Influence of Library Composition on SourceTracker Predictions for Community-Based Microbial Source Tracking.

Area of Science:

• Environmental microbiology
• Molecular ecology
• Bioinformatics

Background:

• Microbial source tracking (MST) is crucial for identifying fecal contamination.
• High-throughput DNA sequencing enables community profiling for MST.
• SourceTracker is an emerging computational tool for community-based MST.

Purpose of the Study:

• To evaluate SourceTracker's efficacy in identifying known fecal contamination sources.
• To assess the impact of autochthonous taxa on SourceTracker predictions.
• To determine factors influencing SourceTracker's accuracy in real-world scenarios.

Main Methods:

• Utilized in situ mesocosms spiked with known fecal sources.
• Applied SourceTracker to analyze microbial community data from sources and sinks.
• Compared SourceTracker predictions with known source contributions.
• Investigated the effect of including or excluding autochthonous taxa in the analysis.

Main Results:

• SourceTracker successfully predicted most fecal sources in the mesocosms.
• Fecal source library composition significantly affected prediction accuracy.
• Predictions were more reliable with curated libraries, accounting for autochthonous taxa, and low source variability.
• SourceTracker identified abundant sources but struggled with closely related ones.

Conclusions:

• SourceTracker is a valuable tool for community-based microbial source tracking.
• Library curation and accounting for local microbial communities enhance SourceTracker's reliability.
• Further refinement may be needed for differentiating sources with similar microbial profiles.