DNA expert Professor Angela Van Daal leaves court. Picture: Campbell Brodie Source: The Advertiser

THE case against a man accused of rape is built on inflated DNA evidence and flawed statistics, a court has heard.

Prosecutors claim Peter Tasman Cannell raped an elderly woman in her home and escaped detection for 17 years. A DNA sample, they claim, proves there is a 600-billion-to-1 chance of anyone other than Cannell being the rapist.

Yesterday, Bond University forensics expert Associate Professor Angela van Daal told the District Court that statistic was unreliable.

Giving evidence on Cannell's behalf, she said Forensic Science SA had not revealed its "confidence interval" - an expression, as a percentage, of faith in its conclusion. She said the odds of Cannell being the rapist could drop between 20 and 40 per cent once that figure was included in calculations.

"The confidence interval takes into account the number of people in the database ... one with 100 or 200 people isn't going to give the same confidence as one with 1000 or one million people," Prof van Daal said.

"That is a very fundamental concept of statistics, something that any student would have done ... it's very elementary.

"Any estimate can, obviously, be incorrect ... given you do not know the true answer, it's important to determine the confidence had in the estimate."

Cannell, 41, of Victoria, has pleaded not guilty to one count of rape and one count of burglary.

Prosecutors have alleged he raped a woman, 81, in her Wright St home in October 1993. The victim suffered extensive injuries requiring six weeks in hospital.

Original post:
DNA rape case evidence 'flawed'