LAUSR

The efficacy of SourceTracker software to attribute contamination from a variety of fecal sources spiked into ambient freshwater samples was investigated. Double-blinded samples spiked with ≤5 different sources (0.025-10% vol/vol) were evaluated against fecal taxon libraries characterized by next-generation amplicon sequencing. Three libraries, including an initial library (17 nonlocal sources), a blinded source library (5 local sources), and a composite library (local and nonlocal sources), were used with SourceTracker. SourceTracker's predictions of fecal compositions in samples were made, in part, based on distributions of taxa within abundant genera identified as discriminatory by discriminant analyses but also using a large percentage of low abundance taxa. The initial library showed poor ability to characterize blinded samples, but, using local sources, SourceTracker showed 91% accuracy (31/34) at identifying the presence of source contamination, with two false positives for sewage and one for horse. Furthermore, sink predictions of source contamination were positively correlated (Spearman's ρ ≥ 0.88, P < 0.001) with spiked source volumes. Using the composite library did not significantly affect sink predictions ( P > 0.79) compared to those made using the local sources alone. Results of this study indicate that geographically associated fecal samples are required for SourceTracker to assign host sources accurately.