Species sorting and community-level functionality in a natural microbial community: experiments using a traditional fermented dairy ecosystem model from Zambia

Microbial communities are continuously faced with environmental changes, leading to unpredictable states in their structure and functioning. Species sorting has been one of the important concepts pursued to understand the microbial community assembly process, which views member species as being driven by their differential fitness traits in response to selection. Several model systems have been applied regarding the study of species sorting, yet the complexity of field-based models often attracts rigorous and costly procedures during data capture, and often overlooks sampling over long temporal scales, while rarely providing community-functional relationships. This motivates the use of traditional fermented milks as simple model systems, given that they constitute self-assembled microbial communities through spontaneous fermentation, therefore representative of natural systems, yet are culturable even under laboratory conditions, and their community-level functionality can be tracked over a prolonged period. This study explored species sorting and community functionality using mabisi, a traditional fermented milk found in Zambia. Specifically, we tracked how microbial community structure shifts following exposure to novel substrates (Chapter 2 and 3) and processor location (Chapter 4) as different ecological selective factors through 16S rDNA sequencing. Alongside, we tracked community-level functionality through proxy analysis of volatile organic compounds (VOC), pH, consistency, and B-group vitamins. We found that the microbial community shifted with variation in novel substrate treatments and processor location. This shift was also time-dependent, conforming to early and late phases of propagation. However, the overall functionality was redundant, although it occurred to varying degrees. For instance, microbial communities from Mongu outperformed those from Kaoma for the acidification process and B-group vitamin production, manifesting as relatively lower median pH, and contributing more Vitamin B9 and B1 through PCA analysis. The findings in this thesis provide a broad significance spanning the microbial ecology and biotechnology interface. Firstly, it illustrates species sorting and prediction of community-level functionality through the interaction of substrate, time, and geographical location. Secondly, novel milk substrates can be used for mabisi processing, given mabisi microbial communities have adapted and exhibited functions similar to mabisi from raw cow milk based substrates. As such, mabisi microbial communities can be a bio-enrichment source for prebiotic and probiotic properties for infant formula, and we highlight the feasibility and effectiveness of the current practice of its use in treating malnourished children.