Automation of multi-omics data generation and interpretation for exploratory metabolism research

Biology is slowly changing from a qualitative discipline to a data-driven quantitative field. Especially molecular bioscience is driven by high throughput- or large scale studies which generate a lot of data. Manual sample generation and -preparation become limiting factors because of the scale of laboratory work a person can perform and since operator effects limit reproducibility. This creates a need to minimize variability while still increasing throughput. Laboratory automation can help; specialized equipment minimizes experimental variation and can increase the scale of possible studies.

The produced data needs to be analyzed, which is why robust methods to process, interpret and contextualize data need to be established. Focussing on metabolism research, both a detailed understanding of specific mechanisms and big picture systems-level view are required to get a good understanding of the studied metabolic effect or -rearrangement. In this thesis I present a combination of different approaches to streamline multi-omics data driven exploratory metabolism research. The presented methods were designed for cell systems and are largely organism agnostic. They include automated genome-scale system
perturbation, automation of cultivation and multi-omics sampling and the subsequent analysis of multi-omics data. Furthermore, a comparison of different approaches to integrate multiple types of biochemical information is presented using thermogenic lipolysis in murine brown adipocytes as a biological model. Brown adipocytes are a type of fat cell that is involved in non-shivering cold response by producing heat through increased energy expenditure. The results of the multi-omics comparison suggest multiple biochemical elements that might be involved in this mechanism as targets for follow up studies.