Since the 1930s, scientists have collected bacteria from the environment and used their associated specialized metabolites (SM) as a source for drug discovery. However, phylogenetic and chemical redundancy in bacterial libraries resulted in a high re-discovery rate of known compounds, and consequently a large divestment in natural product drug discovery. The success of a discovery program is dependent on using libraries with a high diversity of taxa and natural products, however current practice relies entirely on colony morphology and/or 16S rRNA gene sequencing analysis to decide which isolated strains to retain for addition to a library. Importantly, this is not indicative of a strain?s ability to produce SMs. Therefore, the development of a semi-automated, open-access platform to rapidly organize unknown bacterial isolates based on phylogeny and SMs would greatly increase the efficiency of the front-end of drug discovery. This technology has broad future applications, given the interconnectedness of microbiomes, such as in humans and in botanical science. Our lab recently developed such a platform, called IDBac, which uses MALDI-TOF MS and custom software that allows fingerprinting of up to 384 bacterial colonies in four hours by a single user.15 However, while proven to work on relatively small subsets of strains, we have yet to establish precedence for this technology in building low-redundancy bacterial libraries from high numbers of samples (hundreds to thousands). The current proposal aims to develop a framework within IDBac for large, personalized database creation and spectra searching/matching. We will also enable smarter sample collection strategies by constructing IDBac tools to allow researchers the ability to a) build spectra databases of known ?seed? strains for identifying unknown isolates, and b) visualize and study the interplay of bacterial phylogeny and specialized metabolism. Successful completion of these aims will contribute one of the few major advances to the front end of microbial library generation in nearly eight decades, and will enable researchers studying the microbiome of humans, plants, botanicals, and other sources to build custom microbial libraries for rapid in-house strain identification and characterization.27, 28
While natural products have been a successful source of new medicines, current methods to find new drug-leads provide inadequate prioritization and result in unmanageable rates of compound rediscovery. This project seeks to build an innovative framework to maximize the taxonomic and chemical diversity of natural product libraries while minimizing the number of microbial strains entering the pipeline. The successful completion of this project will lessen the effort and resources needed to build, screen, and maintain a diverse bacterial natural products library and address a major unmet need in natural product drug discovery and microbiome analysis.
