DNA Mixture Analysis: Deconvolution & Likelihood Ratio Calculator

simulator advanced ~10 min
Loading simulation...
LR_mix = 4.2 × 10⁶ — strong support for inclusion

A 2-person mixture at 3:1 ratio with 3000 RFU total yields a major contributor at ~2250 RFU and minor at ~750 RFU, both above the stochastic threshold, supporting reliable deconvolution.

Formula

Major_RFU = Total_RFU × R / (R + 1) where R is the mixture ratio
LR_mix = P(Evidence | H_p) / P(Evidence | H_d) across all loci
P(dropout) = exp(-k × RFU) where k is lab-specific logistic regression parameter

The Mixture Challenge

When DNA from multiple individuals is combined in a single sample, the resulting STR profile becomes a superposition of contributor genotypes. At each locus, allele peaks overlap — two contributors can produce 2 to 4 distinct peaks per locus, three contributors up to 6, and so on. The challenge lies in determining how many people contributed, what their individual genotypes are, and how strongly the evidence supports or excludes a person of interest.

Peak Heights and Contributor Ratios

Peak heights in Relative Fluorescence Units (RFU) are roughly proportional to the amount of DNA template. In a two-person mixture, if one person contributed three times as much DNA, their peaks will be approximately three times taller. By examining peak height patterns across loci, analysts can estimate the mixture ratio and assign alleles to major and minor contributors. When the ratio exceeds 10:1, the minor contributor's alleles may be indistinguishable from analytical noise.

Stochastic Effects and Dropout

At low template amounts, PCR amplification becomes stochastic — one allele of a heterozygote may fail to amplify sufficiently, a phenomenon called allele dropout. The stochastic threshold defines the peak height below which dropout cannot be excluded. For minor contributors in heavily skewed mixtures, dropout rates can exceed 30%, making manual interpretation unreliable and necessitating probabilistic approaches.

Probabilistic Genotyping

Modern forensic laboratories increasingly use probabilistic genotyping software that considers all possible genotype combinations, weights them by population allele frequencies, and models peak height variance and stochastic effects. These systems output likelihood ratios comparing the prosecution hypothesis (person of interest is a contributor) against the defense hypothesis (an unknown random person contributed instead), providing quantitative weight to mixture evidence.

FAQ

What is a DNA mixture in forensic genetics?

A DNA mixture occurs when biological material from two or more individuals is co-extracted and amplified. The resulting electropherogram shows overlapping peaks at STR loci, making genotype assignment more complex than single-source profiles. Mixtures are common in sexual assault, touch DNA, and multi-person crime scenes.

How is the mixture ratio determined?

The mixture ratio is estimated by comparing average peak heights of major versus minor contributors across multiple loci. At single-source loci (where contributors share no alleles), peak height ratios directly reflect the DNA quantity ratio. Ratios below 1:1 to about 1:20 can sometimes be resolved.

What is the stochastic threshold?

The stochastic threshold is the peak height (in RFU) below which allele dropout becomes a significant concern. Peaks below this threshold may represent only one of two alleles at a heterozygous locus. Most laboratories set this between 100-200 RFU based on internal validation studies.

What is probabilistic genotyping?

Probabilistic genotyping uses statistical models (continuous or semi-continuous) to evaluate the probability of observing mixture data under competing hypotheses. Software like STRmix and TrueAllele use MCMC methods to consider all possible genotype combinations weighted by allele frequencies and peak height variance.

Sources

Embed

<iframe src="https://homo-deus.com/lab/forensic-genetics/mixture-analysis/embed" width="100%" height="400" frameborder="0"></iframe>
View source on GitHub