Microbial communities may now be studied in more detail using culture-independent methods such as metagenomics (directly analyzing genomes from an environmental sample). One of the many potential applications of metagenomics is in the assessment of water quality. Current methods for detection of fecal pollution in water rely on culture-based microbial testing which is slow and can lack sensitivity and specificity. For the WatershedDiscovery.ca project, it was hypothesized that a molecular-based test, developed based on metagenomics analysis, may be more rapid, and accurate for characterizing fecal pollution. Many bioinformatics methods for metagenomics sequence classification have been developed, but when initiating this research, no comprehensive evaluation of method accuracy had yet been published. Thus, using both in silico and in vitro simulated communities, a comprehensive evaluation of metagenomics taxonomic sequence classification methods was performed. Utilizing knowledge gained from this comparative evaluation, a study was undertaken of microbial community dynamics in monthly water samples from sites in urban, protected, and agricultural watersheds collected over a one-year period. Freshwater samples collected from sites affected by agricultural activity showed distinct microbial profiles versus samples collected from unaffected sites, and a notable presence of Legionella was discovered in all watershed sampling sites (the largest study of Legionella in watersheds to date). Furthermore, biomarkers were developed that could distinguish agriculturally affected samples from pristine samples collected in our watershed study. Finally, there is a lack of methods for the prediction of subcellular localization (SCL) from metagenomics sequences—of interest for the identification of cell surface/secreted proteins for development of ELISA-based diagnostics and other applications. Thus, PSORTb, a precise bacterial and archaeal SCL program, was modified to enable the classification of metagenomics sequences, and applied to the analysis of the watershed samples. A database of protein SCL associated with PSORTb was expanded to make it suitable for a wider diversity of microbes, particularly those with atypical cell envelopes. Collectively this work expands our understanding of metagenomics software accuracy, and available analysis tools, and provides insight into freshwater microbial community dynamics, with potential application in water quality test development.
Copyright is held by the author.
This thesis may be printed or downloaded for non-commercial research and scholarly purposes.
Supervisor or Senior Supervisor
Thesis advisor: Brinkman, Fiona S. L.
Member of collection