Application of Metabolomics and Machine Learning for the Prediction of Postmortem Interval.

Determining postmortem interval (PMI) during forensic investigations is essential and challenging. The traditional methods used to predict PMI, such as algor mortis, rigor mortis, livor mortis, and decomposition changes, involve large margins of error, particularly when the person's death has occurred more than 48 hours ago. Organs and tissues experience profound biochemical and metabolomic changes after death. As such, new approaches are required to enhance the prediction of PMI. Novel developments in forensic sciences are focusing on identifying and analyzing postmortem metabolomics, which are biomarkers found in different body fluids and tissues serving as a "fingerprint" of continuous processes affected by both external and internal factors. This variability complicates the dataset, making examination challenging. Hence, the application of machine learning technology offers the capability to navigate through the complexities of metabolomic data, uncover hidden correlations, and enhance the accuracy of PMI prediction in forensic science. This article explores and assesses the new methodology that has recently been used to enhance the prediction of PMI by analyzing postmortem metabolomics' changes and applying these data to machine learning models. This development provides a significantly more reliable process that could potentially decrease the margin of error compared to the traditional methods used for PMI prediction.