8-Oct-2018

New foundational paper on forensic error and validation

Comparing forensic evidence with a suspect gives the match strength. But what are those match numbers for everyone else? Depending on the DNA, strong evidence may be common, or weak evidence rare. How is match strength distributed?

This problem, and others, is solved in this week’s Heliyon paper, “Efficient construction of match strength distributions for uncertain multi-locus genotypes.”

Author Dr. Mark Perlin of Cybergenetics uses exact genotype probability, and rapidly combines DNA loci. A computer can map out all trillion-trillion match statistics – accounting for every possible person – in a fraction of a second.

The “convolution” method enables rapid and accurate determination of error rate. With a defendant match, scientists can report the chance of evidence matching as strongly to an innocent person. For a validation study on many DNA samples, error rates are immediately found. The underlying science is the same.

The published solution has applications “in the forensic, natural and social sciences,” wrote Dr. Perlin. “Computer-derived match strength distributions elicit the information inherent in DNA evidence, often overlooked by human analysis.”

Sequential convolution of probability functions.