Detecting STR peaks in degraded DNA samples

Human identification from DNA is typically based on 13 short-tandem repeat (STR) alleles. Commercial kits used in forensic casework rely on the detection of these alleles in DNA samples acquired from an individual. However, the process itself is slow (it can take up to 2 days when conducting a laboratory analysis or 1 hour when using Rapid DNA systems) and has been designed to operate on pristine DNA samples. The need for achieving fast and accurate DNA processing has spurred efforts in developing portable systems that can reduce the processing time to less than 1 hour. But such systems are expected to operate on degraded DNA samples due to the architecture and process used by the instrument. Consequently, detecting the alleles in such degraded DNA samples can be a challenging problem. In this paper, we present an algorithm to detected allelic peaks from degraded DNA signals based on an adaptive signalprocessing scheme. The performance of the algorithm is evaluated on two datasets: 1) data collected at the WVU Department of Forensic and Investigative Sciences, obtained by performing a controlled DNA degradation using ultraviolet radiation, 2) data provided by NIST obtained by varying cycle counts for the PCR processing step. Experiments indicate the efficacy of the algorithm in allelic peak detection and reiterate the need for approaching the problem in a systematic manner.