Manitoba

Top DNA expert has 'no confidence' in DNA results linking accused killer to Candace Derksen's murder

A top American DNA expert found so many problems in the lab work which linked accused killer Mark Grant to Candace Derksen's 1984 murder he testified he considered the results "fatally flawed."

Says Molecular World report should never have been published

Mark Edward Grant was convicted in 2011 of second-degree murder in connection with the 1984 death of Candace Derksen, 13. In 2013, the Manitoba Court of Appeal ordered a new trial, ruling the trial judge was wrong to exclude evidence that the defence argued suggests Derksen might have been killed by someone else. (CBC)

A top American DNA expert found so many problems in the lab work that linked accused killer Mark Grant to Candace Derksen's 1984 murder he testified he considered the results "fatally flawed."

"I have absolutely no confidence in the results of this testing or the relevance to this case," Frederick Bieber told court Monday. 

Bieber has an extensive resume when it comes to forensic identification. The medical geneticist and Harvard University professor sat on the FBI advisory board to oversee the quality of DNA testing in the United States. He was also one of a handful of DNA experts appointed to identify the remains of victims from the 9/11 terrorist attacks in New York and hurricane Katrina in Louisiana, court heard.

He was called by Grant's defence lawyer to review the DNA results from Molecular World, a Thunder Bay lab which ran tests on key evidence from the Derksen cold case, including the twine Candace was bound with. 

The 13-year-old's body was found tied up and frozen on the floor of an Elmwood storage shed in 1985, seven weeks after she went missing. 

DNA on the twine linked Grant to the teen's murder in 2007. He is now being tried in her death for the second time, after his first conviction was overturned in 2013.

Data that would have excluded Grant left out

After reviewing the report from Molecular World, Bieber told court "this was among the most egregious departures from proper lab practices" he thinks he has ever seen.  

Bieber said the report prompted concern about lab methods and the way DNA results were interpreted. 

He said scientists ignored three DNA markers which would have excluded Grant as a match on the twine samples. A judgement call that he said could not be explained.

"It's a textbook example of trying to make data fit the hypothesis," he said, suggesting there was evidence of suspect bias.

Also troubling, he said, was that lab staff made three attempts to obtain a DNA result from the twine and the results all came back different.

Bieber said one of the foundations to ensure quality in DNA testing is that results are reproducible.

"Certainly the results are not reproducible, which means they're not reliable," he told court.

Results should never have been published

Bieber also said lab staff failed to obtain a clean DNA profile of Grant from the blood sample submitted by police and therefore should not have been used for comparison. 

"That raises further concern about the practice patterns in the laboratory," he said. 

Defence lawyer Saul Simmonds asked Bieber if he "found anything right" within the lab file.

Bieber said the notes and previous lab work provided to Molecular World "very thorough and very complete."

But the report from Molecular World sent back, he told court, was "unfounded, inaccurate and inappropriate" and should never have been published. 

This is the second expert called on by the defence to raise serious concerns about the DNA results.

Crown Attorneys Michael Himmelman and Brent Davidson will begin cross-examination tomorrow.

The judge-only trial is scheduled to run until March 3 and is being presided over by Justice Karen Simonsen.