Pennsylvania prosecutors use TrueAllele in homicide guilty plea

Back to Presentations

Profiles in Productivity: Greater Yield at Lower Cost with Computer DNA Interpretation

M.W. Perlin and B.W. Duceman, "Profiles in productivity: Greater yield at lower cost with computer DNA interpretation", Twentieth International Symposium on the Forensic Sciences of the Australian and New Zealand Forensic Science Society, Sydney, Australia, 9-Sep-2010.


PowerPoint presentation with live audio recording of the Australian and New Zealand Forensic Science Society 2010 talk.

Download Transcript
Download Handout
PPTDownload PowerPoint


DNA evidence can pose interpretation challenges when it is mixed, damaged, degraded or in low amounts. The resulting human interpretation bottleneck introduces backlogs. With an "inconclusive" result, reprocessing additional DNA items to produce a reportable match score entails further reagent cost, laboratory time and human effort.

We examined 88 DNA mixture items from reported cases. We classified these items into three groups: simple (N=35), intermediate (20) and complex (33). As item complexity increased, the fraction that yielded a reportable match score decreased from 49% (simple) to 25% (intermediate) and 21% (complex). For example, complex items would need five DNA assays and reviews in order to produce one reportable match score.

Cybergenetics TrueAllele® Casework uses a probability modeling approach with statistical computer search. TrueAllele objectively infers genotypes from DNA data, and subsequently matches them with other genotypes (e.g., suspect or database). The model includes common data artifacts, such as PCR stutter, relative amplification, degraded DNA, mixture weight, peak uncertainty and stochastic effects. Uncertainty in variables (including genotypes) is represented using probability.

TrueAllele must always produce a genotype and make comparisons -- it is programmed that way. The average log likelihood ratio (LR) match information scores were 16.3 (simple), 13.1 (intermediate) and 11.9 (complex). Thus complex items had a 100% computer information yield averaging a trillion LR, whereas the 21% human review yield entailed 5-fold redundant experimentation with lower match scores. We conclude that computer interpretation of DNA evidence can produce greater identification information for lower laboratory time, cost and effort.