A Tool for Determining the Number of Contributors: Interpreting Complex, Compromised Low-Template DNA Samples
Abstract
In forensic DNA analysis, the interpretation of a sample acquired from the environment may be dependent upon the assumption on the number of individuals from which the evidence arose. Degraded and damaged samples often exhibit signal that is lower than expected at the high molecular weights, resulting in a ski slope effect. In this work we developed a method, and corresponding software tool, that provides a mechanism to determine the number of contributors (NOC) to an unknown stain, regardless of condition. This was accomplished by assessing all known confounding factors related to complex, low-template DNA mixture interpretation. The software, named NOCIt, considers signal interferences from reverse stutter products, baseline noise and allele drop-out. It also models the expected allele peak heights or areas at all molecular weights and utilizes this information to provide an APP (a posteriori probability) that a certain NOC gave rise to the evidence. We extended the capabilities of NOCIt to provide laboratory and algorithmic solutions to the issues associated the interpretation of compromised, complex and low-template DNA samples. We test the new implementation of NOCIt on a set of 221 samples containing 15 forensically relevant STR (short tandem repeat) regions of interest and 248 samples containing 21 forensically relevant STR loci. We validate NOCIt by examining: 1) accuracy; 2) repeatability; and 3) effects of low-level DNA quantities on the ability to infer the number of contributors to the evidence. We show that the accuracy of NOCIt is roughly 90 percent across all samples tested.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 28, 2017
- Accession Number
- AD1051279
Entities
People
- Catherine M. Grgicak
- Desmond S. Lun