Autosomal DNAm profiles from two-person DNA mixtures can be successfully deconvoluted when one contributor is known, enabling age prediction accuracy comparable to single-source samples except at low suspect-to-victim ratios (1:4 and 1:10).
Key Findings
Results
The UnMixMe approach successfully deconvoluted DNAm profiles from two-person DNA mixtures to predict suspect age with accuracy comparable to single-source DNA samples at most mixture ratios.
Age prediction accuracy was comparable to single-source suspect DNA for mixture ratios of 10:1, 4:1, 2:1, 1:1, and 1:2
Accuracy degraded at low suspect-to-victim ratios of 1:4 and 1:10
Mock trace DNA mixtures were prepared from blood samples using two male-female pairs
Seven suspect-to-victim ratios were tested: 10:1, 4:1, 2:1, 1:1, 1:2, 1:4, and 1:10
DNAm levels (β-values) were measured using the Illumina EPIC v2.0 microarray
Methods
Four array-based epigenetic clocks were used for age prediction from deconvoluted DNAm profiles.
The four clocks tested were: BLUP, EN, Horvath, and skinHorvath clock
All clocks were applied to deconvoluted DNAm profiles retrieved from the DNA mixture
The clocks were also applied to single-source suspect DNA samples as a reference comparison
Results
Three main factors were identified as affecting age prediction accuracy in deconvoluted DNA mixture profiles.
The three factors identified were: precision of the DNAm technology, accuracy of DNA mixture ratio estimation, and accuracy of the prediction model used
The age difference between DNA mixture contributors did not influence prediction accuracy
DNA mixture ratio was estimated via the traditional STR profile for human identification
Methods
The UnMixMe method requires knowledge of the DNAm profile of one contributor (the victim) and the DNA mixture ratio to deconvolute the suspect's profile.
The approach relies on knowing the DNAm profile of the other contributor (the victim)
It also requires knowledge of the DNA mixture ratio between the two contributors
The mixture ratio was estimated using traditional STR profiling
This constitutes a proof-of-concept study establishing feasibility of the approach
Background
Current autosomal age-predictive DNAm approaches are inapplicable to multi-donor DNA mixtures, which are frequently encountered in forensic casework.
DNA methylation profiling has been established as a reliable method for estimating chronological age of an unknown sample donor
Multi-donor DNA samples (DNA mixtures) represent a major limitation for current approaches using autosomal age-predictive CpGs
DNA mixtures are described as frequently encountered in forensic casework
Poggiali B, Børsting C, Kampmann M, Wiuf M, Kelager A, Jepsen A, et al.. (2026). Age prediction of contributors to two-person DNA mixtures via deconvolution of autosomal DNA methylation profiles.. Forensic science international. Genetics. https://doi.org/10.1016/j.fsigen.2026.103438