Related Experiment Video
Updated: Jun 10, 2025
08:39
Simulating Temperature in a Soil Incubation Experiment
Published on: October 28, 2022
2.8K
Environmental predictors impact microbial-based postmortem interval (PMI) estimation models within human
Allison R Mason1, Hayden S McKee-Zech2, Dawnie W Steadman2
1Department of Microbiology, University of Tennessee-Knoxville, Knoxville, TN, United States of America.
Plos One
|October 11, 2024
Summary
Soil microbes can help estimate postmortem interval (PMI). This study found that while microbial data predicts PMI, environmental factors and specific markers impact accuracy, with high error rates observed.
Area of Science:
- Forensic Science
- Microbiology
- Bioinformatics
Background:
- Established postmortem interval (PMI) estimation methods have limitations.
- Microbial succession in soil offers a potential supplementary PMI estimation tool due to microbes' presence throughout decomposition.
- Previous machine learning models for PMI estimation from soil microbiomes did not incorporate environmental factors.
Purpose of the Study:
- To evaluate the impact of including environmental data on microbial-based PMI estimates from soil decomposition samples.
- To compare the predictive performance of different biological markers (16S, ITS, combined) and taxonomic levels for PMI estimation.
Main Methods:
- Random forest regression models were developed to predict PMI using relative microbial taxon abundances.
- Data included bacterial 16S, fungal ITS, and combined 16S-ITS marker data at various taxonomic levels (phylum, class, order, OTU).
- Environmental predictors (temperature, pH, conductivity, enzyme activities) were assessed for their influence on model accuracy (MAE).
Main Results:
- Mean Absolute Error (MAE) ranged from 804 to 997 accumulated degree hours (ADH) across models.
- Bacterial 16S marker models showed better performance than fungal ITS models (p = 0.006).
- Environmental data inclusion reduced MAE for ITS models and improved 16S models at higher taxonomic levels (phylum, class).
Conclusions:
- Soil microbial succession demonstrates some predictability for human decomposition stages.
- Environmental factors significantly influence the accuracy of microbial-based PMI estimations.
- Further research is needed to reduce error rates for reliable forensic application, especially with diverse donor populations.
