Tracing recent outdoor geolocation by analyzing microbiota from shoe soles and shoeprints even after indoor walking

The microbial communities on shoe soles and shoeprints could carry microbial information about where someone walked. This is possible evidence to link a suspect in a crime case to a geographic location. A previous study had shown that the microbiota found on shoe soles depend on the microbiota of the soil on which people walk. However, there is a turnover of microbial communities on shoe soles during walking. The impact of microbial community turnover on tracing recent geolocation from shoe soles has not been adequately studied. In addition, it is still unclear whether the microbiota of shoeprints can be used to determine recent geolocation. In this preliminary study, we investigated whether the microbial characteristics of shoe soles and shoeprints can be used to trace geolocation and whether this information can be destroyed by walking on indoor floors. In this study, participants were asked to walk outdoors on exposed soil, then walk indoors on a hard wood floor. High-throughput sequencing of the 16S rRNA gene was performed to characterize the microbial communities of shoe soles, shoeprints, indoor dust, and outdoor soil. Samples of shoe soles and shoeprints were collected at steps 5, 20, and 50 while walking indoors. The PCoA result showed that the samples were clustered by geographic origin. The shoeprint showed a more rapid turnover of microbial community than the shoe sole during indoor walking. The result of FEAST showed that the microbial communities of shoe sole and shoeprint were mainly (shoe sole, 86.21∼92.34 %; shoeprint, 61.66∼90.41 %) from the soil of the outdoor ground where the individual recently walked, and a small portion (shoe sole, 0.68∼3.33 %; shoeprint, 1.43∼27.14 %) from the indoor dust. Based on the matching of microbial communities between geolocation and shoe sole or shoeprint, we were able to infer the recent geolocation of the individual with relatively high accuracy using the random forest prediction model (shoe sole: 100.00 %, shoeprint: 93.33∼100.00 %). Overall, we are able to accurately infer the geolocation of an individual’s most recent outdoor walk based on the microbiota of shoe sole and shoeprint, even though these microbiotas show a turnover when walking indoor floor. The pilot study was expected to provide a potential method for tracing recent geolocation of suspects.