Flexible reading of single-primer extension microhaplotypes improves degraded DNA mixture interpretation

Microhaplotypes (MHs) are informative forensic markers for degraded and mixed DNA because they combine short target regions, haplotypic diversity and stutter-free allele detection. However, degraded templates often generate reads that cover only part of an MH target. Discarding these partial reads may reduce the information available for mixture interpretation. We evaluated a single-primer extension-based MH massively parallel sequencing system (SPE-MH-MPS) for low-template DNA, simulated degraded DNA, two-person mixtures and degraded two-person mixtures. A 91-locus autosomal panel developed from Chinese Southern Han data was analyzed using three quantitative inputs: complete-read counts, flexible-read counts and flexible UMI counts. Flexible reading increased allele recovery in low-template and degraded samples. At a DNA input of 31.25 pg, the mean allele detection rate increased from 1.5% with complete reads to 20.0% with flexible reads. In simulated degraded samples, complete reads detected fewer than 3% of alleles, whereas flexible reads recovered 50.6-72.0%, 44.5-47.8% and 21.4-36.3% of alleles after 60, 120 and 180 min of degradation, respectively. In non-degraded two-person mixtures, flexible-read counts generated higher log10 likelihood ratios than complete-read counts (Wilcoxon p < 0.001). Flexible-read counts also showed smaller deviations from the theoretical minor-contributor proportions and higher genotype deconvolution accuracy. Similar trends were observed in degraded mixtures, although minor-contributor-specific allele recovery decreased when degradation and mixture imbalance occurred together. UMI purity filtering reduced reference-discordant reads, but UMI-based counts did not consistently outperform flexible-read counts. These results indicate that constrained flexible reading can recover informative partial MH signals and improve the interpretation of defined degraded two-person DNA mixtures.