| Popis: |
In forensic DNA applications, characterizing DNA signal to noise resolution is needed to establish effective and reasonable analytical thresholds (AT). Studies have shown the significance of understanding the behavior of baseline noise such that it can be effectively modeled or can be used to compute an analytical threshold that minimizes Type I and II detection error. Previous studies on noise have described electropherogram noise as well-described by normal, log-normal and Gamma distributions, but there still exist differences of opinion on which distribution class to use. PROVEDIt single source and mixture samples amplified using the Powerplex® Fusion 6C kit were used for noise characterization. First, we determine whether a normal or log-normal distribution class best fits the noise data. We also ascertain whether noise distributions are significantly different between colors and between loci. Our findings demonstrate that the log-normal distribution fits the noise data better than the commonly employed normal-class. In addition, noise peak heights were dependent on both dye and locus. Lastly, noise peaks showed an increase in peak height with increase in injection time, suggesting there may be two sources of noise: that originating from instrumentation and that from amplification. |
