Ex vivo metabolomics

Functional annotation is still the bottleneck to make genome data suitable for biological interpretation.
To overcome this gap we developed an approach, where natural substrates and products of uncharacterized enzymes can be identified by bringing the purified enzyme back into its quasi-native substrate environment. The extraction of almost all metabolites from an appropriate tissue at a specific developmental and physiological status, where the protein of interest is expressed (e.g., after stress), ensures the availability of the native substrate(s). This can make substrates accessible in an un-biased way for the enzyme of interest and even provide substrates that would not be available commercially or by chemical synthesis.
Next, non-targeted metabolomics is used to identify characteristic substrate-product-pairs by comparing the metabolite pattern of the extracts incubated with and without active enzyme. Identification of substrate-product-pairs from large and noisy LC-HRMS data sets is supported by a computational framework for the recognition of pairs with an inverse intensity pattern and a particular mass shift.
Ex vivo metabolomics therefore provides a hypothesis-free approach for functional gene annotation in model and non-model organisms.