The utility of drowning site inference through metagenomic diatom analysis

Abstract

The diatom test is one of the methods used to diagnose drowning in forensic autopsies. Metagenomic diatom analysis may reveal where a drowning occurred. We evaluated whether metagenomic diatom analysis could be used to infer waters, watersheds, and geographic locations using 166 water samples from 64 locations (freshwater: 55; seawater: 9). Principal component analysis (PCA) in all samples revealed no specific clusters for waters or watersheds. In one river, the three samples at the same site generally tended to be in close clusters, but there were some cases where the three sites were far from each other. The precise geographic location could thus not be reliably identified. However, PCA of data from dams, lakes, and retention basins revealed sites with independent clusters, suggesting unique diatom compositions. Diatoms of seawater were not detected in freshwater. The high number of Actinoptychus, Chaetoceros, and Skeletonema detected in seawater samples suggested that they are useful for seawater identification. This method required only 2 mL of water; it suggests that this method can be applied to actual samples. In summary, it was difficult to infer the geographic location and waters or watersheds, but the freshwater/seawater distinction could be easily made, and depending on the application, it may be useful in forensic science practice.