Mapping the chemical space of natural biomes by combining non-targeted LC-HRMS datasets

Understanding the microbial world depends on measuring its biological and chemical features. While metagenomics has enabled charting microbes in complex communities at genomic resolution, methods to comprehensively describe the chemical space are still missing. Given the diversity of chemical compounds, only a fraction (<5%) of the compounds detected in non-targeted metabolomics experiments can be reliably identified. Also, due to instrumental limitations, the uncovered chemical scape does not fully align with the desired chemical space. To compile metabolomics data from samples measured under different experimental conditions, a comprehensive comparison needs to be done. Since the identification of the compounds is the bottleneck in this process, I will use molecular fingerprints and chemical classes to define the chemical space beyond individual annotations. Thus, I will provide a comprehensive representation that can be used in models of microbiome functioning and dynamics.