Experimental and computational approaches for deep metabolome annotation with application to the ecotoxicological...

Comprehensively characterizing the metabolomes of model organisms with high coverage and confidence is a critical step towards interpreting the metabolic basis of human and environmental health, yet there are formidable challenges involved in annotating metabolomes. A wide range of genotypes and phenotypes should be sampled with multiple complementary analytical approaches to cover the large and dynamic biochemical space they exhibit. In addition, multiple computational tools and approaches are required to annotate the metabolites from raw analytical data.

To address this, we developed the deep metabolome annotation (DMA) workflow. Applied to the ecological sentinel species, Daphnia magna, a pooled sample comprising 10 distinct strains exposed to both normal and stressed environmental conditions was extracted and systematically physicochemically separated via solid-phase extraction and liquid- and gas-chromatography prior to extensive multiple-stage mass spectrometric fragmentation, generating >8,000 raw data files, and supplemented by nuclear magnetic resonance spectroscopy. An extensive Galaxy-based computational approach was built to analyse these data, comprising >30 tools. The overall DMA efforts resulted in 8,181 annotated polar metabolites and lipids in D. magna, with the raw and processed data, tools, and annotations disseminated freely via public data repositories and a custom web-based interface to maximize reusability.

The DMA workflow has generated one of the largest metabolome annotation datasets for any non-human model organism and provides the first in-depth characterization of the D. magna metabolome, serving as both a resource and a valuable catalyst for future DMA studies of other model organisms.