Chemical space exploration: targeting the undiscovered

Medicinal chemists and pharmaceutical companies have often relied on natural products (NPs) as a source of biologically active molecular scaffolds.  While genomics and metabolomics allow for high throughput characterization of potentially novel NPs from large culture collections and environmental isolates, we lack tools for directly targeting unexplored chemical space, as evidenced by high rates of rediscovery.  We hypothesize that microbial cultures containing potentially novel chemistries can be prioritized for in-depth characterization based on molecular patterns revealed in a high throughput mass spectrometry survey of the sampled chemical space.  We surveyed the metabolomic profiles of 1000 marine Actinobacteria isolates with a wide range of phylogenetic and geographic diversity.  Small scale cultures were extracted and subjected to profiling with high resolution tandem mass spectrometry to generate over 1 million fragmentation patterns. Molecular fragmentation patterns were grouped, scored, and organized into a molecular network on GNPS (gnps.ucsd.edu).  Visualizing this network allowed us to prioritize several strains for further characterization based on their contribution to the chemical space of the sample set.  Our network contained 835 (19%) strain specific chemistries. Nearly one third of these unique signals were contributed by just five strains, in which the classically prodigious genus of Streptomyces was overrepresented.  This molecular networking based analysis of the sampled chemical space can rapidly highlight potential sources of novel chemistry within a sample set.   We believe high throughput mass spectrometry guided exploration of chemical space will significantly accelerate NP discovery through simple sample prioritization and its effectiveness should increase with sample size.