mHolmes improves cross anatomical cadaveric microbiome forecasting for postmortem interval estimation

Researchers developed mHolmes, a Transformer-based AI model trained on daily microbiome data from 34 cadavers, to improve estimation of postmortem intervals (PMI) with less than 2 days error across different body sites. This model enhances forensic analysis by forecasting microbial succession with higher precision and interpretability, identifying key bacterial markers linked to PMI.