EPT+ Show Spoiler +
Forensic DNA typing relies on software programs to process and display data obtained from capillary electrophoresis of DNA fragments. Common functions performed by these software programs include baselining, color separation, and smoothing, which precede analytical functions such as peak detection, sizing, and allele calling. The performance of many of these functions can be altered by changing user-defined parameters, which are determined by a laboratory’s validation studies. For example, a commonly used analytical peak detection threshold for allele calls is 100RFU (relative fluorescence units). The analyst confirms allele calls resulting from the involved computations performed by the software. In the case of the example, a peak height below the detection threshold is deemed statistically undependable, and the allele information it represents would not be reported in the DNA profile. Therefore, peak height is an important factor when interpreting DNA profiles. Because each program has its own algorithms, the peak height can vary based on differences in raw data processing.
This study investigated the disparities in peak heights produced by four software programs, Genemapper® ID-X (Applied Biosystems, Foster City, CA), GeneMarker® HID (Soft Genetics, State College, PA), FSS-i3™ (Promega Corporation, Madison, WI), and TrueAllele® Casework (CyberGenetics, Pittsburgh, PA), operating on the same raw data. We found correlated differences in peak heights between these programs. Variations in the reportable profile from a single sample exhibiting alleles near the peak detection threshold were observed. We conducted a simulation to test the efficacy of adjustments based on regression analysis of our peak height data. Using these adjustments, we were able to significantly reduce discrepancies in profiles exhibiting near-threshold peaks.
