Faecal source apportionment using molecular methods: A proof of concept using the FEAST algorithm

Abstract

Faecal contamination of freshwater and marine environments represents a significant risk for public health, recreational activity and food safety, and tools for evaluating complex multi-source contamination remain largely in the development phase. We evaluated the efficacy of the Fast Expectation Maximization (FEAST) microbial source tracking (MST) algorithm to apportion sources of faecal contamination among four mammalian species of interest in coastal waters in New Zealand. Using 16S ribosomal DNA metabarcoding of faecal samples from cows, fur seals, and sheep, as well as human wastewater, we aimed to differentiate and quantify the contribution of these sources in mixed faecal samples. Multivariate analysis confirmed significant differences in the microbial communities associated with each mammalian source, with specific bacterial classes indicative of different sources. The FEAST algorithm was tested using mixed DNA and mixed faecal samples, and we found that the algorithm correctly assigned the dominant source from all samples, but underestimated the dominant source's proportional contribution. This underestimation suggests the need for further refinement and validation to ensure accurate source apportionment in environmental samples where the faecal signal is likely to be a minor component. Despite these limitations, the findings of our study, in combination with the evidence from others who have tested the FEAST algorithm in environmental settings, indicates that it represents an advance on existing tools for microbial source tracking and may become a useful addition to the toolbox for environmental management.