate the DNA for a larger sample.
TrueAllele’s analysis of the second data set showed a slightly lower likelihood that Wade was a contributor to the evidence, Sontz said.
“Isn’t it true you cherry-picked your data to be more consistent with Allen Wade?” he said.
Perlin said the second data set was larger, which created more variations and lower-quality data, which is why his report stuck to the first set. He said the likelihood Wade was in the samples was still pretty high.
Allegheny County Common Pleas Judge Edward Borkowski recessed the trial before Sontz finished his cross-examination. Perlin had a scheduling conflict and will resume his testimony Monday; other prosecution witnesses are scheduled for Friday.
Matthew Santoni is a Tribune-Review staff writer. Reach him at 412 391 0927 or [email protected].