Defense Date
2025
Document Type
Directed Research Project
First Advisor
Sarah Seashols-Williams
Second Advisor
Tracey Dawson-Green
Third Advisor
Paul Brooks
Fourth Advisor
Xiaomeng Yang
Abstract
Forensic DNA Phenotyping (FDP) traditionally predicts stable, externally visible characteristics (EVCs) such as eye, hair, and skin color, but estimating dynamic traits, such as body mass index (BMI) and weight, remains challenging and unexplored. MicroRNAs (miRNAs), which are highly stable in degraded forensic samples and regulate adipogenesis and metabolic pathways, offer a promising alternative biomarker class. This study evaluated the performance of identified miRNAs miR-16-5p, miR-126-3p, miR-486-3p, miR-15b-5p, miR-142-5p, and miR-145-5p- using RTqPCR analysis of DNA extracts (QIAamp DNA Investigator Kit) from dried blood collected from 100 donors with diverse BMI and weight ranges. Reference markers miR-16-5p and miR-126-3p showed stable expression, with miR-126-3p exhibiting superior performance. Most ΔCq values for target miRNAs showed weak downregulation with increasing BMI and weight, though marker behavior varied across gender subgroups. Several of these associations reached statistical significance, including let-7i-5p for overall BMI (p = 0.027) and weight (p = 0.047), as well as miR-145-5p across multiple stratifications (p < 0.05). Normalization to miR-126-3p improved the linear fitness in approximately half of the evaluated markers, particularly let-7i-5p and miR-106a-5p, with statistically significant improvements observed for let-7i-5p in female BMI (p = 0.0085) and weight (p = 0.00092), and for miR-106a-5p in gender-combined analyses (p < 0.05). Across all analyses, miR-145-5p demonstrated the strongest and most consistent associations, explaining up to 12.6% of BMI variance and 10.5% of weight variance when normalized to miR-126-3p, supported by highly significant p-values for both BMI (p = 0.00029) and weight (p = 0.0010). These findings suggest that weight-responsive miRNAs, particularly miR-145-5p, may be incorporated into future forensic phenotyping prediction tools. Additional statistical assessments and validation using expanded marker sets and larger population cohorts will be essential for improving predictive relevance.
Rights
© The Author(s)
Is Part Of
VCU Master of Science in Forensic Science Directed Research Projects
Date of Submission
12-21-2025