DEVELOPING NOVEL MICROBIAL TRACKING CAPABILITIES TO IDENTIFY FECAL POLLUTION SOURCES (Invited Presentation)

Urban waters, many of which are used as a source of drinking water or recreation are contaminated frequently by fecal pollution. There are many different sources that can contribute to this fecal contamination, but these sources carry different human health risks. For instance, in urban waters sewage, birds, pets, urban wildlife, or upstream agricultural areas are all common contaminating sources. Among these sources, sewage represents the biggest risk to human health because humans are both the primary contributor to sewage and the primary reservoir of human pathogens. Since mixed fecal sources with varying human heath risks contaminate urban waters, identifying the fecal pollution source is key to understanding the risk and identifying avenues for pollution mitigation strategies. To improve our ability to source identify fecal pollution in the environment, we used 16S rRNA gene sequencing to deeply characterize the microbiota of sewage systems across the U.S. and compared these data to relevant microbiota profiles of animal fecal samples. From these data, we used the random forest algorithm to build fecal source classifiers capable of identifying source-specific fecal pollution signatures in water samples. Here I will discuss the building of our 16S rRNA gene databases from sewage and animal fecal sources, the application of random forest to pollution source tracking, and our tool’s performance on artificial and real water samples containing mixed pollution sources.