Quantifying Allelic Dropout in Degraded and Mixed Forensic STR Profiles: An In-Silico Ground Truth Study

Abstract

Allelic dropout in degraded and mixed short tandem repeat (STR) profiles remains one of the most persistent obstacles to reliable forensic DNA interpretation, yet most of the empirical evidence describing it comes from casework or dilution-series data in which the true underlying genotype can only be inferred rather than directly verified. This study used a fully reproducible, controlled in-silico simulation, in which every contributor's genotype and every mixture's composition was fixed and known exactly, to quantify how degradation, contributor imbalance, and the number of contributors affect allelic dropout and the reliability of profile interpretation. A synthetic ten-locus panel covering a range of 120-380 base pairs was simulated under four degradation levels, four two-person mixture ratios (1:1 to 1:10), and one to three contributors, resulting in 8,800 profiles and 261,824 expected-allele observations. Allelic dropout rose from 2.29% under no degradation to 40.61% under high degradation, with an almost eight-fold gradient between the smallest and largest loci at the highest degradation level. Minor-contributor allele recovery fell from 65.96% at a balanced 1:1 ratio to 12.53% at 1:10, and a standard maximum-allele-count estimator of the number of contributors, while accurate for single-source profiles (>99%), succeeded in only 3.50% of severely degraded three-person mixtures. In a multivariable logistic regression, degradation and mixture imbalance showed a significant interaction (likelihood-ratio χ2(1) = 2,279.46, p < 0.001). A strict ground-truth reconstruction analysis further showed that even a balanced two-person mixture yielded a fully recoverable profile in fewer than half of the non-degraded replicates.

 The results, which were obtained from a known ground truth, support decades of casework-based forensic genetics research and provide a quantitative, mechanistic understanding of how degradation and mixture complexity interact, further justifying the use of probabilistic, rather than threshold-based, DNA interpretation methods.