Dokument

Summary:
While it is known that microbes play many indispensable roles in ecosystems, the relationship between microbiomes and their environment is far from being well-understood. In part, this is the case because the methods necessary for studying environmental microbiomes, such as Next- Generation Sequencing and high-dimensional Machine Learning, have been developed relatively recently. However, the complex nature of ecosystems and environmental microbiomes acts as a further barrier to progress in this field of research. This thesis develops methods and concepts used to gain insight into the ecology of micro- biomes in lakes. It is based around two metabarcoding datasets sampled from lakes in Austria and the whole of Europe, respectively, and attempts to elucidate the microbiome’s relationship to environmental parameters. To this end, a tool for GPS-based dataset enhancement and a ma- chine learning framework for measuring microbiome covariation is developed. Building on this, the latent structure of the microbiome is estimated. In the discussion, a novel theory of informa- tion transmission in complex environments is described. Taken together, the work included herein presents a thorough analysis of the European lake microbiome that takes the complexity of the study object into account. The results point to- wards parameters that act as drivers of lake microbiome structure as well as microorganisms that might act as keystone species for ecosystem functioning. Furthermore, this work might provide the basis for considerable future progress in the study of environmental microbiomes.

Das Dokument ist im Internet frei zugänglich - Hinweise zu den Nutzungsrechten