Author ORCID Identifier

https://orcid.org/0000-0003-3224-1223

Defense Date

2022

Document Type

Directed Research Project

First Advisor

Dr. Tracey Dawson Green

Second Advisor

Dr. Sarah J Seashols-Williams

Third Advisor

Andrea Williams

Abstract

At present, the forensic DNA workflow is not capable of providing information about the contributor status (single source vs. mixtures) of evidentiary samples prior to end-point analysis. This shortcoming can exacerbate the challenges inherent to mixtures and low-template DNA samples. Provided additional sample information earlier in the workflow, protocols could be implemented to mitigate these challenges. High-resolution melt (HRM) curve analysis is a technique used to detect genetic variation in DNA fragments and in the last decade has been evaluated for use in differentiation of samples by genotype and/or contributor status. To this end, a proof-of-concept HRM assay using the STR loci D5S818 and D18S51 and EvaGreen® intercalating dye was integrated into the Qiagen Investigator Quantiplex® qPCR kit. When tested on the ABI QuantStudio 6 Flex qPCR platform, resulting melt curve datasets and statistical analyses were capable of distinguishing single-source samples from mixtures at an 87.88% rate. Given this initial success, integration of the HRM assay into a more commonly used chemistry, the Quantifiler Trio kit, was pursued. Unfortunately, the presence of EvaGreen® dye caused substantial increase in small autosomal and Y-target quantification values, rendering this data unreliable. SYTO 17 and SYTO 64 fluorescent intercalating dyes with spectral emissions in the IPC target channel were tested in the assay to minimize spectral overlap with quantification target dye channels. However, dye channel sharing of the SYTO dyes with the IPC target dye, JUN, resulted in a change in the expected cycle threshold values. Additionally, testing with SYTO 17 revealed inflation of the passive reference dye Mustang Purple and was not pursued further.

Integration of the HRM assay into the Quantifiler Trio was tested with two different reaction condition chemistries. After optimization of reaction condition, additional testing was conducted with adjustments in the assays data analysis settings. The standard QuantifilerTrioassays inter-run variation was compared to the percent differences of quantification values obtained from the standard Quantifiler Trio and integrated Quantifiler Trio HRM assays. Overall, percent differences between the assays were comparable to one another. Further, DNA profiles generated with the standard Quantifiler Trio and integrated Quantifiler Trio HRM assays were evaluated and revealed fully concordant profiles that were not statistically different from one another. Moreover, both assays reported similar male-to-female ratios and degradation indices. Lastly, an inhibition study conducted to discern whether dye channel sharing with the IPC target dye influenced its ability to detect inhibition was conducted and revealed no drastic difference in inhibition detection between the assays.

For the classification of single source vs. mixture samples in the optimized Quantifiler TrioHRM assay, the best performing prediction model was used for D5S818 (SVM linear) and D18S51 (SVM radial) loci. D5S818 reported a single source prediction accuracy of 86.76% and a mixture prediction accuracy of 25%. D18S51 reported a single source prediction accuracy of 79.41% and mixture prediction accuracy of 62.5%. The overall prediction accuracy of the assay using a combined metric was 73.8%.

Rights

© The Author(s)

Is Part Of

VCU Master of Science in Forensic Science Directed Research Projects

Date of Submission

5-7-2022

Available for download on Sunday, May 07, 2023

Share

COinS