Abstract

Background Microbial clocks have shown high accuracy in estimating time since death, or post-mortem interval (PMI), across diverse contexts, including controlled laboratory settings, outdoor environments, and both human cadavers and animal models. However, the limited number of studies performed under different environmental and climatic conditions, particularly in Europe, restricts the broader applicability of these models, as data from distinct climates may not reliably translate across geographic regions. This study investigates the nasal thanatomicrobiome of pig carcasses exposed to outdoor summer conditions in the United Kingdom to assess the potential of microbial succession as a PMI estimator in a temperate climate. Model generalisation was further evaluated through an independent external validation on a published human thanatomicrobiome dataset spanning three geographically distinct states of the USA and all four seasons. Results Three pig carcasses were exposed for 10 days in June 2023, with duplicate swabs collected from both internal and external nares across 26 time points, yielding 312 samples in total. Microbial communities were profiled by 16S rRNA gene amplicon sequencing on an Illumina MiSeq platform; reads were processed in QIIME2 and taxonomy was assigned against the SILVA v138 database. A Random Forest regression model trained on combined internal and external nasal data achieved the highest predictive accuracy for PMI estimation (R² = 0.96, MAE = 8.83 h), outperforming models built on external or internal swabs alone. Ambient temperature, which remained overall stable throughout the study period, was not a significant predictor under these conditions. Dominant taxa associated with decomposition succession included Moraxella, Myroides, Ignatzschineria, Acinetobacter, Vagococcus, Pasteurella, Clostridium, and Savagea; the prominence of Ignatzschineria and Myroides in particular is consistent with insect-associated microbial transfer, suggesting that carrion fly activity significantly shaped the microbial community succession at this site. External validation on a published human facial skin thanatomicrobiome dataset, sampled across all four seasons, confirmed robust model performance with an R² = 0.594 and MAE = 30.3 ADD. Conclusions This study demonstrates that thanatomicrobiome-based PMI estimation can achieve high accuracy under temperate European conditions, and that results can be extended beyond the original experimental conditions, with model performance remaining robust across host species, geographic locations, anatomical location and inter-laboratory variation. However, predictive accuracy was optimised when models were trained on locally derived data, underscoring the importance of generating climate-specific models to improve the robustness and forensic applicability of microbial clocks in casework. By expanding reference datasets across diverse climatic environments and performing external validations, thanatomicrobiome-based microbial clocks have the potential to become a reliable and court-admissible tool for PMI estimation in forensic investigations.