Discipline Article DNA and Serology Analytical Methodology Legal & Evidentiary Standards Level: Advanced 🔒 Login Required (Pro)

Probabilistic Genotyping of Complex Low-Template STR DNA Mixtures: Likelihood Ratios, Stutter Modeling & Drop-Out

Mathematical formulation of continuous peak-height models vs. binary inclusion statistics under SWGDAM guidelines

Dr. Aris Thorne, Ph.D. — DNA & Serology Study Guide

Published: October 19, 2026 • Updated: Oct 21, 2026 • 394 words (2 min read)

Overview & Learning Objectives

Multi-contributor, low-template forensic DNA mixtures frequently exhibit stochastic amplification artifacts including allelic drop-out, locus drop-in, and backward/forward polymerase slippage stutter. This Pro Member article derives the Bayesian Likelihood Ratio (LR) framework used in continuous probabilistic genotyping software and demonstrates how peak height gamma distributions weight genotypic combinations.

Key Takeaways:

Formulate prosecutor ($H_p$) and defense ($H_d$) propositions for 2- and 3-person touch DNA mixtures.
Model electropherogram RFU peak heights and backward stutter ratios ($n-4$ bp) using continuous gamma distributions.
Apply SWGDAM and ISFG verbal equivalence scales and Daubert admissibility safeguards.

Table of Contents

1. The Likelihood Ratio ($LR$) Framework in Forensic Genetics

In evaluating an evidentiary short tandem repeat (STR) electropherogram profile $E$ given reference genotype $S$ of a person of interest (POI), forensic geneticists evaluate the ratio of two conditional probabilities under mutually exclusive propositions—the prosecution proposition ($H_p$) and the defense proposition ($H_d$):

$$LR = \frac{\Pr(E \mid S, H_p)}{\Pr(E \mid S, H_d)}$$

For a two-person mixture where victim $V$ is a known contributor, the propositions are typically formulated as:

$H_p$: The evidentiary DNA profile $E$ originated from the victim $V$ and the person of interest $S$.
$H_d$: The evidentiary DNA profile $E$ originated from the victim $V$ and an unknown, unrelated individual $U$ drawn at random from the relevant subpopulation (incorporating Balding-Nichols coancestry coefficient $\theta = 0.01$).

🔒

Scientia Forensis • Pro Member Access

Login to continue reading this article

This full forensic study resource (394 words, complete concept breakdowns, worked examples, and reference notes) is available to Pro Members. Log in to unlock the full text and interactive Table of Contents.

2 min Full Study Time

Advanced Difficulty Level

Full Notes & Code Complete Reference