Determining the postmortem interval or PMI is one of the most difficult issues in forensic medicine. Traditional methods usually rely on visible changes in the body, corpse temperature, and signs of decomposition, but these indicators can be affected by different environmental conditions.
Researchers at Huazhong University of Science and Technology in China, in a new study, attempted to use a different information source: the microorganisms that remain active on and inside the body after death.
After death, the body's biological activity stops, but microbial life does not end immediately. During decomposition, the composition of bacterial communities gradually changes. Some species increase in the early stages of decomposition and others become dominant in later stages.
The researchers likened these changes to a kind of “biological clock”; a clock that, as decomposition progresses over time, shows a different pattern of microbial activity.
Training mHolmes with daily sampling from corpses
The research team used data from 34 human corpses to develop mHolmes. Skin microbiome samples were collected daily over a 21-day period from two regions: the face and the pelvis.
Using these data, the AI model learned how the composition of microbial communities changes over time and then used these patterns to estimate the time elapsed since death.
One interesting feature of this system is its ability to reconstruct some microbial changes related to periods for which there was no direct sample. Thus, mHolmes does not merely compare existing data with one another; it can also model the trajectory of microbial changes.
The system can also use information obtained from one body region to predict likely changes in another region; a capability that can be important when only a limited part of a corpse’s remains is available.
Error under two days in some tests
Test results showed that the mean error of mHolmes in estimating the time since death across different body locations was less than two days. This rate represents a significant improvement compared to some previous microbiome-based methods.
The researchers also identified seven bacterial communities associated with different stages of decomposition. This is important because the model does not provide only a final number; it is possible to examine the relationship between some of its predictions and the biological changes that occurred during decomposition.
mHolmes is not yet ready for courtroom use
Despite promising results, mHolmes still cannot definitively determine the exact time of death. The current study was conducted on only 34 corpses, two body regions, and a 21-day period, and multiple environmental factors can affect the rate of decomposition and the microbiome composition.
Temperature, humidity, type of body covering, insect activity, and environmental conditions are among the factors that may alter the trajectory of microbial changes. Therefore, to determine how effective this technology is in real-world conditions, broader studies with more samples and varied environments are needed.
If these trials succeed, microbiome analysis could become, alongside existing methods, another tool for forensic medicine. In such an approach, bacterial changes in the body, instead of merely being part of the decomposition process, could be used as a temporal marker to estimate the time elapsed since death.