PhD Defence Mathies Brinks Sørensen

Principal Supervisor
Associate Professor Charlotte Held Gotfredsen, DTU Chemistry
Co-supervisor
Associate Professor Mikael Lenz Strube, DTU Bioengineering
Examiners
Professor Jens Ø. Duus, DTU Chemistry
Professor Tim D. W. Claridge, Exscientia and University of Oxford, England
Professor Morten Arendt V. Rasmussen, University of Copenhagen, Denmark
Chairperson
Associate Professor Sebastian Meier, DTU Chemistry

The remarkable natural products produced by microbial communities, including secondary metabolites that hold great promise within for example medicine and agriculture, have long intrigued scientists. While studying these communities in controlled laboratory settings has provided valuable insights, it falls short of capturing the complexity of interactions occurring within their natural habitats. Consequently, researchers have sought to understand natural microbial interactions to unlock the potential of novel secondary metabolites, leading to the development of groundbreaking techniques.

In a recent thesis, an innovative approach utilizing nuclear magnetic resonance (NMR) spectroscopy has emerged as a powerful tool for studying microbial metabolites. The thesis addresses two key challenges in targeted NMR in-situ detection. Firstly, to eliminate bias and ensure accurate analysis of complex metabolomic spectral samples, the author developed the NMR-Onion framework. This Python/Pytorch-based framework enables automatic detection, quantification, and uncertainty evaluation of peaks within NMR spectra. Its wide range of capabilities accommodates varying signal-to-noise ratios and minimizes the risk of drawing false conclusions by accounting for sample-to-sample variations.

The second challenge tackled in the thesis involved establishing a robust workflow for metabolomic data generation and analysis. By combining Design of Experiments (DoE), statistical quality control (SQC), minimal preprocessing, automatic detection (via NMR-Onion), and statistical analysis, the author devised an efficient metabolomic workflow. This workflow was successfully applied in a case study, where it linked targeted amplitudes to specific amino acids, highlighting the potential for further discoveries in this field.

In conclusion, the thesis presents an exciting breakthrough in NMR-based metabolite detection and analysis. The NMR-Onion framework revolutionizes the field by automating the detection, quantification, and evaluation of NMR signals, significantly reducing operator bias. When coupled with the proposed metabolomic workflow, optimal results can be achieved in metabolomic data analysis. This approach opens doors for exploring the metabolomic landscape of natural microbial communities, offering insights into the diversity and functions of secondary metabolites. Furthermore, the NMR-Onion framework holds immense potential for applications beyond microbial studies, including disease diagnostics, agriculture, and food science. The future looks promising as scientists delve deeper into the secrets hidden within microbial interactions, unraveling the extraordinary world of natural products